ABSTRACT
Title of dissertation: HOVER AND WIND-TUNNEL TESTING
OF SHROUDED ROTORS FOR IMPROVED
MICRO AIR VEHICLE DESIGN
Jason L. Pereira
Doctor of Philosophy, 2008
Dissertation directed by: Professor Inderjit Chopra
Department of Aerospace Engineering
The shrouded-rotor configuration has emerged as the most popular choice for
rotary-wing Micro Air Vehicles (MAVs), because of the inherent safety of the design
and the potential for significant performance improvements. However, traditional
design philosophies based on experience with large-scale ducted propellers may not
apply to the low-Reynolds-number (?20,000) regime in which MAVs operate. An
experimental investigation of the effects of varying the shroud profile shape on the
performance of MAV-scale shrouded rotors has therefore been conducted. Hover
tests were performed on seventeen models with a nominal rotor diameter of 16 cm
(6.3 in) and various values of diffuser expansion angle, diffuser length, inlet lip radius
and blade tip clearance, at various rotor collective angles. Compared to the baseline
open rotor, the shrouded rotors showed increases in thrust by up to 94%, at the same
power consumption, or reductions in power by up to 62% at the same thrust. These
improvements surpass those predicted by momentum theory, due to the additional
effect of the shrouds in reducing the non-ideal power losses of the rotor. Increasing
the lip radius and decreasing the blade tip clearance caused performance to improve,
while optimal values of diffuser angle and length were found to be 10? and 50% of
the shroud throat diameter, respectively. With the exception of the lip radius, the
effects of changing any of the shrouded-rotor parameters on performance became
more pronounced as the values of the other parameters were changed to degrade
performance.
Measurements were also made of the wake velocity profiles and the shroud
surface pressure distributions. The uniformity of the wake was improved by the
presence of the shrouds and by decreasing the blade tip clearance, resulting in lower
induced power losses. For high net shroud thrust, a favorable pressure distribution
over the inlet was seen to be more important than in the diffuser. Strong suction
pressures were observed above the blade-passage region on the inlet surface; taking
advantage of this phenomenon could enable further increases in thrust. However,
trade studies showed that, for a given overall aircraft size limitation, and ignoring
considerations of the safety benefits of a shroud, a larger-diameter open rotor is
more likely to give better performance than a smaller-diameter shrouded rotor.
The open rotor and a single shrouded-rotor model were subsequently tested
at a single collective in translational flight, at angles of attack from 0? (axial flow)
to 90? (edgewise flow), and at various advance ratios. In axial flow, the net thrust
and the power consumption of the shrouded rotor were lower than those of the open
rotor. In edgewise flow, the shrouded rotor produced greater thrust than the open
rotor, while consuming less power. Measurements of the shroud surface pressure
distributions illustrated the extreme longitudinal asymmetry of the flow around the
shroud, with consequent pitch moments much greater than those exerted on the
open rotor. Except at low airspeeds and high angles of attack, the static pressure in
the wake did not reach ambient atmospheric values at the diffuser exit plane; this
challenges the validity of the fundamental assumption of the simple-momentum-
theory flow model for short-chord shrouds in translational flight.
HOVER AND WIND-TUNNEL TESTING OF SHROUDED
ROTORS FOR IMPROVED MICRO AIR VEHICLE DESIGN
by
Jason L. Pereira
Dissertation submitted to the Faculty of the Graduate School of the
University of Maryland, College Park in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
2008
Advisory Committee:
Professor Inderjit Chopra, Chair/Advisor
Professor James D. Baeder
Professor Derek Boyd
Professor Darryll Pines
Professor Norman M. Wereley
c? Copyright by
Jason L. Pereira
2008
Preface
This research was primarily funded by Multidisciplinary University Research
Initiative (MURI) grant W911NF0410176 from the United States Army Research
Office (ARO).
ii
Dedication
To my parents, Owen and Sherry, who showed the greatest understanding and
restraint by never once uttering the words, ?How is your thesis was coming along??,
always instead leaving it to me to tell them if I so chose.
iii
Acknowledgements
Thanks are due to several people, without whom this research, and the lessons
that I have learned as a graduate student, would never have occurred. First and
foremost, to my advisor, Dr. Inder Chopra, who stood by me and continued to
support and guide me through all the years it took me to complete this research,
despite the many, many mistakes that I made. Stern, yet also gentle and forgiving,
and with uncommon patience, he has been a father-figure to me in my professional
academic career. I am grateful also to the other members of my advisory committee
? Dr. Darryll Pines, Dr. Norm Wereley, Dr. Jim Baeder and Dr. Derek Boyd ?
for their insightful comments and suggestions regarding my research. To Dr. Pines
I owe special thanks for his enthusiasm, encouragement and motivational pep talks,
and for many discussions about my future professional career.
To my fellow graduate students and co-workers: Andy Bernhard, the first
graduate student I ever worked with, and who has been my inspiration ever since;
Paul Samuel, Felipe Bohorquez, Jayant Sirohi, Beerinder Singh, Ashish Purekar
and Jeanette Epps, from whom I learned the fundamentals of conducting labo-
ratory research; Shreyas Ananthan, Manikandan Ramasamy, Sandeep Gupta and
Karthikeyan Duraisamy, who patiently answered all my questions about fluid dy-
namics; Paul Samuel (again), Jinsong Bao, Anubhav Datta and Shaju John, for sup-
port, encouragement and general ?graduate-student-advice?, especially when times
were rough; Nicholas Rosenfeld, Carlos Malpica, Joseph Conroy, Brandon Bush,
Brandon Fitchett, Monica Syal, Abhishek Abhishek, and Felipe, Karthik and Mani
(again), for the camaraderie and ?extra-curricular?, recreational bonding that helped
make my graduate-school experience an enjoyable one. These people, and numer-
ous others, too many to mention here but remembered nontheless, have given me
countless happy memories of my years at the Alfred Gessow Rotorcraft Center.
Similarly, to Bernard ?Bernie? LaFrance and Howard ?Howie? Grossenbacher,
iv
who taught me how to work with metal and wood, and thereby helped me become
a more well-rounded engineer, and to Dr. Vengalattore ?VT? Nagaraj, Dr. Marat
Tishchenko and Dr. Jewel Barlow, for their support and professional advice. Also to
Pat Baker, Becky Sarni, Debora Chandler and the rest of the Aerospace Department
staff, who smoothly managed all the administrative paperwork involved in managing
a graduate student, and who always greeted me with smiling faces and friendly waves
whenever I entered the AE Main Office.
Finally, and most importantly, to my family and friends, who rejoiced in my
achievements and who provided me with the emotional support and encouragement
to get through periods of immense bleakness and exhaustion: Minyoung Kim, Erin
Loeliger, Deepti Gupta, Payal Ajwani, Les Yeh and Arun Arumugaswamy; my par-
ents, Owen and Sherry, and my siblings, Tracy and Darren; my cousins, Kyle and
Lauren, and my aunt and uncle, Letty and Loy, who took me into their family and
gave me a home away from home; and my other aunts and uncles ? Lisa-Ann, TJ,
Bryan, Madeline, Neal and Rosemary. Without all of you, I would never have made
it to this finish line.
v
Table of Contents
List of Tables ix
List of Figures x
Nomenclature xvi
1 Introduction 1
1.1 Background: Micro Air Vehicles . . . . . . . . . . . . . . . . . . . . . 1
1.2 The need for more efficient hover-capable MAVs . . . . . . . . . . . . 3
1.3 The shrouded-rotor configuration: Potential for improved performance 10
1.4 Previous research in ducted propellers and shrouded rotors . . . . . . 26
1.4.1 Historical overview . . . . . . . . . . . . . . . . . . . . . . . . 26
1.4.2 Experimental work: Effects of variations in shroud design . . . 32
1.4.2.1 The early work . . . . . . . . . . . . . . . . . . . . . 42
1.4.2.2 Helicopter tail rotors . . . . . . . . . . . . . . . . . . 58
1.4.2.3 Unmanned aircraft . . . . . . . . . . . . . . . . . . . 68
1.4.3 Experimental work: Tests of a single shrouded-rotor model . . 86
1.4.4 Analytical methods for performance prediction . . . . . . . . . 87
1.4.4.1 Blade-Element and Potential-flow methods . . . . . . 88
1.4.4.2 Computational-Fluid-Dynamics methods . . . . . . . 93
1.4.5 Other shrouded-rotor research . . . . . . . . . . . . . . . . . . 96
1.4.5.1 Noise considerations . . . . . . . . . . . . . . . . . . 96
1.4.5.2 Tip-gap flow physics . . . . . . . . . . . . . . . . . . 100
1.4.5.3 Shrouded-rotor UAV stability and control . . . . . . 101
1.4.5.4 Behavior of annular wings . . . . . . . . . . . . . . . 103
1.5 Low-Reynolds-number rotor aerodynamics . . . . . . . . . . . . . . . 103
1.6 Objectives and approach of current research . . . . . . . . . . . . . . 104
2 Experimental Setup 107
2.1 Shrouded-rotor models . . . . . . . . . . . . . . . . . . . . . . . . . . 107
2.2 Hover test setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
2.3 Wind-tunnel test setup . . . . . . . . . . . . . . . . . . . . . . . . . . 118
2.4 Test procedure and uncertainty analysis . . . . . . . . . . . . . . . . 126
3 Experimental Results: Hover Tests 131
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
3.2 Simplifying assumptions . . . . . . . . . . . . . . . . . . . . . . . . . 132
3.2.1 Assumption regarding rotational speed . . . . . . . . . . . . . 132
3.2.2 Assumption regarding rotor radius . . . . . . . . . . . . . . . 135
3.3 Performance measurements . . . . . . . . . . . . . . . . . . . . . . . . 136
3.3.1 General characteristics . . . . . . . . . . . . . . . . . . . . . . 138
3.3.2 Effects of changing shroud parameter values . . . . . . . . . . 155
3.3.2.1 Blade tip clearance . . . . . . . . . . . . . . . . . . . 163
vi
3.3.2.2 Inlet lip radius . . . . . . . . . . . . . . . . . . . . . 167
3.3.2.3 Diffuser angle . . . . . . . . . . . . . . . . . . . . . . 170
3.3.2.4 Diffuser length . . . . . . . . . . . . . . . . . . . . . 173
3.3.2.5 Diffuser expansion ratio . . . . . . . . . . . . . . . . 177
3.3.2.6 Relative effects of the different parameters . . . . . . 181
3.4 Shroud surface pressure measurements . . . . . . . . . . . . . . . . . 184
3.4.1 General characteristics . . . . . . . . . . . . . . . . . . . . . . 184
3.4.1.1 Inlet thrust distribution . . . . . . . . . . . . . . . . 191
3.4.2 Effects of changing shroud parameter values . . . . . . . . . . 194
3.4.2.1 Blade tip clearance . . . . . . . . . . . . . . . . . . . 196
3.4.2.2 Inlet lip radius . . . . . . . . . . . . . . . . . . . . . 198
3.4.2.3 Diffuser angle . . . . . . . . . . . . . . . . . . . . . . 200
3.4.2.4 Diffuser length . . . . . . . . . . . . . . . . . . . . . 202
3.4.2.5 Inlet thrust distribution . . . . . . . . . . . . . . . . 205
3.5 Wake axial velocity measurements . . . . . . . . . . . . . . . . . . . . 206
3.5.1 General characteristics . . . . . . . . . . . . . . . . . . . . . . 206
3.5.2 Effects of changing shroud parameter values . . . . . . . . . . 213
3.5.2.1 Blade tip clearance . . . . . . . . . . . . . . . . . . . 213
3.5.2.2 Diffuser length . . . . . . . . . . . . . . . . . . . . . 214
4 Experimental Results: Wind-Tunnel Tests 219
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
4.2 Shroud surface pressure measurements . . . . . . . . . . . . . . . . . 222
4.3 Performance measurements . . . . . . . . . . . . . . . . . . . . . . . . 231
4.3.1 Thrust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
4.3.2 Normal force . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
4.3.3 Lift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
4.3.4 Drag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
4.3.5 Location of center of pressure . . . . . . . . . . . . . . . . . . 241
4.3.6 Pitch moment . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
4.3.7 Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
5 Vehicle Configuration Trade Studies 249
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
5.2 Shroud weights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
5.3 Trade studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
5.3.1 Shrouded vs. open, using the baseline 6.3-inch rotor . . . . . . 251
5.3.2 Shrouded vs. open, using the cambered 6.3-inch rotor . . . . . 253
5.3.3 6.3-inch shrouded rotor vs. 9.5-inch open rotor . . . . . . . . . 254
6 Concluding Remarks 255
6.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
6.1.1 Hover: Performance measurements . . . . . . . . . . . . . . . 257
6.1.2 Hover: Flow-field measurements . . . . . . . . . . . . . . . . . 259
6.1.3 Translational flight . . . . . . . . . . . . . . . . . . . . . . . . 261
vii
6.2 Contributions to state of the art . . . . . . . . . . . . . . . . . . . . . 264
6.3 Recommendations for future work . . . . . . . . . . . . . . . . . . . . 266
A Theoretical Derivations 269
A.1 Momentum Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
A.1.1 Hover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
A.1.1.1 Open rotor . . . . . . . . . . . . . . . . . . . . . . . 271
A.1.1.2 Shrouded rotor . . . . . . . . . . . . . . . . . . . . . 273
A.1.1.3 Comparisons of shrouded-rotor and open-rotor per-
formance . . . . . . . . . . . . . . . . . . . . . . . . 277
A.1.1.4 Pressure distributions . . . . . . . . . . . . . . . . . 279
A.1.2 Climb (Axial Flight) . . . . . . . . . . . . . . . . . . . . . . . 286
A.2 Combined Blade-Element-Momentum Theory (Hover) . . . . . . . . . 292
B Measures of Rotor Efficiency 297
B.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
B.2 Figure of Merit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
B.3 CT/CP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302
B.4 Comparison of FM and CT/CP . . . . . . . . . . . . . . . . . . . . . 304
B.5 Shrouded rotors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
Bibliography 312
viii
List of Tables
1.1 Previous experimental work: Shrouded-rotor configurations . . . . . 36
2.1 Matrix of shrouded-rotor models tested in hover . . . . . . . . . . . . 112
2.2 Momentum theory predictions . . . . . . . . . . . . . . . . . . . . . . 113
2.3 Rotor parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.1 Shrouded-rotor model comparison series for analyzing effects of the
shroud geometric parameters on performance characteristics . . . . . 160
3.1 Shrouded-rotor model comparison series (contd.) . . . . . . . . . . . . 161
3.2 Quantitative evaluation of sensitivity of shrouded-rotor performance
to changes in parameter values . . . . . . . . . . . . . . . . . . . . . . 183
3.3 Shrouded-rotor model comparison series for analyzing effects of the
shroud geometric parameters on shroud pressure distributions . . . . 195
4.1 Airspeed ratio (?prime) values tested . . . . . . . . . . . . . . . . . . . . . 221
ix
List of Figures
1.1 MAV missions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 The MAV flight regime. . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2 The MAV flight regime (contd.) . . . . . . . . . . . . . . . . . . . . . 7
1.3 Effect of decreasing Reynolds number on airfoil lift and drag. . . . . . 8
1.4 Maximum figures of merit achieved by MAV-scale rotors. . . . . . . . 10
1.5 Cross-section of a shrouded rotor in hover. . . . . . . . . . . . . . . . 12
1.6 Applications of shrouded rotors: Powered-lift V/STOL aircraft. . . . 13
1.6 Applications of shrouded rotors: STOL aircraft and compound heli-
copters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.6 Applications of shrouded rotors (contd.): Helicopter tail rotors. . . . 14
1.6 Applications of shrouded rotors (contd.): Unmanned VTOL aircraft. . 15
1.7 Shrouded rotor thrust components . . . . . . . . . . . . . . . . . . . 17
1.8 Pressure variations in hover flow-fields of open and shrouded rotors . 19
1.9 Comparison of characteristics of open and shrouded rotors . . . . . . 21
1.10 Shrouded rotor in non-axial flight. . . . . . . . . . . . . . . . . . . . . 24
1.11 An illustration from Hamel?s 1923 patent. . . . . . . . . . . . . . . . 28
1.12 An illustration from Kort?s 1936 patent. . . . . . . . . . . . . . . . . 28
1.13 Stipa?s 1933 venturi-fuselage monoplane design . . . . . . . . . . . . . 29
1.14 Nord 500 ?Cadet?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.15 Ring-Fin tail rotors tested at the Bell Helicopter Company. . . . . . . 31
1.16 Principal shroud parameters affecting shrouded-rotor performance. . . 34
1.17 Stipa?s 1931 venturi-tube wind-tunnel tests. . . . . . . . . . . . . . . 43
1.18 Shroud shapes tested by Kr?uger. . . . . . . . . . . . . . . . . . . . . . 46
1.19 Shroud shapes tested by Platt. . . . . . . . . . . . . . . . . . . . . . . 48
x
1.20 Hubbard?s test arrangement. . . . . . . . . . . . . . . . . . . . . . . . 49
1.21 Shrouded propeller tested by Parlett. . . . . . . . . . . . . . . . . . . 50
1.22 Shrouded-propeller tests conducted by Taylor. . . . . . . . . . . . . . 52
1.23 Shroud models tested by Black et al. . . . . . . . . . . . . . . . . . . 54
1.24 Shroud design of the Sikorsky S-67 Blackhawk?s fan-in-fin. . . . . . . 61
1.25 Ducted Tail Rotor tested by Bell Helicopter Textron. . . . . . . . . . 64
1.26 Shroud configurations tested for the AROD by Weir. . . . . . . . . . 69
1.27 MIT/Draper Perching UAV. . . . . . . . . . . . . . . . . . . . . . . . 74
1.28 Auxiliary control devices tested by Fleming et al. . . . . . . . . . . . 76
1.29 Shrouded rotors tested by Martin and Tung. . . . . . . . . . . . . . . 79
1.30 Shrouded rotors tested by Martin and Boxwell. . . . . . . . . . . . . 83
1.31 Inlet lip shapes tested by Graf et al. . . . . . . . . . . . . . . . . . . . 85
1.32 Shrouded rotor tested by Sirohi et al. . . . . . . . . . . . . . . . . . . 87
2.1 Principal shroud parameters affecting shrouded-rotor performance . . 108
2.2 Shrouded-rotor models . . . . . . . . . . . . . . . . . . . . . . . . . . 109
2.3 Close-up view of shrouded-rotor model LR13-D20 . . . . . . . . . . . 110
2.4 Hover test setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
2.5 Wind tunnel test setup . . . . . . . . . . . . . . . . . . . . . . . . . . 119
2.6 Four-component wind-tunnel sting balance . . . . . . . . . . . . . . . 120
2.7 Sting balance with shrouded-rotor model mounted . . . . . . . . . . . 120
2.8 Forces and moments acting on model in translational flight . . . . . . 121
2.9 Angle of attack definition . . . . . . . . . . . . . . . . . . . . . . . . . 122
2.10 Center of pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
2.11 Pressure tap locations (wind-tunnel model) . . . . . . . . . . . . . . . 124
xi
2.12 Wind tunnel model instrumented with pressure taps . . . . . . . . . . 125
3.1 Effects of changing rotational speed on performance . . . . . . . . . . 134
3.2 Effect of increasing rotational speed on power consumption . . . . . . 135
3.3 Effect of changing diameter on the performance of open rotors . . . . 137
3.4 Thrust coefficient vs. collective angle . . . . . . . . . . . . . . . . . . 139
3.5 Power coefficient vs. collective angle . . . . . . . . . . . . . . . . . . . 140
3.6 Theoretical prediction of CT vs. ?tip for different values of ?d . . . . . 143
3.7 Thrust coefficient vs. power coefficient . . . . . . . . . . . . . . . . . 144
3.8 Ratio of thrust coefficient to power coefficient vs. collective angle . . 148
3.9 Ratio of thrust coefficient to power coefficient vs. thrust coefficient . 149
3.10 Figure of merit vs. collective angle . . . . . . . . . . . . . . . . . . . 150
3.11 Figure of merit vs. thrust coefficient . . . . . . . . . . . . . . . . . . 151
3.12 Generalized figure of merit vs. thrust coefficient . . . . . . . . . . . . 152
3.13 Rotor thrust coefficient, as normalized by shroud thoat area (piD2t/4),
vs. collective angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
3.14 Rotor thrust coefficient, as normalized by rotor disk area (piD2/4),
vs. collective angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
3.15 Shroud thrust coefficient vs. collective angle . . . . . . . . . . . . . . 157
3.16 Principal shroud parameters affecting shrouded-rotor performance . . 159
3.17 Effect of blade tip clearance . . . . . . . . . . . . . . . . . . . . . . . 164
3.17 Effect of blade tip clearance (contd.) . . . . . . . . . . . . . . . . . . 165
3.17 Effect of blade tip clearance (contd.) . . . . . . . . . . . . . . . . . . 166
3.18 Effect of lip radius . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
3.18 Effect of lip radius (contd.) . . . . . . . . . . . . . . . . . . . . . . . . 168
3.18 Effect of lip radius (contd.) . . . . . . . . . . . . . . . . . . . . . . . . 169
xii
3.19 Effect of diffuser angle . . . . . . . . . . . . . . . . . . . . . . . . . . 170
3.19 Effect of diffuser angle (contd.) . . . . . . . . . . . . . . . . . . . . . 171
3.19 Effect of diffuser angle (contd.) . . . . . . . . . . . . . . . . . . . . . 172
3.20 Effect of diffuser length . . . . . . . . . . . . . . . . . . . . . . . . . . 174
3.20 Effect of diffuser length (contd.) . . . . . . . . . . . . . . . . . . . . . 175
3.20 Effect of diffuser length (contd.) . . . . . . . . . . . . . . . . . . . . . 176
3.21 Effect of diffuser expansion ratio . . . . . . . . . . . . . . . . . . . . . 177
3.21 Effect of diffuser expansion ratio (contd.) . . . . . . . . . . . . . . . . 178
3.21 Effect of diffuser expansion ratio (contd.) . . . . . . . . . . . . . . . . 179
3.22 Shroud surface pressure distributions in hover: LR13-D20 models . . 185
3.22 Shroud surface pressure distributions in hover (contd.): LR13-D10
models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
3.22 Shroud surface pressure distributions in hover (contd.): LR06 and
LR09 models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
3.23 CFD prediction of secondary suction peak by Rajagopalan and Keys. 190
3.24 Sectional inlet resultant force vector . . . . . . . . . . . . . . . . . . . 192
3.25 Inlet thrust distributions for model LR13-D10-?0.1 . . . . . . . . . . 193
3.26 Effects of blade tip clearance on pressure distribution . . . . . . . . . 197
3.27 Effects of inlet lip radius on pressure distribution . . . . . . . . . . . 199
3.28 Effects of diffuser included angle on pressure distribution . . . . . . . 201
3.29 Effects of diffuser length on pressure distribution . . . . . . . . . . . . 203
3.30 Effect of lip radius on inlet thrust distributions . . . . . . . . . . . . 205
3.31 Rotor wake development at ?0 = 20? . . . . . . . . . . . . . . . . . . 208
3.32 Comparisonofrotorwakedevelopmentofconventional-scaleandMAV-
scale rotors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
3.33 Wake profiles at diffuser exit plane of shrouded rotors . . . . . . . . . 211
xiii
3.34 Effects of collective angle on induced power . . . . . . . . . . . . . . . 213
3.35 Effect of blade tip clearance on shrouded-rotor exit-plane wake profiles215
3.36 Effects of changing blade tip clearance on induced power . . . . . . . 216
3.37 Effect of diffuser length on shrouded-rotor exit-plane wake profiles . . 217
3.38 Effects of changing diffuser length on induced power . . . . . . . . . . 218
4.1 Translational flight: variable definitions . . . . . . . . . . . . . . . . . 220
4.2 Coordinate system for pressure distribution plots . . . . . . . . . . . 223
4.3 Effects of changing airspeed on pressure distributions, at fixed angle
of attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
4.3 Effects of changing airspeed on pressure distributions, at fixed angle
of attack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
4.4 Effects of changing airspeed on pressure distributions, at fixed angle
of attack: 2-D depictions . . . . . . . . . . . . . . . . . . . . . . . . . 226
4.4 Effects of changing airspeed on pressure distributions, at fixed angle
of attack: 2-D depictions . . . . . . . . . . . . . . . . . . . . . . . . . 227
4.5 Effects of changing angle of attack on pressure distributions, at fixed
airspeed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
4.6 Translational flight: variations in thrust . . . . . . . . . . . . . . . . 234
4.7 Translational flight: variations in normal force . . . . . . . . . . . . . 237
4.8 Translational flight: variations in lift . . . . . . . . . . . . . . . . . . 238
4.9 Translational flight: variations in drag . . . . . . . . . . . . . . . . . 240
4.10 Translational flight: variations in location of center of pressure . . . . 242
4.10 Translational flight: variations in location of center of pressure . . . . 243
4.11 Translational flight: variations in pitch moment . . . . . . . . . . . . 245
4.11 Translational flight: variations in pitch moment . . . . . . . . . . . . 246
4.12 Translational flight: variations in power . . . . . . . . . . . . . . . . . 247
xiv
5.1 Trade studies comparing open- and shrouded-rotor configurations for
MAVs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
A.1 Flow-field models for open and shrouded rotors in hover. . . . . . . . 270
A.2 Pressure variations in hover flow-fields of open and shrouded rotors . 281
A.3 Sphere-cap model for inlet flow geometry of a shrouded rotor. . . . . 282
A.4 Theoretical predictions of shrouded-rotor surface pressure distributions.284
A.5 Shroud surface pressure distributions: comparison of theory with ex-
perimental measurements. . . . . . . . . . . . . . . . . . . . . . . . . 285
A.6 Effect of expansion ratio on pressures at the rotor disk plane; CT =
0.02. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
A.7 Predicted variations in ideal induced velocity and power in climb, for
open and shrouded rotors. . . . . . . . . . . . . . . . . . . . . . . . . 291
A.8 Comparison of open and shrouded rotor power requirements in climb. 291
B.1 Variation of power loading with disk loading. . . . . . . . . . . . . . . 301
B.2 Variation of power loading with disk loading for MAV-scale rotors . . 304
xv
Nomenclature
A shroud throat cross-sectional area = pi(Dt/2)2
AR rotor disk area, corrected for blade root cut-out = pi(R2 ?R20)
Ae diffuser exit area
a airfoil lift-curve slope = dCL/d?
CD drag coefficient = D/?A(?R)2
CL lift coefficient = L/?A(?R)2
?CL mean lift coefficient of rotor = 6CT
rotor/?
CM pitch moment coefficient = CN ?xcp/R
CN normal force coefficient = N/?A(?R)2
CP power coefficient = P/?A(?R)3
Cp pressure coefficient = (p?patm)/qtip
CT thrust coefficient = T/?A(?R)2
c rotor blade chord
cS shroud chord
D drag
DR rotor diameter
Dt shroud throat diameter (minimum inner diameter)
J propeller advance ratio = v?/nDR
L lift
Ld shroud diffuser length
M pitch moment (at rotor hub)
?m mass flow = ?Av
N normal force
Nb number of blades
n rotor rotational speed, rev/s
P rotor shaft power = ?Pi +Po
Pi ideal power
Po profile power
p local static pressure
patm ambient atmospheric pressure
Q rotor torque
qtip dynamic pressure at blade tips = 12?v2tip
R rotor radius = DR/2
R0 rotor blade root cut-out
Re Reynolds number
r non-dimensional radial coordinate = y/R
rlip shroud inlet lip radius
S shroud diffuser slant length = Ld/cos(?d/2)
s non-dimensional distance along shroud surface
T thrust
Ttotal total shrouded-rotor thrust = Trotor + Tshroud
t shroud wall thickness
xvi
v air velocity
vi (ideal) induced velocity at rotor plane
vtip rotor tip speed = ?R
v? free-stream velocity (airspeed)
w induced velocity in far wake of rotor
xcp location of model center of pressure
y radial coordinate
z axial coordinate, positive downstream from rotor
Greek symbols
? model angle of attack; blade element aerodynamic angle of attack
?rotor conventional rotor angle of attack = 90? ??
?tip blade tip clearance = (Dt? DR)/2
? induced inflow angle = ?/r
? induced power correction factor
? induced inflow ratio = vi/(?R)
? rotor advance ratio = v? cos(?rotor)/?R
?prime airspeed ratio = v?/?R = J/pi
? rotor rotational speed, rad/s
? air density
? rotor solidity = Nbc/piR
??d shroud diffuser expansion ratio = Ae/A
?d expansion ratio, corrected for rotor hub blockage = Ae/AR
?0 rotor collective (blade pitch angle)
? blade element pitch angle = ?+?
?d diffuser included angle
?lip shroud inlet lip angular coordinate
? rotor azimuth angle
Subscripts
OR Open (unshrouded) Rotor
SR Shrouded Rotor
c value in the climb condition
diffuser component due to shroud diffuser
h value in the hover condition
inlet component due to shroud inlet
rotor component due to rotor
shroud component due to shroud
Shrouded-rotor model nomenclature
LRx model with lip radius = x% Dt
Dy model with diffuser included angle = y?
?z model with blade tip clearance = z% Dt
Lw model with diffuser length = w% Dt
xvii
Abbreviations
AoA Angle of Attack
BEMT Blade-Element-Momentum Theory
CFD Computational Fluid Dynamics
DAS Data Acquisition System
DL Disk Loading = Trotor/A
FM Figure of Merit = Pi/P
FM? Generalized Figure of Merit = FM??d
ID Inner Diameter
MAV Micro Air Vehicle
OAV Organic Air Vehicle
OD Outer Diameter
PL Power Loading = T/P
rpm revolutions per minute
UAV Unmanned Air Vehicle
V/STOL Vertical/Short Take-Off and Landing
xviii
Chapter 1
Introduction
1.1 Background: Micro Air Vehicles
One of the hallmarks of maturing technology is the increase in complexity and
capability of the devices created by that technology, and in many cases the simulta-
neous decrease in size of those devices. In the multi-disciplinary field of air vehicle
design and development, this characteristic is evident in all aircraft components,
from propulsion and power generation systems to sensors, navigation systems and
flight control systems ? many of these in turn dependent on the incredible degree of
miniaturization achieved in the computer and microchip-fabrication industries. This
miniaturization has led to the development and widespread use of Unmanned Aerial
Vehicles, or ?UAV?s ? aircraft that do not carry a human pilot, and are therefore
often much smaller in size than conventional manned aircraft. The commonly-seen
radio-controlled (?RC?) aircraft flown by hobbyists around the world also fall into
this category; however, these aircraft are meant primarily for entertainment pur-
poses, and seldom have the capability of carrying a significant payload or mission
equipment package (MEP). The term ?UAV? is therefore typically used for aircraft
that are designed for a specific mission, either civil or military, and have actual
utility towards that purpose.
The UAVs that are in use today vary in size from a few feet to several tens of
feet1, where, in the case of fixed-wing aircraft, the wingspan is typically regarded
as the characteristic length, while for rotating-wing aircraft like helicopters, it is
1Ryan Aeronautical / Northrop Grumman RQ-4 Global Hawk: 116 ft wingspan; Boeing A160
Hummingbird: 36 ft rotor diameter
1
the main rotor diameter that is used. The natural process of continuing minia-
turization led people to consider the possibility of mission-capable aircraft that
would be even smaller, with maximum spatial dimensions of a foot or less. In 1992,
a DARPA2/RAND Corporation workshop on ?Future Technology-Driven Revolu-
tions in Military Operations? investigated the concept of ?mobile microrobots? at
the 1-cm/1-g scale [1]. This was followed by a series of feasibility studies at the
MIT Lincoln Laboratory and the U.S. Naval Research Laboratory, and led to the
creation of a DARPA Small Business Innovation Research (SBIR) program in the
fall of 1996 to develop this ?new dimension in flight? [2]. According to this program,
the formal definition of a Micro Air Vehicle, or ?MAV?, was an aircraft that would
have no dimension larger than 15 cm (6 in), weigh approximately 100 g ? which
included a payload weight of 20 g ? and have an endurance of one hour. The
payload would typically be some type of sensor ? optical, chemical or radiological,
for example ? and/or a radio transmitter. The envisioned military use of such an
aircraft was as man-portable, ?eye-in-the-sky? flying robot that could be carried and
operated by an individual soldier, for increased situational awareness while mini-
mizing exposure of him- or herself to risk. Because of their small size and weight,
these aircraft would have a much smaller footprint compared to the larger UAVs,
not just in terms of the aircraft itself, but also with regard to their ground station
and logistical trail, and would therefore be faster and easier to deploy, and would
compete minimally with the other equipment that soldiers are required to carry ?
armament, food, water, protective gear, and the like. The primary driver for MAV
development was therefore as a reconnaissance asset for military and paramilitary
applications (Figs. 1.1a?d), but civilian applications such as power-line inspection,
traffic monitoring and disaster-response management were also forseen for this tech-
2U.S. Defense Advanced Research Projects Agency ? called the Advanced Research
Projects Agency (ARPA) before March 1972 and between February 1993 and February 1996
(http://www.darpa.mil/body/arpa darpa.html).
2
(a) Urban operations (b) Bio-chemical sensing
(c) MAV-assisted pilot rescue (d) ?Over-the-hill? reconnaissance
Figure 1.1: MAV missions [3]
nology.
1.2 The need for more efficient hover-capable MAVs
The early developmental work in MAVs led to the creation of highly-successful
fixed-wing aircraft such as Aerovironment?s Black Widow, which had an endurance
of half an hour and a range of almost 2 km while weighing only 80 g [4]. How-
ever, such high-speed platforms ? Black Widow had a loiter velocity of 25 mph ?
are restricted in operation to large, open spaces with a minimum of obstacles ?
outdoors, and at altitudes higher than the tops of trees and buildings. For oper-
ations in highly-congested, highly-cluttered environments like urbanized areas (the
so-called ?urban canyon?), whether indoors or outdoors, aircraft are required that
are capable of high maneuverability at low speeds, and even of hovering. Aircraft
3
that have these capabilities fall into three classes: rotating-wing configurations, like
helicopters and tilt-rotors, flapping-wing configurations, like birds and insects ?
biological flyers, and fixed-wing configurations with powered-lift capability, such as
the Hawker Siddeley Harrier and the Lockheed Martin Joint Strike Fighter (F-35
Lightning II). For any aircraft, low-speed flight and, if it is possible, hovering flight
are inherently much more power-hungry than most other parts of the flight regime.
Of the three classes of aircraft listed above, rotary-wing aircraft exhibit the high-
est efficiency in hover and low-speed flight. The potential for high efficiencies in
flapping-wing ? ornithoptic ? configurations has been demonstrated for aircraft
of the MAV-size and smaller [5, 6], but the mechanical and aeroelastic complexity
of such mechanisms are hurdles that remain to be overcome.
Given that hovering and low-speed flight are already states of high power con-
sumption, the situation is further exacerbated by the highly degraded performance
of conventional airfoils at the MAV-scale. In fluid dynamics, the sense of scale is
best quantified by a non-dimensional parameter called the Reynolds number (Re),
which is proportional to the product of the size and the velocity of the object that
is moving relative to the fluid. Large aircraft, such as commercial airliners, operate
at Reynolds numbers in the tens of millions, whereas MAVs operate in a Reynolds-
number regime of approximately 10,000 to 50,000 ? three orders of magnitude lower
(Figs. 1.2a?c). The Reynolds number is given by the formula:
Re = ?vl?
where ? is the density of the fluid, ? is the dynamic viscosity of the fluid, v is the
relative velocity between the object and the fluid, and l is a characteristic length
of the object. This fluid-dynamic parameter can be considered as a ratio between
4
the inertial forces and the viscous forces that act on fluid elements in the flow.3 At
high Reynolds numbers the inertial forces dominate, while at low Reynolds numbers
the nature of the flow is more strongly affected by the effects of viscosity. Where
airfoils are concerned, this results in two immediate effects: first, a decreased ability
of the fluid to withstand (or maintain) adverse pressure gradients, and therefore to
separate easily from the surface of the airfoil, thereby reducing the maximum lift
capability and increasing the pressure drag; and second, an increase in the skin-
friction drag when the flow does remain attached to the airfoil (Fig. 1.3). Together,
these effects result in extremely low lift-to-drag (L/D) ratios for airfoils in low-
Reynolds-number flows. Certain airfoil shapes, similar to those found in bird and
insect wings, are optimized for low-Re flight, and these do fare better ? in this flight
regime ? than ?conventional? airfoils that have been designed for larger, manned
aircraft; however, the highest L/D ratios achieved by even these optimized, low-Re
airfoils are still substantially lower than those achieved by the conventional airfoils
at higher Reynolds numbers [8].
The degraded performance of airfoils is an obstacle faced by both fixed- and
rotary-wing MAVs, but it is especially critical for the latter, as they spend a large
portion of their mission-profile in the power-intensive hovering and low-speed con-
ditions. The hovering efficiency of a rotor is typically expressed in two different
ways:
(i) By a non-dimensional parameter called the Figure of Merit (FM), which is
the ratio of the ideal amount of power (Pi) required by the rotor to generate
a certain amount of thrust, to the value of the actual amount of power (P) it
requires:
FM = PiP
3An excellent exposition of the characteristics and significance of the Reynolds number is given
by Vogel [7, Ch. 5].
5
(a) From Ref. [2].
(b) From Ref. [5], from data in [9?11].
Figure 1.2: The MAV flight regime.
6
(c) From Refs. [8, 12], from data in [9].
Figure 1.2: The MAV flight regime (contd.)
7
(a) Symmetric airfoils: Effect of thickness on
maximum lift.
(b) Cambered airfoils: Effect of camber on
maximum lift.
(c) Symmetric airfoils: Effect of thickness on minimum drag.
(d) Cambered airfoils: Effect of camber on minimum drag.
Figure 1.3: Effect of decreasing Reynolds number on the maximum lift coefficient
and minimum drag coefficient of airfoil sections [13].
8
The maximum possible value of the figure of merit is therfore 1.0, which would
imply a ?perfect? rotor.
(ii) By a dimensional parameter called the Power Loading (PL), which is the ratio
of the thrust (T) produced by a rotor to the power (P) it consumes to produce
that thrust:
PL = TP
As will be elucidated upon in Chapter 3 and in Appendix B of this dissertation, both
of these formulations have their respective demerits. For example, a restriction
on the former is that a comparison of the figures of merit of different rotors is
only meaningful when they are operating such that they have the same rotor disk
loading (DL = T/A). The power loading, on the other hand, has the undesirable
attibute of being a dimensional quantity, having units of velocity?1, and is inversely
proportional to the tip speed of rotor. A non-dimensional, tip-speed-independent
form of this efficiency measure can, however, be obtained by taking the ratio of
the coefficient forms of the rotor thrust and power (CT/CP = vtip ?T/P). For
now, it suffices to state that rotors are desired to have the highest possible values
of FM and power loading. Larger rotors that are used on conventional, manned
helicopters have been designed that achieve figures of merit of up to 0.8 and power
loadings of around 10 lb/hp [14, p. 43, p. 47]. Assuming a typical rotor tip speed of
700 ft/s (210 m/s) [15, pp. 683?701], this implies a CT/CP ratio of 12.6. MAV-scale
rotors (Fig. 1.4) are currently capable of FM values of up to 0.65 [16], and while
this is a significant improvment from the maximum values of 0.40?0.45 achieved
by the early micro-rotors [17?19], the power requirements remain high: the data
from investigations by Bohorquez and Pines [16] and Hein and Chopra [20], for
example, indicate maximum values of CT/CP between 5.0 and 6.0, less than half
that achieved at the larger scales. Reductions in power consumption would lead to
9
5 6 7 8 9 100.1
0.2
0.3
0.4
0.5
0.6
0.7
Rotor diameter [in]
Maximum FM
Figure 1.4: Maximum figures of merit achieved by MAV-scale rotors [16, 19?25].
increases in aircraft endurance beyond the 10?15 minutes currently achievable by
rotary-wing MAVs [16], and/or reductions in the aircraft weight-fraction taken up by
the energy-storage components like batteries, thereby allowing for larger payloads.
1.3 The shrouded-rotor configuration: Potential for improved per-
formance
Improvements in airfoil and rotor design constitute one means of improving
the efficiency of hover-capable MAVs. Another approach is to investigate alternative
aircraft configurations that have the potential for performance better than that of
conventional rotorcraft designs. One such configuration is the shrouded rotor, which
involves surrounding the rotor with ? in its most basic form ? a cylindrical shroud
or duct.4 More commonly, the shroud resembles an annular airfoil or ?ring-wing?,
with camber and finite thickness that vary along its length, and with a rounded
?leading edge? and a smoothly tapered ?trailing? edge which form, respectively, the
inlet and exit or diffuser sections of the shroud (Fig. 1.5). The shrouded-rotor config-
uration has been extensively investigated for over half a century and found to result
4An arbitrary convention that is sometimes adopted is that the enclosing structure is called
a ?duct? if it is greater in length than the rotor diameter, and a ?shroud? or a ?short-chord duct?
otherwise.
10
in significant gains in aerodynamic performance (increased thrust, reduced power
consumption) compared to the unshrouded or ?open? rotor, and has therefore been
utilized in various forms, from ducted propellers and ducted fans on powered-lift
V/STOL aircraft to ?Fenestron? tail-rotors5 on helicopters and ducted-rotor UAVs
(Figs. 1.6a?p). For a shroud with a profile that is cambered inwards, like that shown
in Fig. 1.5, and which therefore accelerates the flow towards the rotor, the perfor-
mance gain is principally in static (hover) and near-static conditions. However, a
benefit is obtained in high-speed forward flight as well: with the rotor pitched for-
ward to provide the required thrust, the shroud, behaving as an annular wing, can
provide the required lift ? with potentially less than half the induced drag of a
planar wing of the same aspect ratio [26, 27] ? and, because of the induced flow of
the rotor, can safely operate at angles of attack far greater than the stall angle of
the ?unpowered? annular wing [27].
Finally, in addition to these aerodynamic benefits, shrouding the rotor conveys
two other advantages over an open rotor: (1) the shroud can potentially attenuate
the noise signature of the rotor, which is an advantage for MAV missions that require
covert operation, and, (2) the shroud serves as a safety feature, protecting both the
rotating blades from damage by other objects as well as personnel from injury by
the blades. In fact, in most cases, it was this latter reason ? improved safety ?
that the shrouded-rotor configuration was chosen for an aircraft design [28?33].
The performance benefit of shrouding the rotor derives principally from the
ability of the diffuser section of the shroud to restrain the natural contraction of
the flow after it passes through the rotor. For any fluid-dynamic propulsive device,
the increase in velocity of the far wake over that of the free-stream fluid represents
an ?induced? power expenditure that is unavoidable in the generation of thrust.
Additional losses do arise due to the viscosity of the fluid, but even in the ideal case
5Also known as a ?Fan-in-fin?, ?Fantail? or ?Ducted Tail Rotor?, depending on the helicopter
manufacturer.
11
Figure 1.5: Cross-section of a shrouded rotor in hover.
of an inviscid fluid, this induced increase in the kinetic energy of the fluid constitutes
the minumum possible power required to generate a given amount of thrust. For
a conventional open rotor, by combining an actuator-disk model of the rotor with
the conservation laws of fluid dynamics, the classical Momentum Theory is able
to predict the increase in velocity at the far wake as being twice the value of the
induced velocity at the plane of the rotor disk [14, Ch. 2; See also Appendix A of
this dissertation]. By mass conservation, the contraction ratio of the slipstream is
inversely proportional to the increase in velocity, and, in the hover condition, has a
value of 1/2. In other words, the final cross-sectional area of the slipstream in the
far wake is one-half that of the rotor disk area. In the case of a shrouded rotor,
the change in area of the slipstream is determined by the shape of the diffuser. By
using a cylindrical diffuser, or a diverging one, the slipstream can be forced to either
maintain a constant cross-sectional area or to actually increase in area, respectively.
Thus, if the flow is able to stay attached to the diffuser walls, the presence of the
12
(a) Hiller VZ-1 Pawnee. (b) Bell X-22A.
(c) Piasecki VZ-8 AirGeep. (d) Doak VZ-4.
(e) GE/Ryan XV-5 Vertifan. (f) Vanguard Omniplane.
Figure 1.6: Applications of shrouded rotors: Powered-lift V/STOL aircraft.
13
(g) Mississippi State University XAZ-1
Marvelette.
(h) Mississippi State University XV-11A
Marvel.
(i) Piasecki 16H-1 Pathfinder. (j) Piasecki 16H-1A Pathfinder II.
Figure 1.6: Applications of shrouded rotors: STOL aircraft and compound
helicopters.
(k) Boeing/Sikorsky RAH-66 Comanche. (l) A?erospatiale SA365 (HH-65) Dauphin.
Figure 1.6: Applications of shrouded rotors (contd.): Helicopter tail rotors.
14
(m) Sandia AROD. (n) Sikorsky Cypher.
(o) Honeywell Kestrel. (p) Microcraft/Allied Aerospace iSTAR.
Figure 1.6: Applications of shrouded rotors (contd.): Unmanned VTOL aircraft.
15
shroud is able to reduce the increase in velocity of the far wake, and thereby reduce
the induced-power requirements of the rotor.
Just as for an open rotor, momentum theory can be used for a first-order
prediction of the performance and characteristics of a shrouded rotor ? with one
difference: instead of using the actuator-disk model of the rotor, the assumption is
made that the slipstream has fully expanded back to ambient atmospheric pressure
at the exit plane of the diffuser. Therefore, a key parameter in determining the
performance of a shrouded rotor is the diffuser expansion ratio (?d), which is equal
to the ratio of the diffuser exit plane area (Ae) to the area of the rotor disk (A):
?d = AeA
Appendix A (p. 269) contains detailed derivations of the predictions of momentum
theory for a shrouded rotor, but some of the more important results are repeated
here below. In the hover case, for a given total system thrust (Ttotal = Trotor +
Tshroud), the ideal induced velocity at the rotor disk (vi), the ideal induced power
(Pi), the figure of merit (FM)6 and the fraction of the total thrust produced by the
diffuser of the shroud are given by:
vi =
radicalBigg
?d Ttotal
?A (1.1)
Pi = T
3/2
total?
4?d?A (1.2)
FM = C
3/2
Ttotal
2??dCP (1.3)
Tdiffuser
Ttotal = ?
(?d ?1)2
2?d (1.4)
The actuator-disk model of the rotor is used only to derive the fractions of the total
6In the shrouded-propeller/ducted-fan literature, this is often instead called the Static Efficiency
(?s).
16
Figure 1.7: Shrouded rotor thrust components
thrust that are borne by the rotor disk itself and by the shroud inlet:
Trotor
Ttotal =
w
2vi =
A
2Ae =
1
2?d (1.5)
Tinlet
Ttotal =
?d
2 (1.6)
Equations (1.4)?(1.6) are plotted as functions of the expansion ratio in Fig. 1.7.
Note that these thrust fractions are functions of the expansion ratio only, and are
independent of the precise shape of the shroud, and that as?d is increased, the rotor?s
contribution (Trotor) to the total thrust decreases, while the shroud?s contribution
(Tshroud = Tinlet + Tdiffuser) increases.
In the case of an open rotor, Trotor = Ttotal, so, from Eq. (1.5), ?d = 1/2,
which is exactly the value of the ?expansion? that momentum theory predicts for an
open rotor. Equations (1.1)?(1.3) therefore apply equally to both shrouded and un-
shrouded (open) rotors, with the effective value of 1/2 being used for the expansion
ratio in the case of an open rotor.7 In the case of a shrouded rotor with a diffuser
7Note, however, that applying the special-case formula for the figure of merit of an open rotor
17
expansion ratio greater than 1/2, the total thrust is clearly greater than the thrust
acting on the rotor alone. The physical basis for this additional thrust is a region of
low-pressure suction forces acting on the inlet of the shroud, caused by the turning
of the air that the rotor ingests from around the sides of the shroud. This is directly
analagous to the leading-edge suction on the nose of an airfoil. Similarly, just as the
flow around an airfoil experiences a pressure-recovery region on the trailing surfaces
of the airfoil, so too does the flow in the diffuser experience an adverse pressure gra-
dient as it expands back to ambient atmospheric pressure. The inner surface of the
diffuser therefore also experiences suction pressures and generates a negative thrust,
as can be seen from Eq. (1.4). However, the positive thrust from the inlet is always
large enough that the net shroud thrust (Tshroud) remains positive (for ?d > 1/2). A
clear understanding of the preceding discussion can be obtained from Fig. 1.8, which
illustrates qualitatively the variations in pressure through the flow-fields of an open
and a shrouded rotor in hover. The equations used to generate these figures are
also derived using the momentum theory model, and are given in Section A.1.1.4 of
Appendix A (p. 279).
The momentum theory model also allows for comparison of the characteris-
tics and performances of open and shrouded rotors, as functions of the expansion
ratio of the shrouded rotor. For example, it can be shown that if an open and a
shrouded rotor are compared such that both their rotor disk areas and the amount
of (ideal) power they each consume are the same, then, with increasing expansion
ratio, the total thrust produced by the shrouded rotor continuously increases, rela-
tive to the thrust produced by the open rotor. Simultaneously, the mass flow ( ?m)
and induced velocity (vi) through the shrouded rotor also increase, again relative
to the corresponding values for the open rotor, while the final wake velocity (w)
? C3/2T /?2CP ? to the case of a shrouded rotor would be incorrect, and would result in a value
of FM that would be greater than the true value by a factor of ?2?d.
18
(a) Open rotor flow-field (b) Shrouded rotor flow-field
Far upstream Far wakeNormalized distance
Static pressure
Shrouded Rotor, ?d = 1.5Shrouded Rotor, ?
d = 1.0Open Rotor (?d = 0.5)
Rotor disk plane 
Ambient pressure 
Diffuser exit plane(Shrouded rotor)   
Outer edge of inlet lip (Shrouded rotor)
(c) Pressure variations
Figure 1.8: Variations in pressure in the hover flow-fields of open and shrouded
rotors, at the same thrust coefficient
19
and the thrust provided by the rotor itself (Trotor) decrease. These variations are
described by Eqs. (A.27) in Appendix A, and have been plotted in Fig. 1.9a. In a
similar manner, Fig. 1.9b illustrates the variations in these quantities when the open
and shrouded rotors are compared such that their rotor disk areas and total thrust
produced are held to be the same (Eqs. (A.28)). In this case, as in the previous one,
the rotor thrust and wake velocity decrease with increasing expansion ratio, while
the mass flow and induced velocity at the rotor disk increase, but it is also seen that
the ideal power requirement also continuously decreases8. Thus, for example, with
an expansion ratio of 1.0 ? corresponding to a straight-sided, cylindrical diffuser
? a shrouded rotor can theoretically produce 26% more thrust than an open rotor
of the same size while consuming the same amount of power, or consume 29% less
power while producing the same amount of thrust. Increasing the expansion ratio
further causes the performance improvements to further increase, without limit ?
in theory.
In reality, upper limits on the amount of performance improvement are im-
posed by viscosity-dependent phenomena such as surface frictional losses and flow
separation, which the ?simple? momentum theory is unable to account for. As the
diffuser expansion ratio is increased, performance does improve, until at some point
the flow is unable to counter the adverse pressure gradient in the diffuser and sep-
arates from the diffuser wall. Similarly, reducing the radius of the shroud inlet lip
increases the turning that the air flowing into the shroud must accomplish. With
increased turning, the suction pressures on the inlet increase9; but at some point, at
a low enough lip radius, the flow will simply separate from the inlet surface before
8As will be shown in Chapter 3 and Appendix B of this dissertation, a consequence of this change
in ideal power requirement with changing ?d is that the figure of merit is not an appropriate way to
compare the performances of an open rotor and a shrouded rotor, or, in general, shrouded rotors
with different diffuser expansion ratios.
9Note, however, that with decreasing lip radius, the total surface area of the inlet on which
those suction forces act also decreases, so it cannot be said a priori whether the resultant inlet
thrust will increase, decrease, or remain the same.
20
(a) Comparison at same rotor disk area and same ideal power
(b) Comparison at same rotor disk area and same total thrust
Figure 1.9: Comparison of characteristics of open and shrouded rotors
21
even reaching the rotor. This separated flow, when ingested by the rotor, causes
further losses in thrust as well as increases in noise and vibration. Both of these
separation phenomena are governed by viscous effects, and as such are strongly
affected by the Reynolds number of the flow. Flow separation can be delayed by
using a large inlet lip radius and by increasing the length of the diffuser, so that for a
given expansion ratio, the expansion angle and the adverse pressure gradient in the
diffuser are reduced; however, both these methods have the accompanying penalties
of increasing the size and weight of the shroud, and of increasing the surface area
and corresponding skin-friction losses. Careful design of the shroud is therefore of
utmost importance: maximum possible performance improvement is desired while
minimizing the increase in size and weight of the aircraft due to the shroud itself.
For the shrouded-rotor configuration to be a viable option, the increase in thrust
over that of an open rotor must be greater than the weight of the shroud, and/or
a decrease in power required must be realized while producing a thrust increment
equal to the weight of the shroud.
Additionally, theimprovementsordeteriorationinperformanceoftheshrouded
rotor configuration over the baseline open rotor must be examined over the entire
flight envelope and mission profile of the aircraft. In addition to the requirement
for efficient hovering flight, it is important for a MAV to remain stable in gust-
ing cross-winds, to be able to transition quickly to translational flight, and to have
good forward-flight performance characteristics. In non-hovering axial-flow condi-
tions like climbing flight, the natural contraction of an open rotor is less than in
hover. Therefore, the benefits of having a diffuser that restricts this contraction be-
come less significant, while conversely, the drag of the shroud becomes an important
factor. In non-axial flow conditions (Fig. 1.10), besides the bluff-body drag of the
shroud itself, the turning of the oncoming flow from the freestream direction to the
axial direction results in the production of an additional momentum or ?ram? drag.
22
Although this is partly alleviated by the subsequent turning of the rotor wake from
the axial direction back to the freestream direction, the ram drag is still ?usually
the largest [component of the total aircraft] drag, and ... can easily be 95% of the
total drag at low velocities (< 10 knots)? [34].
The other adverse consequence of this turning of the flow from the freestream
direction to the axial direction is that the turning is much sharper at the windward
(forward) side of the shroud than at the leeward (aft) side. Equivalently, the wind-
ward side of the shroud effectively operates at a higher angle of attack to the flow
than the leeward side, and therefore experiences greater suction on the inlet and
produces more lift. For a regular, axisymmetric shroud design, this fore/aft aerody-
namic dissymmetry leads to a strong nose-up pitch moment, much greater than that
experienced by an open rotor10. This moment works in opposition to any command
to pitch the vehicle nose-down into the wind and tilt the thrust vector forward,
whether to maintain position while hovering in a gusty environment or to attain a
forward translational flight. In a gust, for example, the resulting nose-up moment
would actually tend to tilt the thrust vector towards the downwind direction, exactly
the opposite as desired.11 Reference [34] states that ?robust, precise control in tur-
bulent conditions ... including the complex, turbulent flow fields around buildings
and trees ... poses one of the key technical hurdles for the successful operation of
these vehicles. Current analysis shows that even in the best cases, typical position
control of ducted-fan VTOL UAVs hovering in light winds (< 10 knots) is limited
to about 10 feet of error.? Further complicating the situation is the possibility of
10The nose-up pitch moment on an (untrimmed) open rotor is due to different reasons, depending
on the type of rotor. For an articulated rotor, it is due to a lateral dissymmetry on the rotor
? greater lift on the advancing blade than on the retreating one ? and the 90? phase-lag of
a harmonic system forced at its fundamental frequency, while for a rigid rotor it is due to a
longitudinal variation in the rotor inflow ? greater inflow and therefore lower angles of attack for
the blades at the rear of the rotor.
11For a transient gust, the resulting motion in what was originally the downwind direction after
the gust had died out could lead to a reversal of the direction of the pitch moment on the aircraft,
and an ensuing oscillatory motion that might eventually damp out. The possibility ? and details
? of such motion can be ascertained from a dynamic stability analysis of the aircraft.
23
(a) Flow pattern.
(b) Forces and moments.
Figure 1.10: Shrouded rotor in non-axial flight.
24
flow separation on the windward side of the inlet, which would lead to a sudden,
sharp decrease in pitching moment (?moment stall?), perhaps to the point where the
pitching moment reverses sign [35].
Besides the weight considerations of a shroud, it is these detrimental charac-
teristics of a shrouded rotor that are of greatest concern to the aircraft designer.
Clearly, for a shroud of fixed geometry, the design of the aircraft will be optimized
for maximum performance improvement over the open rotor at a single point in the
flight regime, which may be in hover or at some forward/translational speed. At
all other points, the performance of the shrouded rotor will be sub-optimal, and
possibly even worse than that of the open rotor. Variable-geometry inlets, diffusers
and auxiliary control devices such as aft-mounted vanes may be considered, but
these introduce their own associated trade-offs in weight and system complexity.
Compromises such as these are therefore unavoidable in the design of the shroud.
Another example of a compromise is the choice of the inlet lip radius: as men-
tioned earlier, a sharp inlet lip is disadvantageous in hover and in near-edgewise
flight because it promotes early flow separation; on the other hand, such an inlet
is beneficial in axial flight because it creates less drag [36]. Effects such as these,
which depend on the detailed shape of the shroud, the design of the rotor and the
Reynolds number of the flow, cannot be analyzed by the momentum theory model.
More sophisticated theoretical models as well as experimental and computational-
fluid-dynamic (CFD) investigations are therefore essential for the successful design
of shrouded-rotor aircraft.
25
1.4 Previous research in ducted propellers and shrouded rotors
1.4.1 Historical overview
The idea of surrounding a rotor with an enclosing structure is not a recent one,
although the reasons for doing so have been varied, and the question of who origi-
nally invented the concept remains open to debate. A US patent issued to Georges
Hamel, a French citizen, in 1923 [37] describes a fixed-wing aircraft with propellers
embedded in the wings, with their axes in the vertical direction, perpendicular to
the wing chord: the so-called ?fan-in-wing? configuration (Fig. 1.11). The objective
of the design was to ?combine the principle of the aeroplane with that of the he-
licopter,? and thus obtain an aircraft with both good performance in high-speed
forward flight as well as improved safety in low-speed and vertical flight near the
ground. The text of the patent makes no mention, however, of any enhancements
in performance of either the wings or the embedded propellers due to their mutual
interaction, so it is likely that the author was unaware of the existence of such a phe-
nomenon. By 1933, however, as indicated by the application for another US patent
filed by Ludwig Kort in Germany [38] (the patent was awarded in 1936), people had
become aware of the potential for improvements in the the propulsive efficiency of
ship propellers by surrounding them with ?nozzle-shaped appendages attached to
[the] ship?s hull ... rings, straight or conical tubes or the like? (Fig. 1.12). However,
?all of these combinations ... failed in practical use, as none of them were uniting
the proper shape of the nozzle with the proper relation between the propeller, its
revolutions, the areas at the narrowest cross section and at the mouth of the nozzle
and the form, speed and resistance of the ship? [38]. At around the same time, in
Italy, Luigi Stipa was investigating the effects of integrating an air propeller with
a hollow airplane fuselage shaped like a venturi tube on the inside (Fig.1.13), and
found similar increases in thrust and decreases in power consumption compared
26
to the open propeller [39, 40]. Credit is usually therefore given to Kort and to
Stipa for the first serious, scientific, experimental work on optimizing the shape
of a surrounding duct or shroud for improved thrust characteristics of marine and
aeronautical propellers, respectively.12 This was followed shortly afterwards, in the
1940s, by further systematic testing of different shroud profile shapes by Kr?uger at
the Aerodynamics Research Institute (Aerodynamische Versuchsanstalt, or ?AVA?)
in G?ottingen [43] and by the theoretical analyses of K?uchemann and Weber, also in
Germany [44?51], and then by a proliferation of studies around the world, including
the United States. The objective of much of this early work was to improve the
efficiency of regular airplane propellers ? which were usually designed for optimal
performance in high-speed cruising flight ? at the low-speed and take-off flight con-
ditions, by using the shroud to increase the inflow through the propeller [43, 52].
The increased inflow in static conditions also meant that that the changes in inflow
with changing cruise speed would be smaller, resulting in more uniformly efficient
performance of a fixed-pitch propeller over the entire aircraft mission profile, and less
of a need for a mechanically-complex variable-pitch propeller. Fairchild et al. [53]
provide a good historical overview of these research programs, and Sacks and Bur-
nell [54] compiled an exhaustive survey of the experimental work and state-of-the
art in ducted-propeller aerodynamics as of 1962. In the United States, in particular,
the goal of developing viable V/STOL (Vertical/Short Take-Off and Landing) air-
craft resulted in a considerable amount of experimental work that led (Figs. 1.6a?j,
p. 13) to the design of ?flying platforms? such as the Hiller VZ-1 Pawnee and the
Piasecki VZ-8 AirGeep [55], fan-in-wing aircraft such as the GE/Ryan XV-5 and
the Vanguard Omniplane [56?58], tilt-duct aircraft like the Doak VZ-4 [59?68] and
the Bell Aerosystems X-22A [69?76], and aircraft with non-tilting shrouded pro-
12Even today, shrouded marine propellers are frequently referred to as ?Kort nozzles?, a widely-
used series of which were systematically designed in the 1950s and ?60s by the Maritime Research
Institute Netherlands (MARIN) [41], just as the NACA did for airfoil shapes. A review of the
state of the art in marine ducted propellers as of 1966 was compiled by Thurston and Amsler [42].
27
Figure 1.11: An illustration from Hamel?s 1923 patent [37].
Figure 1.12: An illustration from Kort?s 1936 patent [38].
pellers meant solely for forward propulsion, like the Mississippi State University?s
XAZ-1 Marvelette and XV-11A Marvel STOL aircraft [77] and the Piasecki 16H-1
Pathfinder and 16H-1A Pathfinder II compound helicopters. Similar work was also
being carried out in Europe, such as that by the French company, Nord Aviation,
in developing the tilt-duct Nord 500 Cadet aircraft [78] (Fig. 1.14).
From the 1970s onwards, the emphasis in research shifted from V/STOL air-
craft to a different type of fan-in-wing application: the shrouded tail-rotor or ?fan-
in-fin? on helicopters (Figs. 1.6k?l, p. 14). Originally developed by A?erospatiale in
France for the SA.341 Gazelle helicopter, and termed the ?fenestron? [28, 79], the
concept was then investigated and further developed by other helicopter manufac-
turers as well. Research at the Bell Helicopter Company [53] led to development
of a ?Ducted Tail Rotor? on a modified Model 222 helicopter [80, 81]. Earlier tests
28
Figure 1.13: Stipa?s 1933 venturi-fuselage monoplane design [40].
Figure 1.14: Nord 500 ?Cadet?.
29
of a thin (less than 2 inches thick), protective ring called a ?Ring-Fin? around the
tail rotor of a modified Model 206 JetRanger (Fig. 1.15) revealed thrust augmen-
tation effects from even this rudimentary structure [82, 83]. The Sikorsky Division
of the United Technologies Corporation tested a fan-in-fin on its S-67 Blackhawk
prototype, which had originally been designed with a conventional tail rotor [84].
Based on that experience, the Boeing-Sikorsky team incorporated a ?FANTAILTM?
tail rotor in its winning proposal for the US Army?s Light Helicopter Experimental
(LHX) program: the RAH-66 Comanche [31, 85?91]. In Japan, Kawasaki Heavy
Industry, Ltd., used a fan-in-fin system for the XOH-1 observation helicopter [32],
as did, in Russia, the Kamov Company in the design of the Ka-60 ?Kasatka? (?Killer
Whale?) helicopter [92]. Meanwhile, A?erospatiale, which later (in 1992) merged with
Messerschmitt-B?olkow-Blohm (MBB) of Germany to form Eurocopter, continued to
develop the fenestron [29, 30, 33, 93], and incorporated it in several of its helicopter
models: the SA/AS365 and EC155 Dauphin 2, the AS565 Panther, EC120 Colibri,
EC130, and EC135/635.
As interest grew in unmanned VTOL aircraft, which would operate in clut-
tered environments and in close proximity to humans, the shrouded-rotor configu-
ration was an obvious choice, given the safety benefits of the shroud [18]. Indeed, as
Fleming et al. [94] succinctly state: ?For survivability, operational safety, packaging
and transport considerations, the ducted fan configuration has become the preferred
platform for small VTOL UAV development.? For cancelling the torque reaction
of the rotor, these vehicles either used stator vanes in the downwash of a single
rotor or a pair of coaxial, counter-rotating rotors. Examples of such aircraft (see
Figs. 1.6m?p on page 15) include the ?AROD?, or Airborne Remotely Operated De-
vice, developed in the 1980s by Moller and Sandia National Laboratories for the US
Marine Corps? Exploratory Development Surveillance and Ground Air Telerobotic
System (GATERS) programs [35, 95, 96], the ?Cypher?, developed by the Sikorsky
30
(a) Ring-Fin shapes being tested on a Bell 206.
(b) Ring-Fin shapes tested.
(c) Ring-Fin wake structure.
Figure 1.15: Ring-Fin tail rotors tested at the Bell Helicopter Company [82, 83].
31
Aircraft Corporation in the 1990s for the US military?s Air-Mobile Ground Security
and Surveillance System (AMGSSS) program [96?110], and, most recently, the Hon-
eywell ?Kestrel? and Micro Craft13 ?Lift-Augmented Ducted Fan? (?LADF?)14, both
developed in the last decade in response to the DARPA MAV/OAV (Organic Air
Vehicle) programs [27, 112, 113].
1.4.2 Experimental work: Effects of variations in shroud design
As can be seen from Fig. 1.5 (p. 12), the cross-sectional profile shape of a
shroud closely resembles that of an annular wing, albeit with camber inwards rather
than outwards. Design and optimization of the shroud for a shrouded-rotor appli-
cation is therefore very similar to the design of airfoil profiles: parameters involved
include chordwise thickness and camber distributions, leading-edge (inlet lip) radius,
angle of attack to the freestream, and, in this special case of an annular shape, the
ratio of the diameter to the chord (the ?aspect ratio?, Dt/cS). However, a number
of additional parameters specific to the case of a rotor?shroud assembly need to
be considered as well: these are the number and configuration of blade stages, i.e.,
rotors and/or non-rotating stators or guide vanes, the chord-wise locations of the
blade stages within the shroud, and the size of the gap between the blade tips and
the shroud wall, to say nothing of the myriad of variables involved in the detailed
design of the rotor stages and blades themselves. The design space is vast, and
the numerous studies conducted on shrouded rotors invariably focused on different
subsets of these various parameters. Some of these studies are described below, but
before doing so, it is worthwhile to note that, excluding the last aspect of the design
mentioned above, i.e., the configuration and detailed design of the blade stages, it
13A division of Allied Aerospace Industry Incorporated.
14Also known as the ?iSTAR? vehicle ? an acronym for ?Intelligence, Surveillance, Target Aqui-
sition, Reconnaissance?, and seemingly derived from a Hughes Missile Systems Company design,
as suggested by a patent awarded to Ebbert et al. in 1994 [111].
32
was found that the performance and characteristics of shrouded rotors were most
strongly dictated by a small subset of parameters: the diffuser expansion ratio and
expansion angle, the inlet lip radius, and the blade tip clearance (Fig. 1.16). The im-
portance of carefully selecting the value of the inlet lip radius and diffuser expansion
angle was discussed earlier in Section 1.3, namely, the trade-offs between improving
performance, avoiding flow separation and minimizing vehicle weight. The third
variable ? the blade tip clearance ? plays a critical role because of the fact that
the close proximity of the shroud wall impedes the formation of the strong blade tip
vortices which otherwise reduce the thrust produced on an open rotor. Although
some leakage flow from the high-pressure region below the rotor disk to the low-
pressure region above it is unavoidable, minimizing the tip gap was always found to
improve performance. In fact, in several studies [57, 114], the losses due to the tip
vortices on an open rotor caused the performance improvements from shrouding the
rotor to be greater than those predicted by momentum theory, which assumes ideal
conditions for both the open and shrouded rotors.15A large tip gap allows stronger
tip vortices to form and not only reduces the rotor thrust, but also reduces the
shroud thrust due to suction on the inlet by reducing the inflow velocities at the
shroud wall. In addition, as stated by Ahn [115], ?the less energized flow in [the]
blade-tip region generates [a] slower-flow region [that] ... reduces the effectiveness
of the [diffuser], by surrounding the [more-] energized flow and preventing its ex-
pansion.? The minimum size of the tip gap is limited by the danger of the blades
striking the shroud wall, either during ?normal? operation or because of excessive
15Another possible reason for this phenomenon is the off-loading of the rotor by the shroud:
At the same total thrust, the greater induced velocity at the rotor causes the blades to operate
at a lower angle of attack, compared to the open rotor, and therefore to not only produce lower
thrust (Trotor) than the open rotor (Fig. 1.9b), but also consume less profile power. If the factor
of reduction in the profile power is greater than that for the reduction in the ideal induced power,
as predicted by momentum theory, then the net improvement in performance will be better than
that predicted. This, of course, requires consideration of the additional power input required
to offset the losses due to friction from the shroud wall, and requires a more detailed analytical
model. Whichever the reason, several investigations have found that shrouded rotors come closer
to achieving their ideal predicted performance than do their unshrouded counterparts [52, 114].
33
Figure 1.16: Principal shroud parameters affecting shrouded-rotor performance:
diffuser included angle (?d), diffuser length (Ld), inlet lip radius (rlip) and blade tip
clearance (?tip).
vibrations or blade flapping amplitudes. In many applications, though, such as in
some gas turbine engines and also in the Comanche FANTAILTM design [31], the
gap is closed off near-completely by lining the shroud wall with a brush or some sort
of expendable, easily-abradable material at the blade-passage region.
The details of the shrouded-rotor configurations tested in the studies described
in the following pages are tabulated for easy comparison in Table 1.1. Parameters
for which values were not available are marked with a dash (? ? ?). Where possible,
values for some parameters have been approximated from the shroud coordinates
and other information given in the literature, and have been noted as such. For
example, where the precise value of the inlet lip radius was not available, this was
approximated, as an upper bound, as half of the maximum thickness of the shroud
wall. For some shroud models, such as those used on the VZ-4 and the X-22A [68, 72,
116], and those tested by Martin and Tung [117], values of both the true leading-edge
radius and the wall thickness were given. Assuming these cases to be typical, the lip
radii of the other models can be expected to be about one-half of this upper-bound
value. For the diffuser, two values for the expansion ratio are listed: ??d, which is
34
the geometric value of the shroud alone (Ae/A), and ?d, which is the effective value
after correction for the blockage by the rotor hub (Ae/AR):
?d = ??d ? AA
R
= ?
?
d
1?parenleftbigR0Rparenrightbig2
In cases where the diffuser internal surface was very nearly conical, i.e., with straight
sides, and the value of one of the three diffuser variables ? expansion ratio, expan-
sion angle or length ? was not given, it has been calculated using the relation:
??d =
bracketleftbigg
1 + 2LdD
t
tan
parenleftbigg?
d
2
parenrightbiggbracketrightbigg2
Where the diffuser length was not explicitly given, and the exact chordwise position
of the rotor within the shroud also indeterminable, the total length of the shroud
(cS) has been listed, and has been noted as such.
In addition to the shroud parameters, the table also lists some of the rotor
parameters of the test items: the blade root-cutout (R0), the blade tip speed (vtip),
the number of blades (Nb), the amount of blade twist (?tw), as measured from the
blade tip to the blade root, i.e., the attachment point to the hub, and the rotor
collective angles (?75), measured at the 75% radial station, for which test data were
presented in the literature. Finally, unless otherwise specified, the ?diameter? of a
shroud refers to its throat or minimum inner diameter, Dt.
35
Tab
le
1.1:
Previou
se
xp
erimen
tal
work:
Sh
rou
ded-rotor
confi
gur
ation
s
Test
Pro-
gram
D
t
[in
]
?d
[?]
?? d
?d
R0
Ld [%
D
t]
rlip[%
D
t]
?tip[%
D
t]
vtip[ft/s
]
N
b
?tw [?]
?75[?]
Other
notes
St
ipa,
1931
[39]
20.5(in
let)
?
1.0
?
?
300
?
?
160?440
2
?
17
?,
14
?
3
shrou
ds,
2
pr
op
ell
ers
Be
ll,
1941
[118]
21
?
?
?
0.69R
?
?
?
330
24(c?v
anes
:
37)
6 (c?v
anes
:
9)
5?40(c?v
anes
:
40?70)
Upstream
con
-
tra
vanes
Kr
uger,
1944
[43]
9.5(24cm
)
?
0.88,1.14(0.66?1.49)
1.0,1.3(0.75?1.70)
0.35R
34,
22
?
?
500
8
45,37.5
25?65
15
shr
oud
shap
es
Pl
att,
1948
[52]
48
7, 14.4,22.4
1.1,1.3
1.24,1.46
0.33R
33.6,50.2
4.1(1 2
tm
ax
)0.13(1/16?)
420?630
Upp
er:
5, Lo
wer:
7
33
OR:15?50,SR:35?45
Coaxial
Hu
bbar
d,
1950
[119]
48
?
?1.1,1.19,1.25
1.21,1.31,1.37
0.3R
12,24,
36
1.07,1.43,1.49
0.2(3/32?)?4.4
690
2
27
21.5
Sh
rou
d
sh
ap
es
=
NA
CA
4312,
4315,
4318
with
enlar
ge
d
nose
rad
ii
Par
let
t,
1955
[55]
18
0
1.0
?
?
34
1.4?8.3(fw
d.
fligh
t:
1.4onl
y)
0.33(0.06?)
470?825(fw
d.
fligh
t:
470on
ly)
2
?
8
Sh
rou
d
shap
e
=
cylin
dr
ical
shell
36
Tab
le
1.1:
(con
td
.)
Test
Pro-
gram
D
t
[in
]
?d
[?]
?? d
?d
R0
Ld [%
D
t]
rlip[%
D
t]
?tip[%
D
t]
vtip[ft/s
]
N
b
?tw [?]
?75[?]
Other
notes
Ta
ylor
,1958
16
0,
7,
14
1.0?1.5
1.07?1.60
0.25R
3,
28,
53,78,103
0, 1.56,3.13,6.25,12.5
0.4(1/16?)
?
3
18
12?37
Fan-in-win
g
config
Doak
VZ-4
[68,
72,
116]
48
22
1.28
1.33
0.33R
34.4
2.5(2.45forNA
CA
xx18)(1 2
tm
ax
= 5.43)
0.06(0.03?)
380?1000
8 (in
let
gui
de
vanes
:
7,
sta-
tors:9)
33
(in
let
vanes
:
0,
sta-
tors:15)
18?50
Thic
kn
ess
dis
tribu
tion
app
rox.
that
of
a
NA
CA
0018
airf
oil.
Be
ll
X-22A
[72,
75,
116]
84
?17(calc
.)
1.21
1.27(1.17,acc
.
forblo
ck
-
age
by
exitvane)
0.18R
?33
2.36(2.07forNA
CA
xx18)(1 2
tm
ax
= 4.98)
0.45(3/8?)(0.3?2.0tes
ted)
?
3(sta-tors:6)
41
14?49
Thic
kn
ess
dis
tribu
tion
app
rox.
that
of
a
NA
CA
0018
airf
oil.
Blac
k
et
al.,
1968
[114]
30
7 (b
ase),
14,20,9.3(short
-
ch
ord
)
1.1(b
ase),
1.2,1.3
1.17(b
ase),
1.28,1.39
0.25R
30(shortch
ord
),
40(b
ase),
50(p
rop
mo
ved
up
-
stream
)
5 (1 2
tm
ax
)0.12(b
ase),
0.26,0.56
600?1050
3 (b
ase),
4
?
20?55
12
sh
roud
ed
-
pr
op
ell
er
mo
dels
37
Tab
le
1.1:
(con
td
.)
Test
Pro-
gram
D
t
[in
]
?d
[?]
?? d
?d
R0
Ld [%
D
t]
rlip[%
D
t]
?tip[%
D
t]
vtip[ft/s
]
N
b
?tw [?]
?75[?]
Other
notes
Fairc
hild
et
al.,
1973
[53]
12
0,
5,
10,15,
20
?
?
?
20,40,
60
(cS
).
Rotorat 16%,31%,45%,71%c
S.
?
0.52(1/16?)
220
?
19
10?30
?
=
7.3%
,
14.5%,
21.
8%
,
29.0%
Sik
orsky
S-67
Blac
kha
wk
fan
-in-fin
test-
be
d[
84,
85]
56
0
1.0
?
0.39R
?
10
?
?
?
32
25,
35
?
A?erospatial
e
Gaze
lle;
early
Dau
ph
ins
[15,
28,
79,
93]
28
(70
cm
);
35
(90
cm
)
7
?1.0
?
?0.46R
?
?
0.3(2
mm
)
690;745
13
12.5,later7;
8
1?49
?
=
0.46;
0.4
LaterA?erospatial
e
Dau
ph
ins,
Pan
ther
[29,
93]
43(110cm
)
10(7?25tes
ted)
?1.0
?
?
?
?
?
?
11
7
?
?
Bo
eing/S
ikorsky
H-76FANT
AI
LTM
demonstrator[85,
86]
47
10
1.0
?
?
?
7.5(0.9?7.5tes
ted)
?
605
8
7
-20
?
+50
?
=
0.64
Bo
eing/S
ikorsky
RAH-66Comanc
he
[87,
120]
54
10
?
?
?
?
?
?
?
8
7
-2
?
30
?
=
0.62
[88]
38
Tab
le
1.1:
(con
td
.)
Test
Pro-
gram
D
t
[in
]
?d
[?]
?? d
?d
R0
Ld [%
D
t]
rlip[%
D
t]
?tip[%
D
t]
vtip[ft/s
]
N
b
?tw [?]
?75[?]
Other
notes
Be
ll
Du
cte
d
Tail
Rotor
[80,
81]
51.5
?
?
?
?
10,
20
(cS
)
?
0.9(3/8in
=
0.73%?)
600,640,680,720
4,
5
?
?
Ev
en
and
un
ev
en
blade
spacing,
?
=
0.284?0.
422
Ka
was
aki
XOH-1
[32]
43.3(110cm
)
?
?
?
?
?
?
?
662
8
11
?
?
=
0.556.
Kamo
v
Ka-60
[92,
121]
23.6(60cm
)
8
1.1
1.25
0.35R
35
10
0.5
(3
mm
)
245
11
12
?12
?
+47
?
=
0.495
AR
OD
[35]
16(full-scale:24)
7,
14
?
?
?
?
?
?
460
?
?
?
Full-sc
ale
th
rust
=
85
lbs
Sik
orsky
Cyp
her
pro
of-
of-conce
pt
vehicle
[97]
60
?
?1.0
?
?
?
?
?
?
3(
ea.)
?
?
Coax.
W
t.
=
44
lbs.
3.8
hp
engin
e.
Sik
orsky
Cyp
her
bas
e-
lin
e
exp
eri-
me
ntal
mo
de
l
[97]
69(ID/OD
?)?
?1.0
?
?
18
in
(cS
)
?
?
?
4(
ea.)
?
?
Coax.
W
t.
=
350
lbs
.6
5hp
engin
e.
Sik
orsky
Cyp
her
tec
hnology
demonstrator[98?100]
48
?
?1.0
?
?
40(cS
)
15.6(1 2
tm
ax
)?
650
4(
ea.)
?
?
Coax.
Wt.
=
170
lbs
(em
pt
y),
300
lbs
(MTO
W
).
45?58
hp
engin
e.
39
Tab
le
1.1:
(con
td
.)
Test
Pro-
gram
D
t
[in
]
?d
[?]
?? d
?d
R0
Ld [%
D
t]
rlip[%
D
t]
?tip[%
D
t]
vtip[ft/s
]
N
b
?tw [?]
?75[?]
Other
notes
MIT
PUA
V
[122]
10
?
?
?
?
60(cS
)
?
2.5(0.25?)
?
2(
ea.)
?
?
Coax.
Shr
oud
=
NA
CA
0012.
W
t.
=
2.7
lbs
.
Dy
er,
2002
[123]
8.66(22cm
)
0,
12,
23
1.00,1.40,1.82
?
?
86
1.4,5.7,11.5
0.23,0.46(0.5,1.0mm
)
Up
to
230
2(
ea.)
?
0?35
Coax.
?
=
0.17
(ea.)
Wt.
=
12
lbs.
Fleming
et
al.,
2003
[34,
94]
10
?
?
?
?
60(cS
)
5.7(1 2
tm
ax
)1.0(0.1in
)
220?570ft/s
2
?
?
Sh
rou
d
cam-
be
red
ou
t-
wards.
Max
th
rust:
7.6
lbs
.
1.6-hp
motor.
Mar
tin
&
Tu
ng,
2004;
Mar
tin
&
Bo
xw
ell,
2005
[117,
124]
10
?0
?1.0
?1.12
0.33R
58(cS
)
1,7,2.9(1 2
tm
ax
=
5.5,
5.8)
0.5,0.75,1.0,1.25,1.5,2.0(0.05?0.2in)
85?415ft/s
2
?
?
Sh
rou
d
cam-
be
red
ou
t-
wards.
Also
tes
ted
notc
he
d
and
ste
pp
ed
shrou
d
pro-
file
s
with
exte
nded-
diam
eter
rotor
.
40
Tab
le
1.1:
(con
td
.)
Test
Pro-
gram
D
t
[in
]
?d
[?]
?? d
?d
R0
Ld [%
D
t]
rlip[%
D
t]
?tip[%
D
t]
vtip[ft/s
]
N
b
?tw [?]
?75[?]
Other
notes
Gr
af
et
al.,
2005
[125]
?
?
?
?
?
?50(cS
)
?0.8,1.8,1.9,2.2,3.0(1 2
tm
ax
?
3.1,
3.9)
?
?
?
?
?
Differen
t
lip
shap
es,
not
jus
t
diff.
lip
rad
ii.
Mi
cro
craft
LADF
[27]
9 (OD)
?
?
?
?
?
?
?
Up
to
18000rp
m
2
(8
flap
pe
d
sta-tors)
?
?
Sh
rou
d
we
igh
t:
117
gm
(1/4
lb).
GTO
W:
3.1?
3.5
lb.
Max
th
rust:
4.3
lb.
1.2-hp
en
gine.
Hon
eyw
ell
MA
V
?
?
?
?
?
?
?
?
?
?
?
?
GTO
W:
12.
5
lb.
GTRI
?Bage
l?
(Activ
e
Flo
w
Con
trol
test-
be
d)
[126?128]
11
?
?
?
?
31.8
9.1
0.1(0.01in)
170?500
3
(4
sta-tors)
?
?
Thr
ust:
12
lb.
Po
we
r:
2.4
hp
(1.8
kW).
?
=
0.17
Siroh
i
et
al.,
2005
[23,
per-
sonal
comm
u-
nic
ati
on]
6.52
10
1.14
1.20
0.22R
40
12
1.27(2.1mm
)
55?110
3
(4
sta-tors)
0
18
?
=
0.15
Hri
sh
ike
sh
av
an
&
Ch
opra,
2008
[25]
9.6(24.4cm
)
0
1.0
?
?
24
15
0.6(1.5mm
)
105?126
2
0
?
?
=
0.115
41
1.4.2.1 The early work
Stipa?s 1931 experiments [39] consisted of static and wind-tunnel tests of two
different conventional airplane propellers, approximately 20 inches in diameter, in
isolation and coupled with three different duct shapes (Fig. 1.17). The ducts, in
length approximately three times the diameter of the propellers, had the inner shape
of a venturi and the outer shape of a wing profile. The propellers were located at
the inlet plane of the ducts ? not, as later became the convention, at the point of
minimum inner diameter, or ?throat?. Since both the inlet and exit planes of the
ducts had the same diameter, the effective expansion ratio for all three ?venturi-
tubes? was equal to 1.0. Although the differences in the profile shapes of the three
ducts involved changes in more than one parameter, the general changes were a
decrease in thickness ? corresponding to an increase in throat diameter, i.e., less
constriction at the throat, a decrease in the expansion angle, and a decrease in the
inlet lip radius ? and a movement of the chordwise location of throat further back
from approximately 0.25cS to 0.45cS. In all cases tested, the ?intubed? propellers
exhibited higher thrusts, static power loadings and propulsive efficiencies than the
open propellers. Of the ducted configurations, the model with the thinner profile
exhibited lower drag, higher propulsive efficiency and higher static shroud thrust.
In 1941, Bell [118] tested a 21-inch-diameter, 24-bladed, high-solidity fan (?
= 0.86) in a duct with and without contravanes upstream of the fan, at different
fan-blade and contravane angles. When the contravanes were removed, the swirl
angle in the flow increased from approximately 5?10? to approximately 24?, and
42
(a) Wind-tunnel set-up.
(b) Models tested.
Figure 1.17: Stipa?s 1931 venturi-tube wind-tunnel tests [39].
43
maximum figure of merit dropped from 0.88 to 0.80, indicating that ?the rotational
losses without contravanes were greater than the profile drag of the contravanes.?
In the early 1940s, Kr?uger [43] conducted tests on fifteen different annular-
airfoil shrouds with a 9.5-inch-diameter propeller (Fig. 1.18). The parameters varied
between the shroud profile shapes were the chord, thickness, camber, and angle of
incidence between the profile chordline and the propeller axis. Shrouds with both
expanding and contracting ?diffusers? were tested: effective diffuser expansion ratios
varied from 0.75 to 1.70. The report that was translated into English and published
by the NACA contained data on two of the fifteen shrouds: the baseline shroud
(the right-most profile in Fig. 1.18b), which had an expansion ratio of 1.0 and a
diffuser length of approximately 34%Dt, and a much shorter, thinner shroud (the
fourth profile from the left in Fig. 1.18b) with a sharper inlet lip radius, ?d =
1.3 and Ld = 22%Dt. The models were tested at rest and in axial flow up to
airspeed ratios (?prime = v?/?R) up to 1.4, and at different blade pitch angles. In
static and near-static conditions, the sharper lip caused the flow to separate on the
second shroud, so that although the propeller thrust (Trotor) was about the same, the
shroud thrust and total thrust were substantially lower than those of the baseline
shroud. In forward flight, at the higher airspeed ratios, the thinner profile had lower
drag and therefore higher shroud thrust than the baseline shroud. However, the
propeller thrust decreased faster with increasing airspeed ratio than it did for the
baseline shroud, so, for the same blade angle, the total thrust was actually about
the same or slightly less than that of the baseline shroud. Attaching an external
split ring (Fig. 1.18c) to the leading edge of the thinner shroud helped to improve its
44
static characteristics, but worsened performance at airspeed ratios (?prime) above 0.3.
Both the baseline shroud and the thinner shroud with the split-ring attachment
achieved static thrust coefficients more than twice as high as the open propeller,
but also stalled much more sharply with increasing collective compared to the open
propeller. The maximum figure of merit achieved for the open propeller was 0.62.
For the baseline shroud, this increased to 0.69, while for the thinner shroud with
the split-ring attachment it decreased to 0.47. Adding stator vanes to the baseline
shroud caused CT and FM to increase further, while the power coefficient remained
about the same. Due to the increased inflow, both shrouded propellers achieved
their maximum FM at a blade pitch angle 10? higher then that at which the open
propeller achieved its maximum FM.
Measurments of the flow velocity at the diffuser exit plane of the baseline
shroud and at another station inside the shroud indicated that the actual expan-
sion ratio of that model was 1.07, i.e., further expansion took place downstream
of the shroud exit. Furthermore, flow visualization using smoke showed that the
cross-sectional area of the slipstream did not change with varying loading (thrust
coefficient), indicating that the expansion ratio depends only on the shape of the
shroud. The use of smoke for flow visualization was only possible when the model
was equipped with stator guide vanes. Without stators, the extreme rotation of the
flow in the wake made it impossible to obtain good photographs of the smoke. Mea-
surements of total (stagnation) pressure upstream and downstream of the empty
shrouds, that is, without the propeller present, showed that both the shrouds had
about the same, negligible amount of pressure loss in the the internal flow due to
45
(a) Model configuration.
(b) Shroud shapes tested. The arrow marks the location of the propeller.
(c) Leading-edge split-ring configurations tested.
Figure 1.18: Shroud shapes tested by Kr?uger [43].
46
shroud wall friction and viscous effects. However, the report mentions that another
shroud, which had the largest expansion ratio tested (1.7), had pressure losses about
twenty times higher. This translates directly to greater power input requirements
and lower efficiencies. Kr?uger noted that ?since the [diffuser efficiency] will hardly
increase beyond 0.9 unless special steps are taken, ... an attempt to increase the
slipstream cross-section ratio essentially beyond [??d] = 1 would be useless.?
In addition to the baseline, 8-bladed propeller used in these tests, two other
propellers were also tested. Reducing the number of blades from eight to four
resulted in CT, CP and FM decreasing at any given blade angle, because of the
reduced inflow, while using blades with a reduced amount of twist (-37.5? vs. the
baseline -45?), so as to take advantage of the tendency for the inflow velocity to
increase towards the shroud wall, resulted in a negligible change in performance.
In 1948, Platt [52] tested contra-rotating, 48-inch-diameter propellers with
three different shrouds (Fig. 1.19) which had two different values of expansion ratio
(1.1 and 1.3), resulting from three different diffuser angles (7?, 14.4?and 22.4?) and
two different diffuser lengths (33.6% and 50.2%Dt). At the same expansion ratio,
the shroud with the longer diffuser (smaller ?d) resulted in a higher value of CT/CP,
when compared at the same CP. At the same diffuser length, the shroud with
smaller expansion (smaller?d) resulted in lowerCT/CP, but also came slightly closer
to producing the ideal amount of thrust for any given power. At all propeller blade
angles, increasing ?d caused a reduction in power consumption, but the decrease,
which was most pronounced at the highest blade angles, was never more than 10%.
In general, the differences in performance between the three shrouded propellers was
47
Figure 1.19: Shroud shapes tested by Platt [52].
found to be very marginal, within the limited range of blade angles tested (35?, 40?
and 45?). On the other hand, all the shrouded propellers produced CT/CP ratios
that were more than twice as high as the open propeller, at the same CP, and came
much closer to producing the ideal amount of thrust; then again, these results were
obtained at the higher collectives at which the open propeller had already stalled.
In 1950, Hubbard [119] tested a fixed-pitch, 48-inch-diameter propeller with
four shrouds (Fig. 1.20) that had three different diffuser lengths, from 12% to 36%Dt,
and three different inlet lip radii (1.07?1.49%Dt). Shroud thrust, which was mea-
sured by integrating the measured static pressure variation over the shroud surfaces,
was found to increase with increasing Ld. By shortening the propeller blades, tip
clearances of 0.2% to 4.4%Dt were tested with the shroud that had rlip = 1.49%Dt
48
Figure 1.20: Hubbard?s test arrangement [119].
and Ld = 24%Dt. The increase in tip clearance resulted in approximately an 84%
reduction in total thrust, mostly due to a drop in shroud thrust, and a 15% reduc-
tion in power. In both Hubbard?s and Platt?s experiments, the incoming flow was
found to separate at the shroud leading edge at low propeller rotational speeds ?
leading to loss of the inlet suction and large reductions in shroud thrust ? but then
re-attach as the speed was increased.
In 1955, Parlett [55] investigated the effect of changing lip radius on the static
thrust of a fixed-pitch, 18-inch-diameter shrouded propeller (Fig. 1.21). The shroud
was a simple cylindrical shell, with an expansion ratio of 1.0, and was tested with
lip radii varying from 1.4% to 8.3%Dt. The increase in rlip caused a near-linear
increase in figure of merit (at the same collective) from 0.72 to 0.89.
In 1958, Taylor [57] tested configurations with five different inlet lip radii (0?
12.5%Dt), three diffuser angles (0?, 7?, 14?) and four diffuser lengths (3?103%Dt),
49
Figure 1.21: Shrouded propeller tested by Parlett [55].
50
thereby creating expansion ratios from 1.0 to 1.5 and allowing for examination of
the same expansion ratio with different diffuser angles (Fig. 1.22). Although the
test set-up was a propeller embedded in a wall, to simulate the ?upper? surface of a
fan-in-wing configuration, the results are nevertheless also applicable to a shrouded-
rotor configuration in the static (hovering) condition. The maximum figure of merit
of the open propeller was 0.64, at a collective (?75) of 13?. In comparison, for the
model with an expansion ratio of 1.0 and the longest diffuser, the maximum FM
occured at a higher collective angle of 25?. The highest values of FM were achieved
by the models with the inlet lip radii of 12.5% and 6.25%Dt, with a maximum value
of 0.75. For the smaller lip radii, the flow was found to have separated from the
inlet surface, and reducing rlip caused FM to decrease. For rlip less than 3%Dt,
the FM was lower than that of the open propeller, with values around 0.5 for the
smallest lip radius. The decrease in FM with reduction in lip radius coincided with
simultaneous decrease in the shroud thrust fraction (Tshroud/Ttotal), from 50% (in
agreement with momentum theory predictions) with the larger lip radii to 30% at
the smallest lip radius.
At a fixed diffuser angle of 0? (?d = 1.0) and with the larger lip radii, decreasing
the diffuser length from 1.03Dt to 0.03Dt resulted in a decrease in FM from about
0.75 to 0.68. A similar decrease was seen when, with the largest lip radius, ?d was
increased from 1.0 to 1.5, whether by increasing the diffuser angle or its length.
However, even though the FM decreased, the total thrust was found to increase
with increasing expansion ratio. The shroud thrust fraction (Tshroud/Ttotal) increased
exactly in accordance with the momentum theory prediction, and, when compared
51
(a) Test arrangement.
(b) Shroud shapes tested.
Figure 1.22: Shrouded-propeller tests conducted by Taylor [57].
52
at the same power input, the increase in thrust over the open propeller (TSR/TOR)
also increased as predicted, but with values 15?20% greater than the momentum
theory prediction. (This, as was mentioned earlier, was attributed to the tip losses
on the open propeller.) In this set of experiments, these results were found to be
virtually independent of the expansion angle of the diffuser, and depended only on
the value of the expansion ratio.
In 1966, during the development of the 7-foot-diameter ducted fans for the Bell
X-22A, Mort reported [72] that an increase in tip clearance from 0.3% to 2.0%Dt
resulted in approximately a 20% drop in static thrust coefficient on a full-scale
model. This was principally due to a reduction in shroud thrust rather than in
propeller thrust. The power consumed also decreased, but by a lesser amount, so
that the ratio of shroud thrust to power consumed also decreased. A comparison
of these results with those obtained by Hubbard [119] on a smaller, 4-foot-diameter
model revealed that for the smaller model, CTshroud and CP were individually less
affected by changes in ?tip, compared to Mort?s 7-foot model, but the CTshroud/CP
ratio decreased much more rapidly with increasing ?tip than for the larger model.
In axial flow (forward flight), the reductions in the thrust and power coefficients
because of the larger tip clearance became smaller with increasing advance ratio.
At around the same time, a comprehensive, three-year investigation of the ef-
fects of various shroud parameters was conducted at Hamilton Standard by Black,
Wainauski and Rohrbach [114] under a contract for the US Navy. Twelve differ-
ent shrouded-propeller models, with a propeller diameter of 30 inches, were tested
(Fig. 1.23). The parameters investigated included the exit area ratio, lip shape,
53
Figure 1.23: Shroud models tested by Black et al. [114].
shroud chord, propeller location within the shroud, effect of inlet and exit vanes,
propeller blade shape, number of blades, blade tip clearance, and shroud external
shape. Of all these, ?results showed that the most powerful shrouded propeller vari-
able [was the] shroud exit area ratio.? The models were tested in static conditions
and in axial flight up to a Mach number of 0.6. For the baseline model (??d = 1.1),
the increase in static thrust over the open propeller, at the same power consumption,
was greater than that predicted by momentum theory. Compared to the predicted
value of 1.3 for TSR/TOR (for the same ideal power), an experimental sweep of dif-
ferent rotational speeds at fixed collective angle resulted in TSR/TOR values from
1.67 at lower rotational speeds (power inputs) to 1.90 at the higher speeds. This
greater-than-predicted improvement was attributed to the increased efficiencies due
to off-loading of the rotor by the shroud, but could also have been due to lower
blade tip-vortex losses, as has been explained earlier.
The thrust on the shroud was measured by a separate force balance, as well
54
as from measurements of surface pressure distributions. The difference between the
two methods yielded an estimate of the drag on the shroud due to skin friction. In
static conditions, the baseline shroud model came very close to achieving the theo-
retical Ttotal/Trotor ratio (Eq. (1.5)) of 2.2. In axial flight, with increasing freestream
velocity, the drag on the shroud caused Tshroud to reduce and eventually become
negative, resulting in progressively lower values of Ttotal/Trotor and TSR/TOR. In
all cases, though, increasing the input power, either by increasing rotational speed
or by increasing collective, caused Tshroud and Ttotal to increase, by increasing the
suction forces on the shroud inlet. In fact, the report explicitly states that ?through-
out the program, increasing power [input] was found to be a very influential factor
whereby the effectiveness of the shroud generally was increased.? The authors there-
fore concluded that greater benefit from shrouding would probably be seen by using
high-solidity propellers that are capable of absorbing greater amounts of power. In
static conditions, while power loading (T/P) did decrease with increasing collec-
tive, this was probably because the propeller was being operated past its collective
for maximum efficiency. In axial flight, as the freestream velocity was increased,
the variation in power loading changed to increasing with increasing collective. In
general, with increasing axial velocity, the power input and collective for maximum
power loading also increased.
The effects of changing the various shroud parameters were expressed in terms
of the change in total thrust from that of the baseline shrouded propeller, at the
same power consumption. Increasing the area ratio from 1.1 to 1.2 and 1.3, while
keeping the diffuser length constant (i.e., increasing the diffuser angle from 7? to
55
14? and 20?) caused the total static thrust to increase significantly, then level off
and decrease. Compared to the baseline shroud, up to a 15% further increase in
static thrust was obtained at ??d between 1.2 and 1.3. Considering that the baseline
shrouded propeller produced up to 90% more thrust than the open propeller at
the same power input, this means that thrusts up to 2.2 times that of the open
propeller, at the same power, were achieved with these higher area ratios. Here too,
the thrust increments were greater than those predicted by the simple momentum
theory. Based on results from an associated experimental program in which ?a
long-chord, 1.4 area ratio shroud was tested with extremely favourable results,?
the authors surmised that the performance levelling-off was ?more likely due to a
diffusion rate limitation,? i.e., an expansion-angle limitation, rather than to an area-
ratio limitation. In axial flight, as freestream velocity was increased, increasing the
area ratio caused performance to deteriorate severely, and, in the same associated
experimental program mentioned above, ?area ratios of 0.9 and 1.0 were tested
and shown to be beneficial? at the higher speeds. Shortening the shroud chord
? effectively, reducing the diffuser length from the baseline 40%Dt to 30%Dt ?
while maintaining the area ratio at 1.1, and thereby increasing ?d from 7? to 9.3?,
caused little to no effect in static conditions, but resulted in degradations in thrust
at the higher axial velocities. This was attributed to the higher thickness ratio of
the shorter shroud. Two propeller positions within the shroud were tested: the
baseline, with the propeller 40% of the chord-length back from the shroud inlet
plane, and a second case with the the propeller moved forward to the 0.25cS station.
Having the propeller further forward gives the wake more axial length to achieve
56
the desired diffusion, whereas having it further rearward helps to even out any
asymmetries in the inflow. Since the models were only tested in axial flow, the
effects of inflow asymmetries due to operation at an angle of attack to the freestream
were not evaluated; however, it was found that the forward propeller location gave
better results in axial forward flight, while the rearward location performed better
in static conditions. A significant result of these tests was that, from the shroud
surface pressure measurements, it was seen that on almost all the models tested,
the pressure at diffuser exit was greater than atmospheric, and that ?only on rare
occasions did the pressure there equal the ambient value.? This indicated that
the wake had over-expanded and would have to re-contract downstream of the exit
plane. No estimates were given of the true final area of the wake in this study.
The effects of changing the other shrouded-propeller parameters were much
smaller compared to those of changing the diffuser parameters. Three different
propeller blade planforms were tested: rectangular blades, blades with wider tips
which were designed for higher tip loadings, and blades with narrower tips, which
were designed for higher shank loadings. The results varied with input power and
with freestream Mach number, but the differences in total thrust were never more
than 5%. Notably, these variations were more so due to changes in the propeller
thrust rather than the shroud thrust. With narrow-tipped blades, a three-bladed
model fared better in static and near-static conditions, while a four-bladed model
produced more thrust at the higher Mach numbers. For this parameter ? blade
number ? too, though, the variations were within 5% of the baseline. Both inlet
and exit vanes, in the form of untwisted, uncambered, constant-chord airfoils, set
57
at up to ?10? from the axial direction, caused performance to either deteriorate
slightly or hardly change at all. Increasing the blade tip clearance from the baseline
0.12%Dt up to 0.56%Dt caused performance to consistently deteriorate only in static
and near-static conditions, and even there the degradations were by less than 5%
of the baseline thrust. At higher axial velocities, thrust first decreased when ?tip
was increased to 0.26%Dt, and then increased again when the clearance was further
increased to 0.56%Dt.
1.4.2.2 Helicopter tail rotors
For helicopter tail-rotor applications, the shrouded rotor has typically only
been used on small- and medium-sized helicopters. For larger helicopters, the size
and weight of the fan-in-fin assembly grow disproportionately large; also, safety
considerations in ground handling become less critical because a conventional tail
rotor is typically mounted at the top of a pylon, and lateral tilt angles for equilibrium
in hover with a low-mounted fan-in-fin become uncomfortably large [29, 79]. An
important constraint on the application of the shrouded-rotor concept to helicopter
tail rotors is that the length of the shroud is limited by the maximum allowable
thickness of the tailboom and empennage structure. In the ideal case, a uniform
inflow distribution is desired at the plane of the rotor. Therefore, a large shroud
length is desirable, with the rotor placed as far back from the inlet as possible,
so that in forward flight for the helicopter ? which is a sideways-flow condition
for the tail rotor, a long length of shroud is available for the the flow to return
58
to a nearly-uniform profile after turning from the freestream direction to the rotor
axial direction. This requirement is less important in low-speed flight or high-thrust
conditions, when the inflow distribution tends to be fairly uniform already, but it
does become critical in high-speed/low-thrust conditions. A long shroud length, i.e.,
a low shroud thickness-to-chord ratio, also allows for smooth fairing of the trailing
edge of the shroud, for low drag in axial-flow conditions (sideways flight for the
helicopter). On the other hand, in sideways flight for the tail rotor (forward flight
for the helicopter), a long shroud length also results in a larger bluff body being
presented to the oncoming flow, and consequently higher drag.
In 1975, Clark [84] reported on an innovative solution to this problem that
was developed at Sikorsky, in designing the fan-in-fin that was tested on the S-67
Blackhawk helicopter. This solution was a shroud design (Fig. 1.24) that was not
radially symmetric: On the forward half of the shroud, towards the tailboom, the
inlet of the shroud was swollen to increase its chord length, before gradually fairing
into the lesser thickness of the tailboom, while on the opposite half, the shroud was
allowed to taper towards the rear of the empennage. This was possible because, in
sideward flow (helicopter forward flight), the stagnation point that is formed on the
aft side of the inlet ?washes out the sensitivity to duct length seen on the [forward
side],? and so the aft side of the shroud need not be as deep as the forward side. Ad-
ditionally, the cross-sectional profile shape of the aft side of the shroud was changed
from an ovoid shape to a cusped shape, with a concave cut-out on the exit side of the
external profile. This cusp design helped reduce weight by doing away with a long
aft fairing, reduced the strength of the separated flow on the aft surface in sideways
59
flight, compared to the otherwise convex surface of a ?doughnut?-shaped shroud,
and reduced the area around the diffuser exhaust, minimizing negative thrust due
to separated flows in that region.
As has been described already, the quality of the inflow distribution of the rotor
is also affected by the shape of the inlet lip. Based on a survey of experimental fan-in-
wing data, an inlet lip radius of 10%Dt was selected for the S-67?s fan-in-fin, in order
to achieve a uniform inflow distribution as closely as possible, ?while maintaining
fully attached flow in all phases of flight? [84]. Later, during 1/3- and 1/4-scale
tests of the FANTAILTM for the Boeing/Sikorsky LH (Comanche) helicopter, tests
with different lip radii showed that reducing rlip from 6.5%Dt to 0.9%Dt resulted
in separation of the flow from the inlet surface and approximately a 7% decrease
in maximum figure of merit [85]. According to this reference, ?it was found that
to obtain maximum hover performance,? rlip values greater than 6.5?7.5%Dt were
required. A lip radius of 7.5%Dt was finally selected for the LH. It was also found
that a sharp diffuser exit radius was required ?for best static performance,? but
negative thrust ? for helicopter directional control ? and helicopter forward flight
performance ?were compromised.? Based on additional test results, the radius of
the aft 1/3 of the FANTAIL exit lip was increased to that of the inlet, as this was
found to provide ?a measurable drag reduction ... [and] improve reverse thrust
performance.?
In 1973, in investigating the same problem ? the applicability of the shrouded-
rotor configuration to helicopter tail rotors ? Fairchild et al. [53] conducted exper-
iments on short-chord shrouded rotors, with total shroud lengths (cS) varying from
60
(a) Fore-aft taper of shroud length, with swelling at the tailboom joint.
(b) Streamlines around ovoid (?doughnut-shaped?) vs. cusped aft sections of the shroud.
Figure 1.24: Shroud design of the Sikorsky S-67 Blackhawk?s fan-in-fin:
cross-sectional top views [84].
61
20%Dt to 60%Dt and diffuser included angles from 0? to 20?. With the longest
shroud and ?d = 15?, the maximum thrust achieved was twice as much as the maxi-
mum thrust from the open rotor. Power measurements were made but not presented
since excessive scatter in those data rendered them inconclusive. In general, increas-
ing the length of the shroud was found to increase the total thrust and the shroud
thrust fraction, Tshroud/Ttotal. At fixed diffuser angle, the increase in shroud length
from 0.2Dt to 0.6Dt caused Tshroud/Ttotal to increase from approximately 36% to
53%. At low expansion angles, the effect of changing cS/Dt was greater at the lower
values of cS/Dt, while at the higher expansion angles, the effect was more uniform.
At fixed shroud length, increasing the diffuser angle from 0? to 20? was predicted
to cause Tshroud/Ttotal to increase. However, the experimental data showed this pa-
rameter (?d) to not have a strong effect: at cS/Dt = 0.6, the shroud thrust fraction
increased very slightly, while for the shorter shrouds it decreased slightly. This was
attributed to flow separation in the diffuser, due to the low disk loading of the test
model (5?7 lb/ft2).
From the data for the shrouds with ?d = 15?, it was seen that, at the lowest
collective angle of 10?, there was little difference between the total thrust of the open
rotor and the shrouded rotors, while at the higher collectives, the shrouded rotors
produced greater thrust than the open rotor. With increasing collective, the effects
of shrouding and of increasing the length of the shroud became more pronounced,
with the initial increase in thrust due to adding a shroud being greater than the
further increase from increasing the length of the shroud. This was due to the fact
that the open rotor stalled at a collective of around 20?, while the shrouded rotors
62
?did not indicate a stall up to the maximum angles tested? (30?). Increasing the
length of the shroud seemed to increase the linearity of the thrust?collective curve,
and postpone stall to higher collectives.
Tests were also conducted on the effects of rotor solidity and the location of
the rotor within the shroud. At fixed diffuser angle (15?) and fixed shroud length,
increasing the solidity of the rotor from 7% to 30% caused Tshroud/Ttotal to decrease.
For cS/Dt= 0.6, the decrease occurred uniformly over the range of solidities tested,
whereas for the shorter shrouds, the decrease occurred mostly between ? = 7%
and 15%, after which it remained mostly constant. The highest shroud thrust
fraction was therefore achieved with the longest shroud, the largest expansion angle
(although not significantly so), and the lowest solidity rotor. The effect of rotor
location was tested in the shroud with ?d = 0? and cS/Dt = 0.6. Moving the
rotor from the 0.16cS location back to the 0.71cS location resulted in the thrust
measurements showing a slight maximum when the rotor was at the mid-chord
location (0.5cS). However, these data were taken with the rotor at a collective of
10?, at which the shroud had minimal effect on the rotor performance.
In the mid-1990s, tests at Bell Helicopter Textron on Ducted Tail Rotor (DTR)
models (Fig. 1.25) with different shroud lengths (called the ?thickness ratio? in the
reports) ? 10% and 20%Dt ? showed greater reductions in power from the base-
line open rotor and greater ?duct thrust sharing? in hover for the longer (?thicker?)
shroud [80]. Tests of rotor blades with tapered tips showed ?a slight performance
degradation compared to the [baseline] square tip rotor.? Part of the reason for
this was that ?by reducing the blade loading at the tip, the suction on the duct
63
Figure 1.25: Ducted Tail Rotor tested by Bell Helicopter Textron [80, 81].
decreased, resulting in a 6% reduction in thrust produced by the duct.? Other rotor
parameters were also varied, as part of a simultaneous investigation of the acous-
tic characteristics of the DTR, but full-scale tests on a Bell Model 222U helicopter
showed ?no measurable difference? in performance between the baseline, 4-bladed
rotor with equally-spaced blades and a quieter, 5-bladed rotor with unevenly-spaced
blades, increased solidity and reduced tip speed.
With the shroud length limited by the width of the tailboom, helicopter fan-in-
fan tail rotors have typically had expansion ratios of little more than 1.0. Reporting
on the development of the FANTAILTM shroud for the Boeing/Sikorsky LH design,
and its testing on an H-76 demonstrator aircraft, Keys et al. [85] stated that the
shroud was designed with a ?divergence angle? (half the included angle) of 5? to
64
?prevent premature flow separation from the duct walls while maximizing perfor-
mance.? They also state that thrust augmentation (Ttotal/Trotor) values of 1.8 to
almost 2.0 were achieved, compared to the ?theoretical maximum of 2.0,? and this,
along with other analyses presented in this reference and in Ref. [86], indicates an
expansion ratio of 1.0. Tests on the Kamov Ka-60?s fan-in-fin (?d = 8?,??d = 1.1) [92]
also showed a rotor thrust fraction (Trotor/Ttotal) of 0.55, or an augmentation fac-
tor of 1.82. A?erospatiale helicopter fenestrons also had shroud expansion ratios of
approximately 1.0 [93], although tests on fenestron models with diffuser included
angles varying from 7? up to 25? resulted in maximum thrust occurring at ?d =
20? [29, 93]. On the production helicopters such as the later Dauphins, however,
this angle was limited to 10? due to ?flow instabilities? arising from interactions with
the main rotor wake [93]. Adverse interactions with the main rotor wake, leading to
a dead-band in the yaw control authority (the problem did not occur with isolated
fenestron test models), were also the reason for a reduction in blade twist from
?12.5? on the early Gazelle fenestrons to ?7? on the later aircraft [28, 29, 79, 93].
Similar reasoning, based also on experience with ?32? of twist on the fan-in-fin of
the modified Sikorsky S-67 Blackhawk, led to the adoption of -7? of twist on the
Boeing/Sikorsky FANTAILTM [85].
As in other shrouded-rotor applications, the presence of the shroud caused
the stall of the fan-in-fin rotors to be postponed to extremely high blade collective
angles. In tests of the Ka-60?s fan-in-fin [92], it was found that, up to the collective at
which the open rotor stalled, the shrouded rotor produced only slightly more thrust
than the open rotor, at the same power. However, because the shrouded rotor didn?t
65
stall until a much higher collective of around 55?, its maximum achieved thrust was
twice as high as that for the open rotor. Similarly, a 3/4-scale model of the LH
FANTAIL stalled at a collective (?75) of 48? [86], and, in half-scale model tests of a
Gazelle fenestron with?d = 7? and with a rotor with?7? of twist, stall occurred at a
collective (?70) of 45? [93]. In this A?erospatiale test, stall first occurred at the blade
tips and then proceeded radially inwards, as evidenced by measurements of the flow
velocities in the wake, which showed a widening zone of substantially decreased axial
velocities near the shroud wall. Later design iterations of the fenestron incorporated
radial variations in blade thickness and airfoil geometry intended to increase the load
at the blade tip, ?so as to get the maximum depressure level on the shroud and delay
as far as possible the blade tip stall? [93]. Prior to stall, at ?70 = 35?, the average
flow swirl (rotational) angle was about 10?, while at ?70 = 47?, post-stall, the mean
swirl angle was 18? in the core of the wake, increasingly sharply to up to more than
60? near the shroud wall. Surprisingly, the measurements of the shroud surface
pressure distributions show that even after stalling, the rotor continued to maintain
strong suction pressures on the shroud inlet; indeed, the suction on the inlet at
?70 = 47? (post-stall) was greater than that at 35? (pre-stall). The pressure-jump
through the rotor remained approximately the same between the two conditions, so
that the suction pressures on the diffuser were also greater at the higher, post-stall
collective. These pressures, of course, are due to to the total velocity at the shroud
surface, which includes both the axial and the tangential (swirl) components of the
velocity, so it is quite likely that while the axial velocites at the wall decrease after
the rotor blades have stalled, it is the increased swirl velocities that maintain the
66
high suction pressures on the shroud.16
Stator guide vanes can be used to recover this rotational energy and produce
additional thrust, while also functioning as structural support struts for the rotor
centerbody. However, these were only incorporated on the most recent versions of
A?erospatiale?s fenestrons ? on the later Dauphins and subsequent aircraft. Even
the design of the LH FANTAIL only used elliptical cross-sections for the supporting
struts and the driveshaft cover, for the sole purpose of minimizing their drag [85].
With stators, almost complete straightening of the flow was achieved in tests at
A?erospatiale [93], as well as increases in thrust and the ability to reduce the length
of the diffuser required for pressure recovery, and thereby the thickness of the heli-
copter empennage. In half-scale model tests with ?d = 10?, the use of stator guide
vanes caused thrust to increase by 11%; in full-scale ground tests of Dauphin fene-
strons, the addition of guide vanes caused thrust to increase by 26%, and maximum
FM by 4.2% up to a value of 0.74 (the use of blades with newly-developed OAF
airfoil sections led to a further 11% increase in thrust and 3% increase in FM, up
to 0.76) [29, 93]. 3/4-scale tests of the LH FANTAIL resulted in a maximum fig-
ure of merit of 0.69 [85, 86], which is slightly lower compared to the A?erospatiale
data presented in Ref. [93], but also a slightly higher maximum loading coefficient,
CT/?17, in hover of 0.31, which is about twice as high as that typically achieved by
16The same pressure measurements also showed that the flow pressure at the diffuser exit plane
was still slightly less than ambient, indicating that further expansion of the wake took place
downstream of the diffuser exit [29, 93].
17For a shrouded rotor, because CT = CT
total and not CTrotor, this quantity cannot be considered
as a ?blade-loading? coefficient, as is done for an open rotor. The equivalent quantity would be
obtained by multiplying this coefficient by the factor 12?
d
, which requires knowledge of the true
value of the expansion ratio of the wake.
67
an open rotor [86].
1.4.2.3 Unmanned aircraft
During the design of the AROD UAV (Fig. 1.6m, p. 15) for the US Marine
Corps, Weir reported in1988 [35]on tests conducted atSandiaNational Laboratories
on six different shrouded-propeller configurations (Fig. 1.26). The full-size vehicle
had a propeller diameter of 2 feet and was required to produce 85 lbs of thrust,
whereas these tests were conducted on 2/3-scale models (16-inch diameter). The
tests were to investigate the designs of the inlet, diffuser, and stator control vanes.
The baseline shroud had an annular airfoil profile with a diffuser included angle
of 14? and four all-moving vanes attached below the diffuser exit plane in a cross
configuration. The principal changes that were made to this configuration in creating
the other shrouds were (a) to reduce ?d to 7?, (b) to reduce ?d to 7? and extend
the diffuser so that the vanes were shielded from the freestream, (c) to replace the
annular airfoil with a torus, thereby increasing the inlet lip radius, and a conical
diffuser with ?d= 14?, and (d) to remove the conical diffuser, thereby leaving a
shroud that consisted of the torus only. For comparing the different configurations,
only translational-flight data were presented in this reference, and these data were
taken with the models in an ?equilibrium? condition with no net drag. Therefore,
since the models were tested at a fixed blade pitch angle and a fixed rotational speed,
the data at low angles of attack (axial flow) were taken at the higher tunnel speeds
(?prime up to 0.27), while the data at high angles of attack (edgewise flow) were taken
68
(a) Annular shroud. (b) Toroidal shroud.
Figure 1.26: Shroud configurations tested for the AROD by Weir [35].
at the lower tunnel speeds. Additionally, high-speed (low-?) data were only taken
for two of the configurations: the annular airfoil shroud with the diffuser extension,
and the toroidal shroud with no diffuser.
For all the shrouds, the general trend was for the lift (CL) and pitch moment
(CM) coefficients (as normalized byvtip, notv?) to increase with increasing airspeed
(and simultaneously decreasing angle of attack), reach a maximum around ?prime ?
0.125 (? ? 45?50?), and then decrease ? as would be expected. In all cases,
the differences in CL due to the configuration changes were much smaller than the
differences in CM ? almost smaller than the limits of experimental error, even.
Between the annular-airfoil shrouds, decreasing ?d from 14? to 7? led to an increase
in CM but almost no change in CL. Extending the 7? diffuser to shield the vanes
caused a slight decrease in CL, but a large increase in CM. This was because, when
the vanes were exposed, the freestream air caused the exhaust jet to bend away from
69
the propeller axis, increasing the angle of attack at the control vanes and creating a
larger nose-down (negative) pitch moment that opposed the natural nose-up moment
of the shroud. With the vanes shielded from the freestream, the shroud?s pitch
moment is unopposed, and increases to that of a shroud with no vanes. The toroidal
shrouds had higher CM and slightly higher CL than the annular-airfoil shrouds at
low-?prime/high-? ? this was attributed to their larger lip radius; however, the toroidal
shrouds also changed, with increasing ?prime/decreasing ?, to decreasing values earlier
and more abruptly than did the annular shrouds. Surface flow visualization using
tufts showed that this sudden change was not due to any sort of flow separation.
At high-?prime/low-? (near-axial flow conditions), the extended-diffuser/annular-airfoil
shroud had higher CL and CM than the no-diffuser/toroidal-inlet shroud. Between
the two toroidal shrouds, the one with the diffuser had higher values of CL and CM
than the one without ? again, with the differences in CM being much larger than
those in CL.
The Cypher (Fig. 1.6n, p. 15) was a larger, 4-foot-rotor-diameter UAV devel-
oped by Sikorsky from 1986 to 1998 [96?110]. Unlike previous shrouded-rotor UAV
designs, which utilized a single rotor with vanes in the downwash for anti-torque
and yaw as well as pitch and roll control, the Cypher instead used a pair of counter-
rotating, coaxial, bearingless rotors, with each rotor having independent cyclic and
collective controls.18 Use of a coaxial rotor system eliminated gyroscopic coupling in
pitch and roll that would have occurred with a single-rotor system [97, 103]. Also,
18The Hiller VZ-1 flying platform (Fig. 1.6a, p. 13) also used a coaxial rotor system, but the
propellers were of fixed pitch. Yaw and thrust were controlled by varying the propeller rotational
speeds, and pitch and roll by kinesthetic control from the human pilot [103].
70
these six control degrees of freedom provided complete control of the aircraft along
all three axes, ?thereby eliminating the requirement for additional [control] surfaces
or equipment? [103].19 The azimuthal variations in inflow over the shroud inlet
caused by the cyclic variation in blade angle resulted in pitching moments induced
on the shroud that were more than six time higher than those on the rotor alone,
thereby greatly increasing their control authority. For example, in forward flight,
at an airspeed of 70 knots and with the 300-lb aircraft at a 5-degree nose-down
attitude, only 11? of forward cyclic was required to trim the aircraft and neutralize
the 1650 ft-lbs of nose-up pitch moment that was being otherwise produced by the
shroud [103]. In addition, the external shape of the shroud of the final Technol-
ogy Demonstrator (TD) was tailored to itself produce a nose-down pitch moment
in translational flight, so as to reduce the cyclic trim requirements [98]. With this
shaping, the shroud moment ? and rotor cyclic requirement ? increased to a max-
imum at an airspeed of 40 kts and decreased thereafter. As of 1998, the TD aircraft
had achieved a flight endurance of 2?3 hours and a range of 30?50 miles, depend-
ing on the mission profile, a maximum speed of 60 mph (52 kts) and a maximum
altitude of 8000 feet [100]. Table 1.1 shows the evolution of the final TD aircraft
from the initial proof-of-concept vehicle, and Ref. [103], a patent filed in 1990, gives
an indication of some of the experimental work performed in designing the shroud:
19Weir [35], on the other hand, argued for using vanes in the downwash for UAV control, stating
that ?reaction control jets and a cyclic pitch propeller are too complex. Control surfaces or vanes
in the propeller jet stream offer the best solution as they are relatively simple and take advantage
of the high dynamic pressure in the jet under all conditions.? Similar sentiments were expressed by
Lipera [27] in describing the design of the Allied Aerospace iSTAR UAV (Fig. 1.6p): ?Because of the
absence of complex mechanisms common with other VTOL aircraft (counter-rotating propellers,
gearboxes, articulating blades, etc. ...) the iSTAR also benefits from a reduced part count
contributing to improved reliability, low structural weight and low cost.?
71
Two separate sets of figure-of-merit data show that decreasing the inlet lip radius
caused FM to either steadily decrease from about 0.88 to 0.72 when rlip was reduced
from 17% to 3%Dt, or to remain constant at approximately 0.74 as rlip was reduced
from 14%Dt to 5%Dt, and then sharply decrease to 0.50 at a lip radius of zero.
Another noteworthy result shown was that, for two different values of lip radius
tested (6.25% and 12.5%Dt), reducing the shroud length from 100%Dt to less than
5%Dt caused the thrust to decrease by less than 10%, in contrast to what appears
to be an analytical result that predicted a decrease of more than 30%, and also that
there was very little difference in thrust between these two lip radii.
Another shrouded-rotor UAV design that used a coaxial rotor system was
the 10-inch-diameter Perching UAV, or PUAV, developed at MIT?s Charles Stark
Draper Laboratory during 2000?2002 [122, 123, 129] (Fig. 1.27). Unlike the Cypher,
though, the PUAV only had collective control of the rotors, and therefore also used
vanes for pitch and roll control. The unique feature of the PUAV design was that
the direction of rotor thrust was reversed between the hover and forward-flight
conditions (Fig. 1.27b), by changing the blade pitch to negative angles of attack.
The shroud profile shape was that of a NACA 0012 airfoil, arranged so that the
leading edge was the inlet during forward (axial) flight, with the vanes upstream
of the shroud, acting as canard control surfaces. In hover, with the direction of
thrust reversed, the sharper trailing edge of the annular airfoil became the inlet,
and the vanes operated in the rotor downwash. Not surprisingly, in static tests of
the prototype vehicle, the maximum thrust of the shrouded-rotor model was 25%
lower than that of the open-rotor model [122]. These initial tests were followed by
72
an investigation of the various shroud parameters by Dyer [123], who tested models
with three different lip radii (1.4?11.5%Dt), three different diffuser angles (0?23?,
??d = 1.00?1.82) and two different blade tip clearances (0.23% and 0.46%Dt). The
reduction in tip gap caused the power consumption to decrease, but not significantly
so. Increasing the lip radius caused thrust and power loading (T/P) to increase, as
expected. On the other hand, increasing the diffuser angle caused thrust and power
loading to decrease, contrary to predictions. Surveys of the wake with a pitot probe
and with a tufted wand revealed a large reversed-flow region in the core of the wake.
This was due to separation of the flow from the rotor centerbody, and was probably
what prevented the larger diffusers from producing the desired expansion and the
corresponding performance improvements.
In 2003, Fleming et al. [34, 94] investigated the effectivness of seven different
auxiliary control devices (Fig. 1.28), in addition to conventional aft-mounted vanes,
for controlling ducted-fan VTOL UAVs in crosswind turbulence. The devices, some
of which were inspired by the patents related to Paul Moller?s ?Aerobot? line of
ducted-propeller UAVs [130, 131], were tested separately, and included (1) an inter-
nal duct vane, mounted immediately below the rotor on the windward side of the
duct, (2) a retractable ?duct deflector? on the windward side of the internal duct
wall, with and without bleeding of air to the outside of the duct, (3) a trailing edge
(diffuser exit) flap on the leeward side of the duct, to increase the effective camber of
the duct profile, (4) a deflector trough or ?thrust reverser? on the windward side of
the diffuser exit, (5) a spoiler on the windward side of the inlet lip, (6) a lip extension
or ?leading edge slat? on the leeward side of the inlet, and (7) tangential and normal
73
(a) Vehicle cutaway view.
(b) Vehicle flight modes.
Figure 1.27: MIT/Draper Perching UAV (PUAV) [122].
74
blowing of pressurized air on both windward and leeward sides of the inlet. The goal
of the investigation was a device that would augment, not replace, the conventional
aft-mounted vanes, and which would provide as close to a pure pitch moment as
possible, without generating any side-force or additional drag.20 The test model
was similar in form to the 10-inch-diameter Honeywell Kestrel (Fig. 1.6o, p. 15),
although without the two side pods present on that aircraft, and was tested with
a fixed-pitch propeller in hover and in crosswinds (edgewise flow) of speeds up to
30 knots (50 ft/s). In hover, the models were tested at different rotational speeds,
with tip speeds from 200 to 500 ft/s. Above a tip speed of about 300 ft/s, the
baseline shrouded propeller (no vanes, no auxiliary control devices) averaged a 20%
increase in thrust coefficient above the open propeller (0.0205 vs. 0.017). At these
higher tip speeds, CT remained fairly constant, while at the lower tip speeds, CT
decreased near-linearly with decreasing vtip. In the cross-wind tests, it was stated
that the models were maintained at a constant disk loading21 of 12 psf22, which
would imply that the rotational speed was varied with changing airspeed.
In edgewise flow, the total drag of the baseline shrouded rotor was almost
identically equal to the prediction for the momentum drag of the system, clearly
showing how large this component of the drag is, compared to the profile drag of
the shroud. Unlike the drag, which increased as would be expected with increasing
airspeed, the pitch moment data showed an increase in nose-up pitch moment up
20Conventional aft-mounted vanes generate a corrective, nose-down pitching moment by creating
a force that acts in the same direction as the vehicle drag, further exacerbating the situation.
21The authors seem to use the term ?disk loading? to mean the ratio Ttotal/A, not Trotor/A.
22?For comparison to actual flying vehicles, ducted-fan VTOL UAVs of comparable size have
disk loadings ranging from approximately 11 to 25 psf.? [34]
75
Figure 1.28: Auxiliary control devices tested by Fleming et al. [34].
76
to an airspeed of 17 ft/s, and then an unexpected decrease to nose-down pitch
moments. The authors surmised that this was possibly due to separation of the
flow from the windward side of the ?relatively sharp leading edge? (rlip was less
than 6%Dt). However, from the illustrations of the model, it seems that the shroud
profile was cambered outwards, rather than inwards, as is the norm for shrouded
rotor systems that are optimized for low-speed flight, and this may have also been
a factor in causing this phenomenon.
The vanes were tested at deflection angles from ?30? to +30?, and caused
the expected increases in drag and changes in pitch moment. At airspeeds above
15?20 kts (25?34 ft/s), though, the results due to different vane deflection angles
were nearly identical, indicating that the deflection of the jet (the rotor downwash)
by the freestream had increased the angle of attack of the flow at the vanes to large
enough values that they had stalled. This was confirmed by a CFD simulation of
the bending of the jet by the crosswind.
Of the various auxiliary control devices tested, the best performing were the
internal duct deflector with bleed, the flow control at the inlet by blowing, and the
internal duct vane, although flow control by blowing would probably not be a feasible
option on a MAV because of the high plenum air pressures that were required in
these tests to have any significant effect. In general, the various devices were able to
produce the desired nose-down pitch moments, although their effectiveness varied
at different airspeeds. At low speeds, the best performing devices augmented the
control authority of the primary control vanes by up to about 50%, while at higher
speeds, where the vanes had stalled, the auxiliary devices were able to provide
77
changes in pitch moment that the control vanes alone could not.
In 2004, Martin and Tung [117] tested two 10-inch-diameter shrouded-rotor
models (Fig. 1.29) in hover and in forward flight, at angles of attack from 0? (axial
flow) to 110? and airspeed ratios (?prime) up to 0.18 (41 knots, 70 ft/s). The two shrouds
had different inlet lip radii ? 1.7%Dt and 2.9%Dt ? as well as different camber and
thickness distributions. In the following discussion, these will be referred to as the
?LR1.7? and ?LR2.9? shrouds, respectively. Both shrouds were cambered outwards,
likethemodelstestedbyFlemingetal.[34, 94], and?wereconstantinternaldiameter
along the length, except for a very small region near the trailing edge,? i.e., they
had expansion ratios of approximately 1.0. As can be seen from the profile shapes
shown in Fig. 1.29, this approximation is more accurate for the LR2.9 model; ?d
for the LR1.7 model would have been slightly greater than 1.0. Both models were
tested with fixed-pitch propellers of varying diameters, to test tip clearances from
0.5% to 2%Dt.
In hover, the shrouded and open rotors were tested at tip speeds from 87 to
415 ft/s. The open rotor produced a thrust coefficient of 0.019, which was fairly
constant down to a tip speed of 214 ft/s, below which it decreased to ?0.017 at vtip
= 87 ft/s. The shrouded rotors also displayed a reduction in CT with decreasing tip
speed, but at a rate more severe than that for the open rotor, to the extent that for
tip speeds below about 130?170 ft/s, the thrust coefficients of the shrouded rotors
were lower than that of the open rotor. This was attributed by the authors ?either
to loss of the suction peak on the duct lip from laminar separation, or more likely
due to the viscous losses inside the duct.?
78
(a) Shroud profiles. Red = ?LR1.7?, blue = ?LR2.9?. (b) Test model.
Figure 1.29: Shrouded rotors tested by Martin and Tung [117].
Theshrouded-rotormodelwiththelarger, 2.9%Dt lipradius(?LR2.9?)achieved
a maximum thrust coefficient of 0.026 ? 37% higher than that of the open rotor ?
at the smallest tip gap of 0.5%Dt. When ?tip was increased to 2%Dt, the maximum
CT fell by 12% to 0.023, which was still 21% higher than that of the open rotor.
Decreasing the lip radius to 1.7%Dt caused CT to decrease by about 15%: from
0.026 to 0.022 at a tip gap of 0.5%Dt, and from 0.025 to 0.021 at a tip gap of 1%Dt.
With the sharper inlet-lip (?LR1.7?) model, thus, the highest thrust coefficients were,
respectively, 16% and 10% higher than that of the open rotor. A potential flow anal-
ysis showed that a decrease in inlet suction would have been the most likely reason
for the reductions in thrust due to increasing tip gap and decreased lip radius. On
the other hand, the LR1.7 model did display, experimentally, a slightly less-severe
reduction in thrust with decreasing tip speed than did the LR2.9 model.
As with the thrust coefficient, the increase in blade tip clearance also caused a
79
reduction in the figure of merit, as seen from the data given for the LR2.9 model at
the highest tip speed tested. The open rotor achieved a FM of 0.44. The reported
FM for the LR2.9 shrouded rotor was 0.67 at the smallest tip gap (0.5%Dt), which
was 52% higher than that of the open rotor. Doubling the tip gap to 1%Dt caused
FM to drop rapidly by 24% to 0.51, after which a further doubling of ?tip to 2%Dt
caused FM to decrease more slowly, by 7%, to 0.46. However, the authors used the
formula for an open rotor in calculating the figure of merit for the shrouded rotors
as well; hence, for a shrouded rotor with ?d ? 1.0, their reported values would be
40% higher than the actual values. The correct values of FM for the LR2.9 model
would therefore be 0.47 for the smallest tip gap (?0.5), which is only 7% higher than
that of the open rotor, and 0.33 at the largest tip gap (?2.0), which is 25% lower
than that of the open rotor.
Given FM and CT, the power coefficients for the open rotor and the LR2.9
shrouded rotor can be calculated, and the values of CT/CP obtained as follows: 4.52
for the open rotor, 5.91 for LR2.9 at the smallest tip gap (?0.5), which is 30% higher
than for the open rotor, a 23% decrease to 4.55 for ?1.0, and a further 5% decrease
to 4.26 for ?2.0, which is 6% lower than that for the open rotor. The variation in
CT/CP with changing tip clearance is thus similar to that of the figure of merit.
However, even these values are not entirely accurate, because even though a six-
component balance was used to measure the forces and moments on the models in
this investigation, the values of power used in the analysis by the authors were the
measurements of electrical power supplied to the motor, and so include electrical
and transmission losses in addition to the pure aerodynamic torque on the rotors.
80
Although the measurements were taken at the same rotational speed, the motor effi-
ciency also depends on the loading conditions, and is therefore different at different
rotor torques. Therefore, not only would the true values of FM and CT/CP have
been higher than those presented above, for both the open and shrouded rotors, but
their indicated variations also be questionable.
In forward flight, in the low angle-of-attack range (0?45?), the various shrouded
rotors exhibited very similar lift (CL-vs.-?) characteristics: an initial slope 75%
greater than that of the ?unpowered? shroud (annular wing) alone, a slight kink in
the curve at an angle of attack between 16? and 23?, and then a further increase
with almost the same slope before stalling at around ? = 40?. The unpowered
LR2.9 shroud, in comparison, stalled at ? = 14?, and had a maximum CL that
was 3.75 times lower than that of the powered model (LR2.9-?1.0). Decreasing
either the tip gap or the lip radius caused the kink in the curve to be postponed
to a slightly higher angle of attack. The pitch moment (CM) characteristics, on
the other hand, were quite different for the two shroud models. From axial flow (?
= 0?) up to an angle of attack of 12?, the two models had nearly identical pitch
moment coefficients. Beyond ? = 12?, CM for the sharper-inlet model (LR1.7)
remained roughly constant, exhibiting some scatter, while CM for the LR2.9 model
increased dramatically to values roughly ten times higher than those of LR1.7. At
? = 40?, coincident with the lift stall, both models displayed a sudden, sharp drop
in pitch moment. For the LR1.7 model, for which CM was already fairly low, this
resulted in negative, nose-down pitch moment between ? = 40? and 45?. These
preceding measurements were taken at an airspeed ratio of 0.15 (60 ft/s). At a
81
slightly lower airspeed ? ?prime = 0.1 (40 ft/s) ? and in the higher angle-of-attack
range (60??110?), the pitch moments were once again positive (nose-up) for both
models, with CM for the LR2.9 model being 37% higher than that for the LR1.7
model at angles of attack near 90? (edgewise flow). The lower pitch moments of the
sharper-inlet model was attributed to separation of the flow from the inlet, and this
is also suggested by the greater scatter in theCM data for this model, in comparison
with the much smoother variations in the data for the LR2.9 model. However, since
the two shroud profiles also had different thickness and camber distributions, in
addition to different leading-edge radii, it is quite possible that those factors also
affected theCM characteristics. The effect of changing airspeed on the pitch moment
was investigated by a sweep from 12 to 54 ft/s (?prime = 0.03?0.14) for the LR2.9-?1.0
model in edgewise flow (? = 90?). The CM data increased to a maximum value
at ?prime = 0.125 before starting to decrease again, although the ?decrease? may have
been an artifact of possible scatter in the measurements and the curve-fit that was
applied to them.
Following up on the results obtained in this investigation on the beneficial
effects of minimizing the blade tip clearance, Martin and Boxwell [124] tested two
other shrouded-rotor models that were designed to effectively eliminate the tip clear-
ance completely (Fig. 1.30). Both models were derived from the baseline, 10-inch
inner-diameter LR2.9 shroud of the earlier investigation, and were tested with a
10.35-inch-diameter propeller. In one shroud, the extended-diameter propeller fit
into a notch that was cut into the inner shroud surface at the propeller tip-path
plane. In the second shroud, a rearward-facing step was cut into the inner shroud
82
(a) Notched shroud.
(b) Stepped shroud.
Figure 1.30: Shrouded rotors tested by Martin and Boxwell [124].
profile just above the tip-path plane to accomodate the propeller. The potential-flow
analysis predicted increased inlet suction and a ?dramatically increased performance
increment? for the stepped shroud. However, the experimental (hover) thrust and
power data showed no difference between the performance of the baseline model and
the notched-shroud model, and degraded performance for the stepped shroud. The
failure of the analytic model was attributed by the authors to ?viscous dominated
flow in the area of the [blade] tip along the duct wall.?
During the same time period, Graf et al. [125] tested the effects of five different
lip geometries (Fig. 1.31) on the hover and forward-flight performance of a similarly-
scaled shrouded-rotor UAV model, the goal being to obtain ?a duct lip design that
83
provides good static thrust performance while also avoiding the generation of adverse
moments as much as possible.? Three lip geometries were extensively tested: a
?baseline1? lip with rlip ? 1.8%Dt and a shroud profile thickness of 12.5%cS; a
?step? lip with a rearward-facing step above the rotor plane, similar to that tested
by Martin and Boxwell [124], and with slightly increased shroud profile thickness
(15.8%cS) and lip radius (?2.2%Dt); and a ?baseline2? lip, which was a blending
of the designs of the other two shapes ? similar in shape to the stepped lip, but
without the step, and with reduced thickness (12.5%cS) and lip radius (?1.9%Dt).
The baseline2 lip was tested on a shroud model that was 15% larger than that for
the other models, and also had stator vanes of a different design. Two additional
lip shapes were initially also tested ? an elliptical lip, with a reduced lip radius
of ?0.8%Dt, and a circular-arc lip, with an increased lip radius of ?3.0%Dt ?
but these were not tested further because they produced less static thrust than the
other three models. The circular-arc lip produced slightly more thrust than the
elliptical lip, and the baseline1, baseline2 and step lips produced almost the same
amount of thrust, with the values for the baseline2 and step being only very slightly
higher than those of baseline1. The test set-up allowed for separate measurement
of the axial load on the shroud only, and it was found that, although their total
static thrust values were nearly the same, the baseline2 model achieved a higher
shroud thrust fraction (Tshroud/Ttotal) of about 0.53, compared to about 0.35 for the
baseline1 model.
In forward flight, the baseline1 and step-lip models were tested in edgewise
flow (? = 90?) at airspeeds from 10 to 40 ft/s. In these conditions, the step lip
84
Figure 1.31: Inlet lip shapes tested by Graf et al. [125]: (a) baseline1, (b) step, (c)
elliptical, and (d) circular arc. The left side of each lip corresponds to the interior
of the shroud. The last shape tested (baseline2) was not shown.
produced slightly higher lift, slightly lower drag, and significantly higher nose-up
pitch moments than the baseline1 lip. For both models, increasing the airspeed
caused lift to remain approximately constant or decrease very slightly, drag to in-
crease significantly, and the pitch moment to initially increase before starting to
level off at the higher airspeeds. For both models, the ram drag averaged 80% of
the total drag, with the ratio slowly decreasing as airspeed increased, indicating the
increasing contribution of the shroud profile drag to the total.23
The baseline2 model was tested at airspeeds from 30 to 85 ft/s, and at angles
of attack from axial flow (0?) to edgewise flow (90?). At all airspeeds, lift and drag
both increased with angle of attack, although the lift data did show a maximum
value near ? = 70?, beyond which it slowly decreased. With increasing airspeed,
the angle of attack at which this maximum occurred slowly decreased. Both lift and
drag increased slowly with increasing airspeed. On the other hand, the pitch moment
23The ram drag is proportional to the airspeed, while the profile drag scales with the square of
the airspeed.
85
was seen to decrease with increasing airspeed, showing nose-down (negative) values
at 85 ft/s, and to also exhibit increasing scatter in the data. At the lower airspeeds,
the pitch moment increased steadily with increasing angle of attack, while at the
highest airspeed it showed a gradually decreasing trend. Both of these phenomena
point to separation of the flow on the windward side of the inlet with increasing
airspeed.
1.4.3 Experimental work: Tests of a single shrouded-rotor model
This section lists experimental investigations in which a single shrouded-rotor
configuration was tested, without changing any of the shroud parameters described
above. Typically, these investigations were for the purpose of determining the effects
of changing the freestream velocity and angle of attack on a shrouded rotor. These
included the work of Kr?uger [43], Parlett [55], Moser [58], Kelly [132], Milla and
Blick [133], Abrego and Bulaga [134], the tests in support of the development of
the Doak VZ-4 aircraft by Parlett, Yaggy, Goodson, Mort, Grunwald, Kelley and
Champine [60?68], and of the Bell X-22A aircraft by Newsom, Maki, Giulianetti,
Mort, Spreemann, Biggers, and Gamse [69, 71?75].
In 2005, Sirohi et al. [23] investigated the effects of adding a shroud with an
expansion ratio (??d) of 1.14 and anti-torque stator vanes to a 6-inch-diameter rotor
(Fig. 1.32). The open rotor achieved its maximum figure of merit of 0.45 at a thrust
coefficient of 0.029. Testing the open and shrouded rotors at the corresponding
collective angle, and over a range of rotational speeds, resulted in increases in thrust
86
Figure 1.32: Shrouded rotor tested by Sirohi et al. [23].
over the open rotor, at the same electrical input power to the motor, of up to 35%, or
reductions in electrical power at the same thrust by up to 44%. More recently, tests
by Hrishikeshavan and Chopra [25] in 2008 on a 9.5-inch-diameter (24 cm) rotor,
with a maximum open-rotor figure of merit of 0.64, showed increases in thrust from
shrouding (??d = 1.0) of around 30% at the same mechanical input power.
1.4.4 Analytical methods for performance prediction
Analytical methods for predicting the performance of shrouded rotors have
ranged from the elementary momentum-theory, blade-element-momentum-theory
(BEMT) and blade-element-vortex-theory (BEVT) models to more sophisticated
potential-flowmodels(theclassical?methodofsingularities?[54]), sometimescoupled
with a separate boundary-layer code in order to provide some modeling capability for
viscous effects, and, in recent years, computational-fluid-dynamics (CFD) methods
using the Euler equations and various other simplifications of the full Navier-Stokes
equations. An excellent account of each of these approaches, along with comparisons
of their relative merits and demerits, was provided by Sacks and Burnell in 1962 [54]
87
? except for the CFD methods, which were developed after their time.
1.4.4.1 Blade-Element and Potential-flow methods
The seminal work in this area is considered to be that of K?uchemann and
Weber [44?51], who, in the 1950s, developed potential-flow solutions for annular
airfoils (?ring-shaped cowlings?) in a uniform flow, by modeling the bodies with
distributions of vortex rings over their surfaces, and also performed wind-tunnel
experiments with which to compare their predictions. In the 1960s, using the same
method, Kriebel [135] derived expressions for the force and pitching moment coef-
ficients of a ducted propeller in steady flight at an angle of attack as well as the
longitudinal stability derivatives in pitching motion, incorporating the effects of duct
length, thickness, camber and taper ratio, but modeling the propeller as a uniformly
loaded actuator disk. Kriebel and Mendenhall [116] developed methods for predict-
ing the forces and moments on a duct in terms of the propeller thrust, as well as the
duct surface pressure distributions and boundary-layer separation, and compared
their predictions with measurements made on the Bell X-22A and Doak VZ-4 air-
craft models. In 1970, Mendenhall and Spangler [136, 137] consolidated the earlier
work and presented a computer program for predicting the forces and moments on
a ducted fan at a specified advance ratio and angle of attack, as well as the radial
inflow variations and duct surface pressure distributions. The program algorithm
consisted of an iterative application of a blade-element-vortex-theory model and the
superposition of two solutions: ?an axial flow solution for the fan-duct-centerbody
88
combination and an angle of attack solution for the duct.? Sheehy [138] presented a
computer algorithm for the inverse task: that of designing the propeller and stator
vanes when the shroud geometry and performance requirements are specified. This
algorithm was later used in 1988 by Weir in designing the AROD UAV [35]. Ear-
lier, also in 1970, Gray and Wright [139] developed a vortex wake model for heavily
loaded ducted fans, in which the ?inner vortex sheets [shed from the blades] move
at a different rate from that of the boundary sheet [shed from the trailing edge of
the duct]?, thereby removing the restriction of light loading ? in which the two
sheets move downstream at the same rate ? inherent in previous analyses. A sim-
pler method, compared to that of modeling the shroud as chordwise distribution of
vortex rings, was suggested by McCormick in 1967 [140], and involved concentrating
the entire circulation of the shroud as a single vortex ring at the shroud quarter-
chord location. This formulation was used by Fairchild et al. in 1973 [53] to show
that even for a diffuser angle of 0?, in axial flow, increasing the shroud chord causes
greater reductions in ideal power compared to an open rotor.
An early attempt at modeling the effect of the finite tip clearance was per-
formed by Goodman [141], who, in 1956, formulated a correction factor for thrust
and efficiency measurements in wind-tunnel tests of propellers. This formulation
was essentially a ?re-working [of] the Prandtl solution ... [which accounts] for a
finite number of blades by considering the radial flow near the blade tips ... taking
into account the presence of a nearby wall.? As the size of the gap approaches in-
finity, i.e., the case of an open rotor, the Goodman correction factor reduces to the
standard Prantl tip-loss correction factor for an open rotor [14, pp.102?105]. The
89
analysis shows that ?the circulation at any [blade] station is larger when there is
a wall present than when there is not. ... Because of the constraint of the wall,
less air is able to flow around the propeller disc and more air must flow through
it.? The gain in thrust is achieved at the cost of a loss in efficiency, however, if a
blade design is used that has been optimized for use on an open rotor, indicating
that blades designed for shrouded rotors should have a different twist schedule than
those designed for open rotors.
In 1974, Gibson [142] conducted theoretical studies of the effects of the tip
clearance and of radial variations in blade loading on the performance of shrouded
rotors, with the shroud modeled using continuous surface distributions of ring vortic-
ity. Decreasing the tip clearance from 5% to 1%Dt showed the expected increases in
thrust and efficiency. However, below 1%Dt, further reducing the clearance caused
the predicted thrust and efficiency to decrease; this was attributed to failure of the
inviscid flow model to accurately model the interaction between the blade tips and
the shroud boundary layer. The highest thrusts and efficiencies were also predicted
with uniform and with tip-biased loading distributions.
In designing the fan-in-fin for the Sikorksy S-67 helicopter in the early 1970s,
Clark [84] used a potential-flow method originally developed for fan-in-wing ap-
plications, combined with a boundary-layer analysis to check for the existence of
separation. The potential-flow method involved a superposition of three basic solu-
tions ? those of (1) a ?closed? or blocked duct in axial flow, (2) an ?open? duct in
axial flow, and (3) an open duct in cross (edgewise) flow ? to simulate the flow at
any thrust level and angle of attack. The predictions of power consumption, sur-
90
face pressure distributions and inflow velocity distributions from this model agreed
extremely well with experimental measurements.
In 1991, Wright et al. [86] described a model ?based on a momentum ap-
proach? used in the design of the H-76 FANTAIL, which seems to essentially consist
of a fixed augmentation factor (Ttotal/Trotor, set equal to 2.0) multiplying the pre-
dictions of the GenHel code for rotor thrust as a function of collective. The same
year, DeCampos [143] gave a review of the analytical methods used in the design
of marine ducted propellers, and presented a three-dimensional theoretical analysis
of the steady interaction between a ducted propeller and a radially and circumfer-
entially sheared axial inflow. In 1996, Page [144] presented a blade-element-based
algorithm for designing and analyzing contra-rotating or single-rotor/single-stator
ducted fans. However, the algorithm does not account for the effects of the blade
tip gap, requiring an empirical correction factor, nor does it include any predictions
of the force (thrust) on the duct itself.
In 1998, as part of the aeroservoelastic model developed for handling qualities
analysis of the RAH-66 Comanche, Kothmann and Ingle [88] presented an ?analytic
model for the dynamic thrust response of a ducted rotor, which captures the effects
of high-speed edgewise flight and high-frequency blade pitch inputs?. The model is
of the BEMT-type, developed for axial flow as well as flow at an angle of attack,
and with additional terms to account for unsteady inflow effects. The formulation
of the model is for the special case of a constant area duct (?d = 1.0), and so
the expansion ratio does not appear explicitly in it. This correction, as well as
other improvements and enhancements, were made in a similar model developed
91
by Basset and Brocard [33] in 2004 for the flight dynamics analysis of fenestron-
equipped helicopters like the Dauphin.
In 2000, Bourtsev and Selemenev [92] presented a computational method de-
veloped at the Kamov company during the development of the fan-in-fin of the Ka-60
helicopter. Their method used momentum theory for modeling the global effect of
the shroud on the system, and blade-element-vortex theory for calculation of the
blade loads. Empirical correction factors were used for the pressure losses due to
flow through the shroud, as well as for the effects of the shroud geometric parameters
? lip radius, blade tip clearance, diffuser angle, diffuser length ? in both positive
and reverse thrust conditions. The effect of the blade tip gap on the blade loads is
accounted for via a ?Prandtl-Shaidakov factor?, which, like the Goodman correction
factor [141], reduces to the regular Prandtl tip-loss correction factor for an open
rotor when the tip ?gap? becomes infinitely large. Extremely good agreement was
achieved between the predictions of this method and experimental test data.
In 2003, Guerrero et al. [145] developed a code called AVID-OAV for ?the
design and analysis of ducted-fan VTOL UAVs?. In this model, the shroud and ro-
tor aerodynamics are considered separately, and their solutions superimposed. The
shroud is modeled as an annular airfoil, and its characteristics in a freestream flow
derived by interpolation from the wind-tunnel measurements made by Fletcher in
1957[26]. Therotorismodeledaseitheranactuatordisk, orwiththrust/power/efficiency
data from a ?propulsive performance map? that is obtained using a separate blade-
element-vortex analysis. As in the case of Wright et al. [86], it is unclear, however,
how the influence of the shroud on the performance of the rotor is modeled, espe-
92
cially since even the expression presented in this reference for the induced velocity
in the shroud is actually the special-case formula for an open rotor, with no mention
of the expansion-ratio factor. Nonetheless, the predictions for the forward flight
performance of the Allied Aerospace 29-inch iSTAR vehicle match fairly well with
the wind-tunnel data for that aircraft.
In 2004, Quackenbush et al. [146] used the CHARM (Comprehensive Hierar-
chical Aeromechanics Rotorcraft Model) code, which uses a free-wake, potential-flow
model for the rotor and lifting panel model for fuselages and other bodies, to analyze
shrouded-rotor configurations for UAV applications. Good agreement was achieved
between their predictions and previously published experimental data on both an-
nular airfoils (shroud-only) [26] and the X-22A shrouded propellers [75], but only in
the pre-stall region. For capturing viscous effects and modeling post-stall behavior,
the RSA3D code was used, which showed varying agreement with the experimental
data.
1.4.4.2 Computational-Fluid-Dynamics methods
Blade element and momentum methods are useful for approximate compar-
isons with experiments and for predicting upper limits of performance, but cannot
represent the geometric characteristics of the shroud, unless empirical correction
factors are used. Potential flow methods can address this issue, but because of their
inherent assumptions ? inviscid, irrotational and incompressible flow ? cannot
predict friction drag and stall characteristics [147]. For these reasons, and with the
93
advances in computing power, CFD-based approaches have assumed a prominent
role in the design and analysis of shrouded rotors.
At A?erospatiale [93], early performance estimations and initial sizing of fen-
estrons were obtained by BEMT methods, while detailed analysis of the effects of
the shape of the shroud was conducted using a code called METRAFLU which
combined CFD with table look-ups for blade airfoil characteristics. This reference
shows comparisons of the predictions of this code and those of the BEMT method
with fenestron test measurements of thrust, power, inflow velocity distributions and
shroud surface pressure distributions.
In 1989, Rajagopalan and Zhaoxing [147] performed a CFD analysis to com-
pare the performances and flow-fields of a shrouded and an open propeller in axial
flow, in which the propeller was modeled by time-averaged momentum sources.
Their code was also used later in the detailed design of the RAH-66 Comanche?s
FANTAIL [87]. A coupled CFD analysis of the entire Comanche fuselage, with the
FANTAIL rotor modeled either as an actuator disk or with a blade-element model,
was described in 2003 by Alpman, Long and Kothmann [89?91, 148]. An improved
momentum source model, coupled with the OVERFLOW-D mesh solver, was used
by Nygaard et al. [120] in 2004 for modeling the FANTAIL. This reference provides
detailed comparisons with experimental data and with a more accurate, but also
more computationally-intensive, discrete-blade CFD model used earlier by Ruzicka,
Strawn and Meadowcroft [149, 150]. The analysis shows how reducing the blade tip
clearance reduces the leakage flow and the size of the reversed-flow zone at the shroud
wall, resulting in increased system thrust and figure of merit, and also demonstrates
94
the sensitivity of the predicted shroud thrust to accurate modeling of the effect of
the blade tip clearance. In this region, the flow is three-dimensional, and the appli-
cability of 2-D airfoil tables ?becomes questionable.? Both the momentum-source
and discrete-blade models show the flow separating at the shroud wall slightly up-
stream of the rotor, due to the tip vortices lifting the flow away from the wall, and
the prevention of the flow from re-attaching downstream of the rotor due to the
adverse pressure gradient in the diffuser. The importance of realistic 3-D tip mod-
eling was also demonstrated in 2003 by Lee, Kwon and Joo [121], who performed a
CFD analysis of the Kamov Ka-60 fan-in-fin, and showed closer agreement with the
experimental results than did the analytical model of Bourtsev and Selemenev [92]
described earlier.
In 2004, Ahn and Lee [115, 151] used a CFD model of a representative fan-
in-fin tail rotor in axial flow to parametrically investigate the effects of varying
the inlet lip radius, diffuser angle and rotor radial ?strength? distribution. The
CFD methodology was an extension of the ISES code developed in the 1980s by
Drela and Giles [152], and the rotor was modeled as an axisymmetric actuator disk,
but with radially varying strength distributions. At fixed thrust and freestream
velocity (sideways flight for the helicopter), increasing ?d from 8? to 12? caused a
1% reduction in power; increasing rlip from 4.75%Dt to 7.5%Dt caused the power
required to decrease by less than 0.1%; and changing the rotor strength distribution
from a uniform one to one that has increasing strength towards the blade tips ?
corresponding to less blade twist and greater blade pitch angles at the tip than
would be obtained with an ideal twist distribution ? caused a 7% reduction in
95
power, ?by facilitating the expansion of the fan duct flow,? and also less sensitivity
to losses due to the blade tip gap.
Aeroelastic stability and blade-flutter analyses for ducted fans, as used in
aircraft engine and turbomachinery applications, were developed in the late 1990s
by Srivastava et al. [153], Keith and Srivastava [154] and Srivastava and Reddy [155].
In these analyses, which used the unsteady Euler equations for modeling the flow,
only the blades were considered elastic, while the hub and the duct were considered
rigid. In that same time period, Elena and Schiestel [156] and Randriamampianina
et al. [157] presented models for the prediction of transitional and turbulent flow
inside shrouded rotor-stator systems, taking into account the effects of rotation on
turbulence.
1.4.5 Other shrouded-rotor research
This section briefly describes some of the work that has been performed on
aspects of shrouded-rotor operation that were not addressed in this investigation,
but are of significant importance in the design of MAVs of this configuration: noise
considerations, detailed analysis of the physics of tip-clearance flow, aircraft stability
and control considerations, and the aerodynamics of annular wings.
1.4.5.1 Noise considerations
In 1950, Hubbard [119] took sound measurements on five shrouded propellers
in static conditions, as well as on the unshrouded (open) propeller. The maximum
96
total sound pressure produced by the shrouded propellers varied from approximately
one-half to two times as much as that for the open propeller, depending on whether
the flow at the shroud surface was attached or separated, respectively, at the inner
surface of the shroud inlet. The models were tested on an outdoor setup, and
although nominally static conditions were tested, it was found that crosswinds and
tailwinds caused the flow to separate more easily, whereas headwinds helped the
flow stay attached. Low propeller rotational speeds also led to the flow separating.
In attached flow conditions, noise levels increased with increasing tip speed, just
as for an unshrouded propeller, and with increasing blade tip clearance. Increasing
the number of blades decreased the noise levels, also as for an unshrouded propeller,
because of the increase in the fundamental frequency (at the same rotational speed).
Increasing the length of the shrouds had no significant effect on the noise levels.
In 1971, Drischler [158] performed analytic studies of sound pressures inside
the shroud of a shrouded propeller in axial flight, by modeling the propeller disk as a
distribution of rotating acoustic pressure doublets, and the shroud and centerbody
as infinite coaxial circular cylinders. The purpose of this study was ?to develop
criteria for light-weight structures of sufficient acoustic-fatigue life, particularly in
the critical region near the propeller disk.? The effects of varying the propeller
loading, loading distribution, tip clearance, tip speed, freestream velocity and hub-
tip ratio were analyzed. The analysis showed that for a fixed torque and total thrust,
increasing the freestream Mach number or decreasing the tip gap by ?only a few
percent? caused up to 6-dB increases in sound pressures in the vicinity of the blade
tips, but ?insignificant variations at distances greater than about one duct radius
97
from the propeller disk,? while ?increasing the hub-tip ratio from 0.2 to 0.5 results
in an increase in sound-pressure level of approximately 2 dB to 3 dB throughout
the duct.? The increase in sound pressures with decreasing tip gap agreed with the
earlier experimental results of Fricke et al. [159], who had tested gap sizes of 0.375
and 1.5 inches on a modified, full-scale model of the X-22A shrouded propeller.
At A?erospatiale, tests on fenestron tail rotors [29] in the mid-1980s showed
increases in noise levels with increasing blade collective, and also faster attenuation
of noise with distance compared to that for a conventional tail rotor. This occurred
because the fundamental frequencies of the fenestron noise were higher by approx-
imately an entire order of magnitude, and was due partly to the higher number of
blades in the fenestron, and partly to the interaction between the rotating blades
and the non-rotating support struts. This blade-strut interaction was found to be
the main origin of fan-in-fin noise. An increase in the spacing between the rotor and
the support struts by 6 cm (approximately 5.5%Dt) caused a rapid reduction in the
noise levels ? by up to 10 dB for the 40?-collective case ? to the same level as the
case with no support struts at all. The masking effect of the shroud itself resulted
in a 5?6-dB reduction in in-plane noise. Reference [29] also stated that ?a simple
theoretical model has been established which provides good acccount of the general
trend of fan-in-fin rotational noise evolution with rotor speed, blade pitch setting,
blade/support spacing and support shapes,? but provided no further information
about this model.
Riley [81] and Andrews et al. [80] reported on tests, conducted at Bell Heli-
copter Textron in the mid-1990s, of the effects of uneven blade spacing on shrouded
98
tail rotor acoustics. Tests were conducted on a 4-bladed evenly-spaced rotor, two
4-bladed scissors rotors with 70?110? and 55?125? spacings, a 5-bladed evenly-spaced
rotor, and three 5-bladed rotors with different amounts of sinusoidally-modulated
spacings. This last concept ?re-distributes the energy of a single tone over a number
of discrete frequencies, ... reducing the amplitude of each blade passage frequency
harmonic and producing a more broadband-like spectrum,? which reduces the per-
ceived annoyance, and had been previously used with success in radiator fan designs
and by Eurocopter in the design of the EC135?s fenestron. Also tested were different
rotor solidities, airfoils, tip shapes, tip speeds, and ?a number of duct treatments and
modifications which included inflow turburlence, simulated driveshaft and support
struts, two spinners and two centerbody fairings, uneven rotor-to-duct gap, unequal
blade collective pitch, and reduced duct diffuser exit angle.? The quietest config-
uration ? one of the 5-bladed rotors with sinusoidally-modulated blade spacing,
reduced tip speed and increased solidity, and an aft-tapered blade tip ? achieved
an 11.3 dBA reduction in harmonic noise compared with the baseline rotor. Of this
reduction, the sinusoidal spacing accounted for 5.4 dBA, the aft-tapered blade tip
for 0.3 dBA, and the reduced tip speed and increased solidity together for 5.6 dBA.
Similarly, the use of unequal blade spacing resulted in a reduction in sound pressure
levels by approximately 4.5 dB, compared to an equally-spaced blade arrangement,
in tests at the Japan Defense Agency?s Technical Research and Development In-
stitute (TRDI); this led to the adoption of a 55?/35? blade-spacing scheme for the
XOH-1 helicopter?s ducted tail rotor [32].
In 2004, Martin and Boxwell [124] measured the noise produced by a 10-inch-
99
diameter shrouded-rotor UAV model. Adding the shroud to the open propeller
did not change the ?rotational noise? ? the harmonics of blade passage frequency,
which are strongest in the plane of the rotor ? but increased the level of the ?broad-
band noise?, which is strongest in the axial direction (called ?vortex noise? by Hub-
bard [119]). Increasing the blade tip clearance from 0.5% to 2.0%Dt caused no
measurable difference in the acoustic signatures. However, separated flow caused by
sharper inlet lip radius, and by a rearward-facing step above the rotor disk in one
of the models, caused broadband noise levels to increase.
1.4.5.2 Tip-gap flow physics
In 1966, Gearhart [160] reported on a study of the mechanics of tip clearance
flow in turbomachinery in an effort to reduce tip clearance cavitation losses in marine
shrouded propulsors. The study characterized the flow patterns associated with
variously-shaped gap configurations, and experimentally examined the effects on the
tip clearance flow due to variations in the size of the clearance, the wall boundary
layer and the blade tip loading.
More recently, in 2005, Oweis and Ceccio [161] used particle image velocimetry
(PIV) to examine the flow-field in the tip region of a marine ducted propulsor,
and revealed significant vortical phenomena in the instantaneous flow-fields that
get masked in the more-common time-averaged measurements of such flows. In
further investigations, both PIV and LDV (laser-doppler velocimetry) were used
to determine how the tip-leakage flow varied with Reynolds number [162], and to
100
compare this flow with the tip vortices formed on an unducted (open) rotor [163].
The presence of the duct was found to reduce the strength of the primary tip vortex,
relative to the other, secondary vortices in the tip-flow region, but the radius of its
core did not vary substantially between the ducted and open cases. The strength
and core size were only weakly dependent on the Reynolds number ? within the
range tested in this study (0.7?106?9.2?106) ? and the core size was found to
be on the same order as the tip clearance (0.26 in, 0.77%Dt).
1.4.5.3 Shrouded-rotor UAV stability and control
A number of investigations have addressed the topic of the stability and con-
trol of shrouded-rotor UAVs. In 1988, Weir [35] presented a sizing analysis for
traditional, aft-mounted control vanes, compared the relative merits of split vanes
(flapped stators) versus all-moving vanes, and presented experimental data showing
the effectivness of the vanes in hover. In 2003, Fleming et al. [34, 94], as described
earlier, tested a UAV model with control vanes as well as with seven other auxiliary
control devices, and presented data comparing their effectiveness in hover and in
crosswind conditions. In 2004, Sirohi et al. [23] tested flat-plate and circular-arc
vanes on a 6-inch-diameter rotor, in both the shrouded and unshrouded conditions,
and found that the vanes produced more anti-torque moment when the shroud was
present than when it was not. As would be expected, the curved vanes produced
more anti-torque than the flat ones, and were able to completely neutralize the
torque produced by the rotor. The power penalty due to the need to overcome the
101
download on the vanes was approximately 10% for both the open and shrouded
rotors.
In 2001, as an alternative to the mechanical complexity of cyclic rotor control
mechanisms, Kondor and Heiges, Kondor et al., and Fung and Amitay [126?128]
advocated using active flow control devices for vehicle control, citing mechanical
simplicity, a minimum of moving parts and a potential reduction in weight as ad-
vantages of this approach. In tests on an 11-inch-diameter model, combinations of
tangential and normal blowing on the inner diffuser wall, near the exit, resulted
in the ability to generate pure sideforces in hover, without any pitching/rolling
moment [126]. Such pneumatic circulation control requires an on-board source of
compressed air, which is not feasible for a small UAV. Therefore, the use of synthetic
jets, which require only small amounts of fluid momentum flux for their operation,
was investigated, with a series of such devices embedded in the fixed stator vanes of
the model [127, 128] . The operation of these devices is binary in nature, causing
the flow to be completely attached or completely separated, resulting not only in
a much simpler on/off control scheme, but also a means for avoiding the ?unpre-
dictable flow separation and attachment tendencies of thick laminar boundary layers
at low Reynolds numbers.? With these devices, successful yaw control was achieved,
but significant coupling occurred between the pitch and roll control moments.
Finally, mathematical models of the dynamics of shrouded-rotor UAVs have
been developed in recent years by Avanzini et al. [164?166] and Salluce [167].
102
1.4.5.4 Behavior of annular wings
The aerodynamic characteristics of annular wings (?unpowered shrouds?) have
been studied by Fletcher [26], who measured the lift, drag and pitch moment on
five different annular airfoils, and Lipera [27], who experimentally compared the lift
characteristics of powered and unpowered shrouds.
1.5 Low-Reynolds-number rotor aerodynamics
While this investigation has focused on the design of the shroud for a shrouded-
rotor MAV, an awareness of the aerodynamics of the rotor itself in this Reynolds-
number regime is of obvious importance. The influence on MAV performance of
rotor parameters such as blade aspect ratio, twist and taper distributions, tip
shape, airfoil shape, solidity, number of blades, collective angle and rotational
speed have been investigated by, among others, Samuel et al. [17], Bohorquez and
Pines [12, 16, 19, 22, 168, 169], Hein and Chopra [20, 170] and Ramasamy et
al. [24]. Flow-visualization studies using particle image velocimetry (PIV) have also
been conducted by Ramasamy et al. [24, 171] with a view to obtaining a better
understanding of the physics of the flow around rotating wings at these small scales.
2-D and 3-D CFD studies of airfoils at low Reynolds numbers, using the OVER-
FLOW and TURNS codes with a low-Mach pre-conditioner, have been conducted
by Schroeder and Baeder [172, 173].
103
1.6 Objectives and approach of current research
As can be seen from the preceding survey of the literature on the subject of
shrouded rotors, a considerable amount of work, both experimental and theoretical,
has been performed in this area. However, the data that are available from these
investigations are predominantly from tests of rotors with diameters from one to
seven feet, typically operating at blade tip Reynolds numbers on the order of 105
to 106 and higher. Rotary-wing MAVs, on the other hand, operate in the laminar
flow regime, with tip Reynolds numbers on the order of 104, where viscous effects
become dominant and the flow separates much more readily in the face of adverse
pressure gradients. Traditional design philosophies ? rules-of-thumb for the optimal
values of different shroud parameters ? based on experience with larger-scale ducted
propellers, while instructive in terms of indicating possible starting points for design,
may or may not apply to MAVs as well. Although this flight regime has, in recent
years, become a burgeoning area of aerodynamics research, and some testing has
been done on smaller, near-MAV-scale models, yet, the data are not comprehensive,
and compared to the collective experience with aerodynamic phenomena at both
much higher (and also much lower) Reynolds numbers, there is still much that
is not known about the physics of the flow at this particular scale. In many of
the research programs conducted, at both MAV- and larger scales, vital pieces of
information are lacking, such as the power consumption of the shrouded rotors,
or direct comparisons between the shrouded and unshrouded (?open?) versions of
the same rotor, or the effects of rotor tip speed and blade pitch angle. In many
104
cases, more than one parameter value was changed between the different models
tested, and so it is not possible to say with certainty which particular parameter
was primarly responsible for the differences in the test results obtained. Even in the
cases where strictly a single parameter value was changed, the results obtained varied
between the different test programs conducted. This indicates that the changes in
the performance characteristics also depended on the values of the other, unchanged
parameters ? which includes the size of the model itself, as well. And finally,
predictive, computational tools, while immensely useful in reducing the amount of
experimental testing that must be performed in designing an aircraft, are themselves
hampered by difficulties in modeling flow separation phenomena, which is of critical
importance in this low-Re regime.
All of these factors point to the need for an extensive, systematic investiga-
tion, at low Reynolds numbers, of the effects of varying the different shrouded-rotor
design parameters on the hover and translational-flight characteristics of aircraft of
this configuration. Such an investigation has been carried out, and is described in the
following chapters of this dissertation. The goal of this research program was two-
fold: first, to investigate the improvements in performance that are achievable from
using the shrouded-rotor configuration for micro air vehicles; and second, to acquire
a body of experimental data that could be used for the validation of analytical, pre-
dictive tools, and as guidelines in the design of more efficient, more capable MAVs.
Chapter 2 describes the experimental setup used in conducting this investigation,
and gives the details of the various shrouded-rotor models tested. Chapter 3 presents
the results obtained from the tests of these models in the hover condition, wherein
105
measurements were made of the performance variables (thrust produced, power con-
sumed) as well as of the pressure distributions over the shroud surfaces and the axial
velocity distributions in the wakes of the models. Chapter 4 presents the results ob-
tained from the wind-tunnel tests of the open rotor and one of the shrouded-rotor
models, at angles of attack from axial flow to edgewise flow, and at various freestream
velocities. In those tests, measurements were made of thrust, power, pitch moment,
normal force, and shroud surface pressure distributions. Chapter 5 presents trade
studies, based on the results obtained in this research, to determine whether it is
worthwhile to choose the shrouded-rotor configuration for the design of a hover-
capable micro air vehicle. Chapter 6 summarizes the work and presents concluding
remarks and suggestions for future work. Appendix A contains detailed derivations
of the momentum-theory and combined blade-element-momentum-theory models
for a shrouded rotor, and Appendix B presents a discussion of the different mea-
sures used to describe the efficiency of hovering rotors, including reasons why the
traditional formulation of the figure of merit is unsuitable for shrouded rotors, and
proposes a new, generalized form of the figure of merit that overcomes this limitation
and is applicable to both open and shrouded rotors.
106
Chapter 2
Experimental Setup
2.1 Shrouded-rotor models
For this test program, the parameters chosen for investigation were those that
were found, from the literature on previous research, to have the greatest influence on
shrouded-rotor performance. These were ? see Fig. 2.1 ? the blade tip clearance
(?tip), the inlet lip radius (rlip), the diffuser angle (?d), the diffuser length (Ld)
and the diffuser expansion ratio (?d). In cross-section, the profile shape chosen for
the shrouds consisted of a simple semi-circular inlet joined to a conical diffuser.1
While advanced shroud designs would have profiles with smoothly and possibly
non-linearly varying camber and thickness distributions, this elementary profile was
chosen for this investigation because it involved a finite number of discrete and easily
changeable parameters, yet would offer universal insight into the behavior of most
shroud shapes.
The nominal inner, or ?throat?, diameter of the shrouds (Dt) was chosen to
be 16.0 cm (6.3 in), to match the diameters of the MAV-scale rotors being tested
in other investigations at the Alfred Gessow Rotorcraft Center, which ranged from
15.2 cm to 17.4 cm [12, 16, 17, 19, 20, 22, 23, 168?170], to facilitate comparisons with
the results of these other research efforts. The outer dimensions of the shrouds, upon
1For these shrouds, therefore, the shroud chord cS = rlip + Ld.
107
Figure 2.1: Principal shroud parameters affecting shrouded-rotor performance:
diffuser included angle (?d), diffuser length (Ld), inlet lip radius (rlip), blade tip
clearance (?tip)
which depend the maximum values of lip radius, diffuser length and diffuser angle,
were constrained by the size of the equipment used to manufacture the shrouds
and by the available sizes of the raw material from which the shrouds were made.
Since a large number of differently-shaped shrouds were to be built and tested, the
fabrication of the shrouds from composite materials was not an option, as this would
require the manufacture of multiple tooling molds and was therefore prohibitive
from time and cost standpoints. Instead, a machineable, low-density, polyurethane
prototyping foam2 was used. The shrouds (Figs. 2.2, 2.3) were machined on a lathe,
coated with a latex-based filler to seal their surfaces, and then coated with several
layers of spray-on polycrylic and extensively sanded to a smooth finish. Attachment
to the rotor centerbody structure was accomplished by two sets of four support
struts and a center support bracket. The struts were brass tubes with a diameter
2LAST-A-FOAMR? FR-6704, produced by the General Plastics Manufacturing Company
(http://www.generalplastics.com/).
108
Figure 2.2: Shrouded-rotor models
of 3/32 in (2.34 mm), and were attached to the shroud by delrin inserts placed in
the shroud wall. The inserts had a collar which rested against the external shroud
surface, and were held in place by the tension between diametrically opposite struts.
Set-screws were used to make the majority of the connections between the various
components of the models.
The objective of this study was to systematically investigate each of the
shrouded-rotor parameters, along with their mutual ?interference? or coupling ef-
fects on each other, instead of focusing on a single parameter. Consequently, in
order to limit the test matrix to a manageable size, only a few, broadly-spaced val-
ues of each parameter could be tested. Three shrouds were fabricated, each with
a different inlet lip radius and a different diffuser angle. By separating the inlet
109
Figure 2.3: Close-up view of shrouded-rotor model LR13-D20
and diffuser sections, and ?mixing and matching? them, nine different shroud mod-
els were obtained. The values chosen for the lip radii and diffuser included angles
were 6%Dt, 9%Dt and 13%Dt, and 0?, 10? and 20?, respectively. These values were
chosen to span the ranges tested by previous investigations of shrouded rotors, and
subject to the constraints described above. The values of lip radius were biased to
the upper end of the range, based on the reasoning that rotary-wing MAVs would be
optimized more for efficient hover performance than for high cruise speeds. Owing to
variability in the manufacturing process, the actual dimensions of the finished mod-
els were Dt = 15.9 cm, rlip = 6.48%Dt, 8.99%Dt and 13.1%Dt (1.03 cm, 1.43 cm
and 2.08 cm, respectively), and ?d = 0?, 9.70? and 20.22?. All nine shrouds had
diffuser sections with a length of 72%Dt (11.5 cm, 4.5 in). To test the effect of
110
varying diffuser length, the 10-degree diffuser was cut into three sections to obtain
shorter diffusers with lengths of 31% and 50%Dt, and these were tested with the
13%Dt-lip-radius inlet. The final parameter, blade tip clearance, was tested by using
rotors of different diameters to obtain clearances of 0.1%?1.6%Dt (0.15?2.50 mm,
0.006?0.100 in). The reductions in the rotor radius ? by up to 2.85 mm (3.6%R) ?
were small enough that there was insignificant difference between the performances
of the (open) rotors.
Thus, inadditiontotheopenrotoritself, atotalofseventeendifferentshrouded-
rotor models were tested, using eleven different shroud profiles. For ease of reference
in the following discussions, the models are identified with labels of the form ?LRx-
Dy-?z-Lw?, where ?x? is the nominal inlet lip radius, expressed as a percentage of the
shroud throat diameter (Dt), ?y? is the nominal diffuser included angle, expressed
in degrees, ?z? is the blade tip clearance, expressed in %Dt, and ?w? is the diffuser
length, also in units of %Dt. An absence of the ?L? suffix in the model name in-
dicates that the shroud had the default diffuser length of 72%Dt. The complete
matrix of test models is shown in Table 2.1. Table 2.2 lists, for each of the diffuser
profiles tested, the values of the corresponding expansion ratio and the values pre-
dicted by momentum theory for Trotor/Ttotal, TSR/TOR and Pi(SR)/Pi(OR), calculated
using Equations 1.5 (A.17), A.27 and A.28, after correction for the blockage by the
rotor hub. This matrix of models enabled comparison of data from multiple series
of models for each of the shroud parameters. For example, for testing the effect
of changing lip radius, data were available from a series of models with a diffuser
angle of 0?, from another series of models with ?d = 10?, and yet another with ?d =
111
Table 2.1: Matrix of shrouded-rotor models tested in hover, indicating models for
which surface-pressure measurements (P) and wake measurements (W) were also
made.
Lip Radius Diffuser Included Angle
D00 D10 D20
Models with the default diffuser length (L72)
LR06 ?0.6 (P) ?0.1 ?0.1 (P)
LR09 ?0.6 (P) ?0.6 ?0.1 (P)
?0.5
LR13 ?0.6 ?0.1 (P,W) ?0.1 (P)
?0.6 (P,W) ?0.6 (P)
?0.8 (P)
Models with reduced diffuser lengths
LR13 ?0.1-L50 (P,W)
?0.1-L31 (P,W)
?0.8-L31 (P,W)
?1.6-L31 (P,W)
20?. In each series, all variables are held constant except for the shroud parameter
of interest. However, as mentioned previously, because the values of these other
parameters do change from one series to the other, their interference or coupling
effects cause the effects of the parameter of interest to be seen to different degrees
in each of the different series.
Of the seventeen models, thirteen were instrumented with pressure taps ?
up to nine over the inlet lip and up to nine over the diffuser inner surface ? to
measure the pressure variation along the shroud surface. The tubing used for the
112
Table 2.2: Momentum theory predictions
Diffuser Expansion Expansion ratio, Trotor/ TSR/TOR Pi,SR/Pi,OR
model ratio (??d) corrected (?d) Ttotal (at same A, Pi) (at same A, T)
D00 1.000 1.033 0.48 1.27 0.70
D10-L31 1.113 1.150 0.43 1.32 0.66
D10-L50 1.184 1.224 0.41 1.35 0.64
D10 1.269 1.311 0.38 1.38 0.62
D20 1.575 1.628 0.31 1.48 0.55
pressure taps had an inner diameter of 0.0650 in (AWG16) and an outer diameter
of 0.0875 in. The taps were staggered in azimuth, i.e., around the circumference of
the shroud, over a range of approximately 90?, to prevent contamination of the flow
at each tap by the presence of the upstream taps. This azimuthal distribution was
possible because of the axial symmetry of the flow-field for a hovering rotor. In the
?axial? direction, the taps were concentrated more closely together in regions of the
greatest change in pressure: on the inner portion of the inlet surface and near the
blade passage region.
For the translational-flight tests, only the open rotor and a single shrouded-
rotor model were tested ? again, due to the prohibitive size of the test matrix
that would have otherwise resulted. This model was one of those that was found to
exhibit the best performance in hover, and was the model with an inlet lip radius of
13%Dt, a blade tip clearance of 0.1%Dt, a diffuser angle of 10? and a diffuser length
of 31%Dt (?LR13-D10-?0.1-L31?). A model with the shortest diffuser length was
chosen so as to minimize the weight and inertia of the model, to prevent overloading
113
of the wind-tunnel balance.
The rotor used for all the models was a three-bladed, variable-collective,
rigid (hingeless) rotor with untwisted, flat-plate, rectangular, aluminum blades, and
driven by a 6-volt/8.5-watt electric motor3 attached to the rotor support struc-
ture . The values of the important rotor parameters are given in Table 2.3. This
blade/airfoil geometry was chosen on the basis of its simplicity, since multiple rotors
had to be fabricated to create different tip clearance values, and the emphasis of this
research was on investigating the effects of different shroud geometries rather than
optimizing the rotor itself. However, in order to test the effects of shrouding on a
more-efficient rotor that is better suited to the low-Reynolds number flight regime,
a single hover test was also conducted of the LR13-D10-L31 shroud with a second
rotor with slightly higher solidity (c = 1.27 cm, ? = 0.15) and cambered, circular-
arc airfoils (maximum camber = 13.4%c). The other parameters of this rotor were
the same as those of the baseline rotor, as given in Table 2.3. The blade tip clear-
ance between this ?cambered? rotor and the LR13-D10-L31 shroud was 0.45 mm, or
0.28%Dt.
2.2 Hover test setup
The setup for the hover tests is shown in Fig. 2.4. The tests were conducted
with a custom-built test-stand, specially designed to measure the low values of thrust
3WES-Technik Model Micro DC 1331 (DC6-8.5) (http://www.wes-
technik.de/English/motors.htm).
114
Table 2.3: Rotor parameters
Nominal rotor radius (R) 8.0 cm
Number of blades (Nb) 3
Rotor solidity (?) 12%
Blade root cut-out (R0) 18%R
Blade planform Rectangular (no taper)
Blade twist (?tw) 0?
Blade chord (c) 1.0 cm
Blade airfoil Flat-plate (uncambered), rectangular profile
(blunt leading, trailing edges), 5% thick
and torque of MAV-scale rotors.4 The models were mounted inverted (thrusting
downwards) to avoid the effects of a ground plane on their performance. The models
were tested at collective angles from 5? to 40?, in steps of either 5? or 10?, and
rotational speeds of 2000 to 4000 rpm, corresponding to blade tip speeds of 55 to
110 ft/s, tip Reynolds numbers of approximately 11,500 to 23,000, and tip Mach
numbers of 0.05 to 0.10. A few initial tests were conducted at rotational speeds from
500 to 5000 rpm, to determine the effects of Reynolds number on the performance.
The atypically high ? for a conventional rotor ? collective values tested are in
accordance with the blade angles used in other shrouded-rotor applications and
test programs, as described in the previous chapter. For the shrouded rotor, these
4The test-stand was designed by Dr. Jayant Sirohi, also of the Alfred Gessow Rotorcraft Center.
The sensing elements were commercial, off-the-shelf (COTS) precision strain-gauge load cells.
Thrust was measured by a Honeywell Sensotec Model 31 Precision Miniature Tension/Compression
Load Cell (http://www.sensotec.com/). Torque was measured by a Transducer Techniques RTS-
Series Reaction Torque Sensor (http://www.transducertechniques.com/). The sensing elements
used for these tests had full-scale ranges of 1000 g (2.2 lb) and 10 oz-in (0.07 N-m, 0.05 ft-lb),
respectively, but other sensors of the same form factor and with different capacities are available
from these vendors and can also be used in this test-stand.
115
high collectives compensate for the greater inflow seen at the plane of the rotor
disk, compared to the case of an open rotor. The open rotors were also tested at
these high collectives to obtain a comparison of the performances of the open and
shrouded rotors at the same collective angles. The rotational speed was measured
by an arrangement involving a Hall Effect sensor5 mounted to the rotor support
structure and two magnets mounted diametrically opposite each other on the rotor
hub. The passage of the magnets past the sensor produced electric pulses that were
counted by the data acquisition program. Two magnets were used instead of one
so as to keep the rotor balanced and to improve the accuracy of the measurements.
The speed measurements were verified with an optical tachometer6, which emitted a
laser beam and counted the pulses reflected from a reflective-tape target positioned
on the rotor hub. The mechanical shaft power of the rotors was then calculated as
the product of the torque and rotational speed.
The shroud thrust (Tshroud) was calculated by integration of the measured
pressure distributions over the shroud surface. The integrated shroud thrust was
then subtracted from the total thrust to arrive at the rotor thrust (Trotor). The
models for which pressure data were not collected ? and for which, therefore, the
division of thrust between the rotor and shroud could not be calculated ? were
the LR06-D10-?0.1, LR09-D10-?0.6, LR09-D20-?0.5 and LR13-D00-?0.6 models. It
was recognized that the measurement of shroud thrust by this method results in
an upper-bound value for the actual shroud thrust, since it does not account for
5Sypris/F. W. Bell Model FH-301 (http://www.fwbell.com/).
6MonarchInstrumentsPhasar-TachseriesPhaser-Laser(http://www.monarchinstrument.com/).
116
Figure 2.4: Hover test setup
drag due to surface friction on the shroud and download on the shroud support
struts. An external frame was therefore constructed to support the shrouds while
the rotor inside was separately mounted on the test-stand, unattached to the shroud.
Difficulties were experienced, however, in accurately positioning the shroud relative
to the rotor, resulting in frequent impacts between the rotor blades and the shroud;
this meaurement procedure were therefore abandoned.
Measurements were also made of the axial velocity (dynamic pressure) dis-
tributions in the wakes of the open rotor and of six of the shrouded rotors ? the
LR13-D10 models ? using a pitot-static probe. For the open rotor, the measure-
ments were made at a collective of 20?, and at 89 locations in the wake: at radial
stations r/R from 0.125 to 1.0, in steps of either R/8 (1 cm) or R/16, and at axial
stations z/R from 0.04 to 1.60, in steps of R/5 (1.6 cm). The density of sampling
117
locations was increased for the regions of large gradients in the measured velocities,
which corresponded to the trajectory of the blade tip vortices. A similar set of
measurements was also made at 60 locations in the wake of shrouded-rotor model
LR13-D10-?0.6, also at a rotor collective of 20?. A third set of tests consisted of
measurements at different radial stations at the diffuser exit plane only, for the
five other LR13-D10 models, at rotor collectives of 10?, 20?, 30? and 40?. Both the
shroud surface pressure distribution and the wake velocity measurements were made
with a variable-capacitance, very-low-pressure transducer7, with a full-scale range
of 0?50 Pa (0?0.007 psi).
2.3 Wind-tunnel test setup
The translational-flight tests were performed in a open-jet wind tunnel with a
22 in x 22 in test section and a turbulence level of 0.3% (Fig. 2.5). A four-component
sting balance (Figs. 2.6, 2.7) was designed and constructed to measure the normal
force, N, pitch moment, M, roll moment (rotor torque, Q), and axial force (rotor
thrust, T) (Fig. 2.8). The lift, L, and drag, D, were determined from the measured
thrust and normal force by a coordinate transformation from the body axes to the
wind axes:
L = N cos(?) +T sin(?) (2.1)
D = N sin(?)?T cos(?) (2.2)
7Setra Systems Model 267 HVAC/R-series (Heating, Ventilation, Air-Conditioning and Refrig-
eration) Very Low Differential Pressure sensor (http://www.setra.com/).
118
Figure 2.5: Wind tunnel test setup
In this study, the angle of attack, ?, is defined as the angle made by the axis
of rotation of the rotor with the freestream velocity vector, and is therefore the
complement of the angle ?rotor which is considered to be the angle of attack in
conventional rotor analysis (Fig. 2.9).
Note that the pitch moment at the rotor hub can be considered as equivalent
to the normal force, which is perpendicular to the rotor axis, acting at a point on
the axis henceforth referred to as the ?center of pressure? and located at a distance
xcp above the plane of the rotor disk (Fig. 2.10). xcp is therefore positive for points
119
Figure 2.6: Four-component wind-tunnel sting balance
Figure 2.7: Sting balance with shrouded-rotor model mounted
120
Figure 2.8: Forces and moments acting on model in translational flight
121
Figure 2.9: Angle of attack definition
Figure 2.10: Center of pressure
122
above (upstream of) the rotor, and negative for points below (downstream of) the
rotor.
The rotational speed and shaft power of the models was measured in the
same way as in the hover tests. Like the hover-stand, the balance was designed to
measure the very low forces and moments generated on MAV-scale models. The
roll moment and axial force were measured using the same torque and force sensors
that were used in the hover-stand, while the normal force and pitch moment were
measured using a beam with two sets of strain-gauged flexures (Fig. 2.6). The full-
scale ranges of the roll-moment and axial-force sensors were 0.6 in-lb and 2.2 lb,
respectively, while the normal-force/pitch-moment sensing element was calibrated
up to 0.7 lb and 4.0 in-lb. To avoid the normal-force sensing element having to
bear the weight of the model, the test setup was designed so that the model would
?pitch? in a horizontal plane instead of a vertical one (Fig. 2.5), so the normal force
and pitch moment are actually the side-force and yaw moment in the ?earth? frame
of reference. The model support structure was designed so that any point on the
model longitudinal axis could be positioned directly above the angle-of-attack pivot
axis. For the tests described in this paper, this point was chosen to be the location
of the rotor hub.
Theshroudwasinstrumentedwithtwenty-fourpressuretaps, asshownschemat-
ically in Fig. 2.11. Twelve of the taps were at the ? = 180? rotor azimuth, corre-
sponding to the windward side of the shroud when the model is pitched forward
into the wind, and twelve were at the ? = 0? azimuth, corresponding to the leeward
or downwind side (Fig. 2.8). Within each set of twelve taps, eight were distributed
123
Figure 2.11: Pressure tap locations (wind-tunnel model)
over the shroud inlet and four along the inner surface of the diffuser, with a greater
concentration of taps near the plane of the rotor disk. Figure 2.12 shows the test
setup after the model had been instrumented with the pressure taps. The bundles of
pressure tubing were adjusted to remain out of the rotor wake and to exert minimal
force on the model due to their own weight. Although it was desired to have a com-
plete picture of the flow-field around the entire model, the small size of the model
resulted in limited room for additional tubing, and therefore it was not possible to
intrument the outer surface of the shroud with pressure taps as well.
For both the open and shrouded rotors, the test matrix consisted of free-stream
velocities of 0, 10, 15 and 20 ft/s, and angles of attack from -45? to +90?, in steps of
either 5? or 15?, where ? = 0? corresponds to axial flow (vertical climb or ?propeller
mode?), and ? = 90? corresponds to hover in edgewise flow. The performance
measurements were made at rotational speeds of 2000, 3000 and 4000 rpm, while
the surface pressure measurements were made at a rotational speed of 3000 rpm
124
(a) Side view
(b) Rear view
Figure 2.12: Wind tunnel model instrumented with pressure taps
125
only. All tests were performed at a fixed rotor collective of 20?, which was the
collective at which the rotor exhibited optimal performance in hover. Note that the
rotor was not trimmed in any way.
2.4 Test procedure and uncertainty analysis
For all tests, the data were acquired and analyzed using the MATLABR?
programming environment8 and a National Instruments data-acquisition system
(?DAS?)9 with 12-bit resolution, or a least count of 1/212 times the full-scale in-
put range of the individual data channels. For the hover tests, the input ranges
were set to ?5 V for all the channels, resulting in a resolution (least count) of
2 g for the thrust measurements, 0.02 oz-in (0.14?10?3 N-m) for the torque and
0.025 Pa for the pressure. For the wind-tunnel tests, depending on the values of the
forces, moments or pressures being produced, the input ranges were reduced so as
to obtain even smaller resolutions in the measurements.
The procedure that was used to acquire the data in both the hover and the
wind-tunnel tests is described below. For each point in the test matrices:
1. The rotor rotational speed was adjusted to within ?10 rpm of the nominally
desired speed.
2. With the rotor running [and the wind tunnel on], four [five] seconds of data
were acquired from the transducers at a rate of 1000 [2000] samples/second
8http://www.mathworks.com/products/matlab/
9National Instruments DAQCard-6062E, 12-bit resolution (http://www.ni.com/). Some of the
earliest tests were conducted with a National Instruments model PCI-6031E DAS which had 16-bit
resolution.
126
from each transducer, and the values averaged to eliminate unsteady effects.
A higher sampling rate and a longer time period were found necessary for the
wind-tunnel tests to compensate for the higher vibration levels in the models
in those conditions.
The sampling rate ? at 1000 Hz, more than ten times higher than the highest
rotational speed at which the models were tested (5000 rpm, 83.3 Hz) ? and
the lengths of time were chosen so as to be able to average out the electrical
noise in the computerized data-acquisition system and any other random or
periodic fluctuations in the operation of the rotor. Although this investigation
was only concerned with the steady values of test variables, the setup al-
lowed for detection of vibrations of frequencies up to the Nyquist frequency of
500 Hz, corresponding to a rotor harmonic of 6/rev or more, with a resolution
of 0.25 Hz10.
3. The rotor [and tunnel] were then stopped, and, after all transients were allowed
to die out, the transducer outputs were recorded again for a period of one
second and averaged. In the hover tests, the time required for to allow the
model to come to rest varied from five to fifteen seconds, depending on the
inertia of the specific model being tested. For the wind-tunnel tests, this
duration was increased to 20?30 s.
4. The difference between the two averaged values ? rotor-on[/wind-on] and
rotor-off[/wind-off] ? was recorded as the measurement for that particular
10The frequency resolution ?f is equal to 1/T, where T is the length of time for which data
were acquired.
127
run. This procedure was used for each run to avoid errors from drifting of the
transducers.
5. Each measurement run was repeated from four to ten times, until the scatter
in the measurements fell to acceptable levels.
The uncertainties in the measurements, and in the ?final? reported values that
are discussed in the subsequent chapters, were estimated by the following procedure.
For each test variable (thrust, torque, pressure):
1. The uncertainty in the measurement value from each run ? the difference
between the two averages, rotor-on[/wind-on] and rotor-off[/wind-off] ? was
considered to be equal to the ?reading error? for the test variable. For these
digitally-acquired data, this would be the discretization or quantization error
as set by the resolution of the DAS, and equal to plus/minus one-half of the
least count for each variable.
Although this was not done, a more rigorous estimate would have been ob-
tained by the following procedure:
(a) For each of the 4000/5000/1000 measurement samples acquired during
the rotor-on[/wind-on] and rotor-off[/wind-off] phases, estimate the sta-
tistical error, due to random variations, in the individual sample value
by the standard deviation of the 4000/5000/1000 values.
(b) Compare the value of this statistical error to the value of the reading
(discretization) error, and choose the larger of the two values as the mea-
surement error for each individual sample.
128
(c) Propagate the measurement error through the averaging and differencing
operations to obtain the estimate for the uncertainty in the final mea-
surement value for that run.
Since a very large number of samples (4000/5000/1000) were acquired, the final
propagated value of the uncertainty, obtained by this more rigorous method,
would have been much lower than the uncertainty in each individual sample
value. Therefore, the value of the uncertainty that was actually used for the
final measurement from the run ? i.e., the reading error ? is quite likely to
have been very conservative.
2. Using this procedure, multiple measurement runs were conducted until the
scatter in the values from each run, as quantified by the standard deviation
of these values over all runs, fell to about the same magnitude as that of
the reading error, or lower. Because of fluctuations in the output of the power
supply controlling the drive motor and other unsteady factors in the operation
of the rotors, the average rotational speed in any given measurement run
typically varied by up to 10?15 rpm between different runs; hence, the values
of the measurements from each run would be expected to be slightly different
from one run to another. Even so, the reduction in the spread ? the standard
deviation ? of the values gave a good estimate for the improvement in the
precision of the measurements. Typically, only about 4?5 runs were required to
obtain satisfactory precision, but in some cases up to 10?12 runs were required.
3. The measurement values of the test variables from each run were converted
129
into their non-dimensional coefficient forms. The uncertainties in the mea-
surements of the rotational speed (?1 rpm), the rotor radius (?0.001 cm), the
air density (?0.05 kg/m3) and the (dimensional) values of the test variables
were propagated through these calculations to obtain the uncertainty in the
value of the coefficient for that run.
4. The final value of the coefficient for that point in the test matrix was obtained
by averaging over the values of the coefficients from all runs. The uncertainties
in the values from each run were propagated through the averaging operation
to obtain the uncertainty in the mean value. Note, however, that this uncer-
tainty is still due only to the reading errors in the different measurements.
5. The statistical (random) error in the coefficient values for each run was esti-
mated by the standard deviation of the values over all runs. This is a valid
operation because, even though the dimensional values of the test variables
would be expected to vary from one run to another, the coefficient values
would be expected to be the same. This statistical uncertainty, which gives a
measure of the spread of the measurements from different runs (the precision,
or, rather, the lack thereof), was similarly propagated through the averaging
operation over all runs.
6. The larger of the two propagated uncertainty values ? the one due to reading
errors and the other due to random effects ? was selected as the estimate of
the uncertainty in the ?final? mean value of the coefficient.
130
Chapter 3
Experimental Results: Hover Tests
3.1 Introduction
This chapter presents the performance and flow-field measurements made in
the hover tests of the open and shrouded rotors, and the comparisons of the dif-
ferent shrouded-rotor models based upon those measurements. The data ? and
the associated analyses ? are presented in three sections: first, the performance
characteristics (thrust, power, efficiency) and comparisons with the predictions of
momentum theory; second, the shroud surface pressure distributions; and third, the
wake velocity distributions. Within each section, the effects of varying each of the
shroud geometric parameters (blade tip clearance, inlet lip radius, diffuser angle,
diffuser length, diffuser expansion ratio) are discussed. In all cases, the data are
presented in non-dimensional form ? as force, power and pressure coefficients and
as normalized distances and velocities. For obtaining the coefficients of the open
rotors, the tip speed and disk area of the individual rotors were used as the reference
velocity and reference area; for the shrouded rotors, since rotors with different blade
tip clearances were tested, the reference velocity was the tip speed of the actual
rotor used for the individual model, while the reference area was that of the shroud
throat.
131
3.2 Simplifying assumptions
Two simplifying assumptions were made prior to performing the analyses pre-
sented in this chapter. The first was that the values of the coefficient/normalized
forms of the various quantities did not vary significantly between the tested ro-
tational speeds of 2000 and 4000 rpm, and that the values used in the analyses
comparing the various shrouded-rotor models could therefore be those obtained by
averaging over all the measurements made at all rotational speeds within this range.
The second assumption was that the reduction in diameter of the rotor, which was
required in order to test models with different blade tip clearances, did not signifi-
cantly change the performance of the (open) rotor itself, and that, for the purpose
of comparison with the shrouded rotors, the different open rotors could therefore be
represented by a single ?representative? open rotor, the diameter and performance
of which were obtained by averaging over the values of all the rotors. The reasoning
and justification for these assumptions are given below.
3.2.1 Assumption regarding rotational speed
For the first open rotor and shrouded-rotor model tested (LR09-D20-?0.5),
the hover tests were conducted at rotational speeds from 500 rpm up to 5000 rpm,
to determine the effects of Reynolds number on the data. The thrust and power
coefficients for these two cases are shown in Figs. 3.1a?d, plotted for the different
blade collective angles as functions of the rotational speed. It can be seen that
the mean values of the coefficients, calculated as described above, remain fairly
132
constant over the range of speeds tested, although some variation is seen at the
lower speeds. More significant, though, is the fact that the uncertainties in the mean
values ? estimated by the method described in Section 2.4 (p. 126), and shown by
the errobars in the figures ? increase significantly as the speed is decreased. This
is due to two reasons: first, the increased sensitivity of the flow, and hence of the
rotor performance, to small random disturbances at lower Reynolds numbers [8];
and second, the decrease in the relative magnitude of the absolute, dimensional
values of the various measured quantities (thrust, torque, rotational speed) to the
uncertainties in their measurement ? effectively, a reduction in the signal-to-noise
ratio. The first increases the uncertainty due to the statistical spread of the values
from different measurement runs ? the ?random error?, while the second increases
the uncertainty due to the inherent limits on the precision in measuring a physical
quantity ? the ?reading error? (see Section 2.4, p. 126). Figures 3.1e?f show a
comparison of these two sources of uncertainty, along with their variations with
changing rotational speed. It can be seen that, in these tests, the uncertainties due
to the reading errors were typically greater than those due to the random errors, and
especially so at the lower rotational speeds. As mentioned in Section 2.4, the final
estimates of the uncertainties shown in the presentation of the data in this chapter
were the greater of the two values for each individual point in the test matrix.
It is thus clear that, to obtain results with good confidence levels, the tests
should be conducted at the higher rotational speeds. However, at the higher blade
collective angles, the increased drag on the blades and the power limitation of the
133
0 1000 2000 3000 4000 5000 6000?0.01
0
0.01
0.02
0.03
C T
Rotational speed [rpm]
?0 = 0o?
0 = 5o?0 = 10o
?0 = 15o?
0 = 20o?0 = 25o
?0 = 30o?
0 = 35o?0 = 40o
(a) Open rotor: thrust coefficient.
0 1000 2000 3000 4000 5000 60000
0.005
0.01
0.015
C P
Rotational speed [rpm]
?0 = 0o?
0 = 5o?0 = 10o
?0 = 15o?
0 = 20o?0 = 25o
?0 = 30o?
0 = 35o?0 = 40o
(b) Open rotor: power coefficient.
0 1000 2000 3000 4000 5000 6000?0.01
0
0.01
0.02
0.03
C T
Rotational speed [rpm]
?0 = 5o?
0 = 10o?0 = 15o
?0 = 20o?
0 = 25o?0 = 30o
?0 = 35o?
0 = 40o
(c) LR09-D20-?0.5: thrust coefficient.
0 1000 2000 3000 4000 5000 60000
0.005
0.01
0.015
C P
Rotational speed [rpm]
?0 = 5o?
0 = 10o?0 = 15o
?0 = 20o?
0 = 25o?0 = 30o
?0 = 35o?
0 = 40o
(d) LR09-D20-?0.5: power coefficient.
0 1000 2000 3000 4000 50000
0.002
0.004
0.006
0.008
0.01
0.012
? C T
Rotational speed [rpm]
Uncertainty due to reading errorUncertainty due to statistical (random) error
(e) LR09-D20-?0.5: uncertainty in CT.
0 1000 2000 3000 4000 50000  
0.0002
0.0004
0.0006
0.0008
0.0010
? C P
Rotational speed [rpm]
Uncertainty due to reading errorUncertainty due to statistical (random) error
(f) LR09-D20-?0.5: uncertainty in CP.
Figure 3.1: Effects of changing rotational speed on performance
134
0 1000 2000 3000 4000 5000 60000
1
2
3
4
5
6
Power [W]
Rotational speed [rpm]
?0 = 0o?
0 = 5o?0 = 10o
?0 = 15o?
0 = 20o?0 = 25o
?0 = 30o?
0 = 35o?0 = 40o
(a) Open Rotor
0 1000 2000 3000 4000 5000 60000
1
2
3
4
5
6
Power [W]
Rotational speed [rpm]
?0 = 5o?
0 = 10o?0 = 15o
?0 = 20o?
0 = 25o?0 = 30o
?0 = 35o?
0 = 40o
(b) Shrouded-rotor LR09-D20-?0.5
Figure 3.2: Effect of increasing rotational speed on power consumption
drive motor1 meant that speeds of only up to 2000 or 3000 rpm could be tested
(Figure 3.2). Therefore, for the subsequent models, the tests were conducted at
rotational speeds from 2000 to 4000 rpm only; and since the mean values of the
coefficients at the different speeds within this range were not significantly different,
the values of the coefficients were therefore averaged over all rotational speeds tested
(between 2000 and 4000 rpm) for the analyses that are presented in the following
sections.
3.2.2 Assumption regarding rotor radius
As mentioned earlier, the different blade tip clearances in the shrouded-rotor
models were created by using rotors with blades of different lengths. Seven different
rotors were used, with diameters varying from 154.0 mm to 159.7 mm ? a range of
3.6%Dt. Figures 3.3a?b compare the performance of six of the rotors at four different
1The electric drive motor had a maximum ?safe? rating of 8.5 watts. Accounting for the motor
efficiency ? which, for these motors, is typically around 50% ? and for transmission losses, the
power available at the rotor would therefore be much lower.
135
collective angles, showing the thrust and power coefficients plotted as functions
of rotor diameter, along with the average of the values of all the rotors at each
collective. In these figures, the uncertainty bounds for the individual rotors are
shown by the errorbars, while those for the mean values are shown by the bands
enclosed by the dashed lines. It can be clearly seen that the variations in CT and
CP with changing diameter are quite random, i.e., the reduction in diameter had no
generalizable effect on the performance of the rotor. Therefore, for the purpose of
comparisons with performance of the shrouded rotors, it was justifiable that, within
the limits of experimental uncertainty, all seven rotors could be represented by a
single ?representative? open rotor, the performance of which ? in non-dimensional
coefficient terms ? was taken to be equal to the averages of the values of the
individual rotors. This is further illustrated by Figure 3.3c, which shows the rotor
polar ? CT vs. CP ? of the resulting representative open rotor, along with the
measurements of the individual open rotors and those of one of the shrouded-rotor
models.
3.3 Performance measurements
Inthissection, thegeneralperformancecharacteristicsoftheopenandshrouded
rotors are first discussed, followed by more detailed analysis of the effects of varying
the individual shroud geometric parameters.
136
153 154 155 156 157 158 159 1600.006
0.01
0.014
0.018
0.022
Rotor diameter [mm]
C T
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(a) Thrust coefficient
153 154 155 156 157 158 159 1600
0.004
0.008
0.012
Rotor diameter [mm]
C P
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(b) Power coefficient
0 0.004 0.008 0.0120
0.01
0.02
0.03
C T
CP
OR: D = 154.0 mmOR: D = 156.9 mm
OR: D = 157.0 mmOR: D = 158.4 mm
OR: D = 158.7 mmOR: D = 159.7 mm
OR: RepresentativeSR: LR13?D10??0.1
(c) Rotor polar
Figure 3.3: Effect of changing diameter on the performance of open rotors
137
3.3.1 General characteristics
The performance data for all seventeen shrouded rotors, as well as for the
representative open rotor, are shown in Figs. 3.4?3.12. In these figures, the measured
values are shown by the symbols and are connected by straight line-segments to
reveal the trends more clearly. Figures 3.4 and 3.5 show the ?raw? thrust (CT) and
power (CP) data plotted as functions of blade collective angle (?0). The maximum
values of the thrust and power coefficients obtained in these tests ? 0.02?0.03 for
CT, 0.010?0.014 for CP ? are substantially higher than those typically seen for
larger-scale rotors. This is partly attributable to the higher collective angles at
which these micro-scale rotors have been tested, but more so due to their much
lower tip speeds: at a typical rotational speed of 3000 rpm, these rotors had a tip
speed of around 80 ft/s, which is almost nine times lower than the typical values of
700 ft/s and higher seen for full-scale manned helicopters [15, pp. 683?701]. Even
though the disk loadings of these micro-rotors (up to 0.5 lb/ft2) are also much lower
than those for large-scale helicopters (typically up to 10 lb/ft2 [14, p. 487]), the
inverse scaling of the thrust coefficient with the square of the tip speed results in
much higher values of CT; similar reasoning applies to the values of CP.
Figures 3.4 and 3.5 also show that, for collectives above 10??15?, the shrouded
rotors all produce higher thrust than the open rotor, at the same collective angle,
while consuming less power at all collectives. The increase in thrust coefficient ?
up to 50% more ? is greatest at the higher collectives, while the decrease in power
? also by up to 50% ? is greatest at the lower collectives. The reduced power is
138
0 10 20 30 40
0
0.01
0.02
0.03
?0 [deg]
C T
Open rotorLR06?D00??0.6
LR09?D00??0.6LR13?D00??0.6
LR09?D20??0.5LR13?D20??0.6
0 10 20 30 40
0
0.01
0.02
0.03
?0 [deg]
C T
Open rotorLR13?D10??0.1?L31
LR13?D10??0.8?L31LR13?D10??1.6?L31
LR13?D10??0.1?L50
0 10 20 30 40
0
0.01
0.02
0.03
?0 [deg]
C T
Open rotorLR06?D10??0.1
LR09?D10??0.6LR13?D10??0.1
LR13?D10??0.6
0 10 20 30 40
0
0.01
0.02
0.03
?0 [deg]
C T
Open rotorLR06?D20??0.1
LR09?D20??0.1LR13?D20??0.1
LR13?D20??0.8
Figure 3.4: Thrust coefficient vs. collective angle
139
0 10 20 30 400
0.004
0.008
0.012
?0 [deg]
C P
Open rotorLR06?D00??0.6
LR09?D00??0.6LR13?D00??0.6
LR09?D20??0.5LR13?D20??0.6
0 10 20 30 400
0.004
0.008
0.012
?0 [deg]
C P
Open rotorLR13?D10??0.1?L31
LR13?D10??0.8?L31LR13?D10??1.6?L31
LR13?D10??0.1?L50
0 10 20 30 400
0.004
0.008
0.012
?0 [deg]
C P
Open rotorLR06?D10??0.1
LR09?D10??0.6LR13?D10??0.1
LR13?D10??0.6
0 10 20 30 400
0.004
0.008
0.012
?0 [deg]
C P
Open rotorLR06?D20??0.1
LR09?D20??0.1LR13?D20??0.1
LR13?D20??0.8
Figure 3.5: Power coefficient vs. collective angle
140
due both to the reduction in the induced power because of the increased expansion
ratio and the consequent lower velocity of the wake, as well as to the reduction in
profile power because of the off-loading of the rotor by the shroud, causing the rotor
blade elements to operate at lower angles of attack. The reason for the reduced
thrust of the shrouded rotors at lower collectives, compared to the open rotor,
can be understood by considering an approximate blade-element-momentum-theory
(BEMT) model of the hovering rotor, in which it is assumed that the induced inflow
(?) is constant over the entire rotor disk and that all blade elements operate at the
same angle of attack, derived from the mean lift coefficient of the rotor ( ?CL) [14,
p. 113]. The pitch angle, ?, of a blade element at any radial station along the blade
is then given by the sum of the aerodynamic angle of attack (?) and the induced
inflow angle (?):
?(r) = ?(r) +?(r)
=
?CL
a +
?
r
=
parenleftbigg1
a
parenrightbigg6C
Trotor
? +
??
dCT
r
=
parenleftbigg 6
?a
parenrightbiggC
T
2?d +
??
dCT
r (3.1)
Thus, for the same thrust coefficient, increasing the expansion ratio causes a re-
duction in the aerodynamic angle of attack of a blade element but an increase in
the induced inflow angle. The corresponding change in the required rotor collective
depends on the relative magnitudes of the two terms in Eq. (3.1). At very low thrust
coefficients, the induced-inflow term (proportional to ??d) dominates, so that in-
141
creasing ?d results in a net increase in the pitch angle required for a given CT, or,
equivalently, a lower CT at the same rotor collective. At higher thrust coefficients,
the angle-of-attack term (proportional to 1/?d) dominates, so that increasing ?d re-
sults in a net decrease in the pitch angle required for a given CT, or, equivalently, a
higher CT at the same ?0. Equation (3.1) has been plotted in Fig. 3.6 for an open
rotor2 and for different values of ?d for a shrouded rotor, and the cross-over seen in
the experimental data in Fig. 3.4 can be clearly seen in these theoretical predictions
as well, occurring between blade tip angles of 3? and 5?. A reduction in either the
expansion ratio or the lift-slope (a) of the blade airfoils causes the first term in
Eq. (3.1) to dominate and the cross-over to occur earlier, i.e., at a lower collective.
The higher values of ?0 at which the cross-overs are seen in the experimental data
are probably due to factors that are not accounted for in this simplified model.
The improved performance of the shrouded rotors is more evident when the
thrusts of the two configurations are compared at the same power (Fig. 3.7). At the
same power coefficient, some of the shrouded rotors ? models LR13-D10-?0.1-L72
and -L50, specifically ? have thrust coefficients up to 94% higher than that of the
open rotor, with the greatest improvements occurring around CP = 0.004. Con-
versely, at the same thrust coefficient, the shrouded rotors have power coefficients
up to 70% lower than that of the open rotor, with the greatest reductions occur-
2 For the special case of an open rotor, Eq. (3.1) becomes:
?OR(r) =
parenleftbigg 6
?a
parenrightbigg
CT +
radicalbigC
T/2
r
Thus, for comparing the aerodynamic angles of attack and the induced inflow angles of the blade
elements of a shrouded rotor and an open rotor at the same thrust coefficient, this yields the
relations ?SR = 12?
d
?OR and ?SR = ?2?d?OR.
142
0 1 2 3 4 50
0.001
0.002
0.003
0.004
C T
?tip [deg]
Open Rotor (?d = 0.5)Shrouded Rotor, ?
d = 0.75Shrouded Rotor, ?d = 1
Shrouded Rotor, ?d = 1.25
(a) Low thrust coefficients
0 5 10 15 20 25 300
0.01
0.02
0.03
0.04
C T
?tip [deg]
Open Rotor (?d = 0.5)Shrouded Rotor, ?
d = 0.75Shrouded Rotor, ?d = 1
Shrouded Rotor, ?d = 1.25
(b) High thrust coefficients
Figure 3.6: Theoretical prediction of thrust coefficient as a function of blade tip
angle, for different values of expansion ratio (Equation (3.1), ? = 0.1, a = 5.73?per
radian, r = 1.)
ring near CT = 0.02, which is approximately the highest value of CT achieved by
the open rotor. Note that these comparisons are made while considering the actual
power, not the ideal power, and therefore would be expected to be different from
the predictions from momentum theory that were given in Table 2.2 (p. 113).
The power loading, PL = T/P, is the quantity that the designer of an aircraft
is ultimately interested in, because it directly gives the obtainable increase in thrust
at the same power, or, conversely, the decrease in power at the same thrust. While
the rotor polars ? the graphs ofCT versusCP ? are useful means of comparing the
performances of the open and shrouded rotors, they are only good for comparisons
when the different configurations are forced to operate at the same thrust or power
coefficient. This is not a realistic circumstance in the design of an aircraft, since
any rotor design would typically be operated as near as possible to its collective
for maximum efficiency (power loading). Since the rotor polar does not directly
indicate the value of this collective, it does not easily enable comparisons of the
143
0 0.004 0.008 0.012
0
0.01
0.02
0.03
CP
C T
Open rotorLR06?D00??0.6
LR09?D00??0.6LR13?D00??0.6
LR09?D20??0.5LR13?D20??0.6
0 0.004 0.008 0.012
0
0.01
0.02
0.03
CP
C T
Open rotorLR13?D10??0.1?L31
LR13?D10??0.8?L31LR13?D10??1.6?L31
LR13?D10??0.1?L50
0 0.004 0.008 0.012
0
0.01
0.02
0.03
CP
C T
Open rotorLR06?D10??0.1
LR09?D10??0.6LR13?D10??0.1
LR13?D10??0.6
0 0.004 0.008 0.012
0
0.01
0.02
0.03
CP
C T
Open rotorLR06?D20??0.1
LR09?D20??0.1LR13?D20??0.1
LR13?D20??0.8
Figure 3.7: Thrust coefficient vs. power coefficient
144
different configurations at their respective points of maximum efficiency. Figures 3.8
through 3.12 therefore show three different measures of efficiency which do enable
such a comparison, plotted as functions of collective angle and of thrust coefficient.
Two of these measures are familiar to rotor aerodynamicists: the ratio of thrust
coefficient to power coefficient, CT/CP (Figs. 3.8, 3.9), which is maximum at the
point of maximum power loading when the tip speed of the rotor is constrained to
a fixed value:
T
P =
1
vtip ?
CT
CP (3.2)
and the figure of merit, FM (Figs. 3.10, 3.11), which shows a maximum at the point
for optimal operation when the tip speed is allowed to vary but the disk area is
constrained:
T
P =
radicalbigg2?
DL ?FM =
radicalbigg
4?d?A
T ?FM (3.3)
where now, in the general case of a shrouded rotor, the disk loading (DL) is equal
to Trotor/A = T/2?dA. These two measures are appropriate for determining the
optimal operating collective of a given rotor, open or shrouded, and can be used for
comparing the performances of different open-rotor designs or of different shrouded
rotors with the same diffuser expansion ratio. However, when comparing designs
with different expansion ratios, use of the traditional formulation of the figure of
merit (Eq. (1.3)) can lead to incorrect conclusions. This can be seen from Eqs. (1.2)
and (1.3): at the same total thrust, increasing the ?d causes the ideal power ?
and hence the figure of merit ? to decrease, even though the total power require-
ment has actually decreased, and the power loading (efficiency) increased. However,
145
Eq. (3.3) suggests a different efficiency measure that would overcome this limitation
? a ?generalized? figure of merit, FM? = FM??d, which is nothing but C3/2T /2CP.
This measure can be used for comparing both open and shrouded rotors, whether
at the same or different value of expansion ratio, as well as for determining the
optimal operating point for a given rotor, open or shrouded, since the maximum
FM and maximum FM? both occur at the same point (collective). Comparisons of
the traditional FM are still permissible when the different designs have the same
rotor thrust (disk loading); however, this is of little practical use, since it is the total
thrust that a designer is interested in. At a desired total thrust and disk area, the
required power is inversely proportional to FM?, while at a given power and disk
area, the thrust that would be produced is directly proportional to (FM?)2/3. Thus,
a graph of FM??d versus CT (Fig. 3.12) would perform the same role as that of the
traditional FM-vs.-CT graph for an open rotor, while also allowing for valid compar-
isons of the performances of configurations with different expansion ratios. The only
disadvantage of this measure is that, unlike the traditional figure of merit, it does
not have a maximum possible value of unity, having instead a maximum possible
value of ??d, and therefore does not give as immediate an idea of how close a given
design is to ?perfection.? For this purpose, the traditional figure of merit remains
the better efficiency measure. Further discussion about these different measures of
rotor efficiency is given in Appendix B (p. 297) of this dissertation.
Figures 3.8 and 3.9 show that, for the open rotor, the maximum value ofCT/CP
occurs at a collective angle of 10? (CT = 0.0075), while for the shrouded rotors it
occurs between collectives of 10? and 20? (CT = 0.0075?0.0225), depending on the
146
specific model. Although measurements were made for only a few of the shrouded
rotors at ?0 = 15?, the trends in the data suggest that the maximum CT/CP may
occur at this collective angle for all the shrouded-rotor models (when used with this
specific rotor). The greatest improvement in performance, therefore, according to
this measure of efficiency, is by around 94%, comparing the maximum value of 3.5
for the open rotor with that of 6.8 for shrouded-rotor model LR13-D10-?0.6. This
corresponds to either a 94% increase in thrust at the same power, or a 49% reduction
in power at the same thrust, compared to the open rotor, if the tip speeds are held
the same (Eq. 3.2). Although data were not taken with this shroud and a smaller
tip clearance (0.1%Dt) at ?0 = 15?, it is quite probable that the smaller tip gap
would have resulted in even greater improvements in performance.
In terms of the figure of merit (Figures 3.10 and 3.11), the shrouded rotors
show maximum values of up to 0.42 (model LR13-D10-?0.1-L50), which is a 68%
improvement over the maximum value of 0.25 for the open rotor, and these maxima
universally occur at a collective angle of 20?, for both the open and shrouded rotors.
The thrust coefficients at which these maxima occured were 0.015 for the open rotor
and 0.020 or slightly higher for the shrouded rotors. As expected, the FM values
for the models with the largest expansion ratio ? the D20 models ? are in some
cases lower than those for the open rotor, which would imply a seemingly reduced
efficiency for these models. The generalized form of the figure of merit eliminates
this misleading effect of the conventional form of FM, and clearly shows (Figure 3.12)
that all of the shrouded rotors are indeed much more efficient than the open rotor
? up to 160% more efficient in the case of some of the LR13-D10 models, which
147
0 10 20 30 40
0
2
4
6
8
?0 [deg]
C T/C
P
Open rotorLR06?D00??0.6
LR09?D00??0.6LR13?D00??0.6
LR09?D20??0.5LR13?D20??0.6
0 10 20 30 40
0
2
4
6
8
?0 [deg]
C T/C
P
Open rotorLR13?D10??0.1?L31
LR13?D10??0.8?L31LR13?D10??1.6?L31
LR13?D10??0.1?L50
0 10 20 30 40
0
2
4
6
8
?0 [deg]
C T/C
P
Open rotorLR06?D10??0.1
LR09?D10??0.6LR13?D10??0.1
LR13?D10??0.6
0 10 20 30 40
0
2
4
6
8
?0 [deg]
C T/C
P
Open rotorLR06?D20??0.1
LR09?D20??0.1LR13?D20??0.1
LR13?D20??0.8
Figure 3.8: Ratio of thrust coefficient to power coefficient vs. collective angle
148
0 0.01 0.02 0.03
0
2
4
6
8
CT
C T/C
P
Open rotorLR06?D00??0.6
LR09?D00??0.6LR13?D00??0.6
LR09?D20??0.5LR13?D20??0.6
0 0.01 0.02 0.03
0
2
4
6
8
CT
C T/C
P
Open rotorLR13?D10??0.1?L31
LR13?D10??0.8?L31LR13?D10??1.6?L31
LR13?D10??0.1?L50
0 0.01 0.02 0.03
0
2
4
6
8
CT
C T/C
P
Open rotorLR06?D10??0.1
LR09?D10??0.6LR13?D10??0.1
LR13?D10??0.6
0 0.01 0.02 0.03
0
2
4
6
8
CT
C T/C
P
Open rotorLR06?D20??0.1
LR09?D20??0.1LR13?D20??0.1
LR13?D20??0.8
Figure 3.9: Ratio of thrust coefficient to power coefficient vs. thrust coefficient
149
0 10 20 30 40
0
0.1
0.2
0.3
0.4
?0 [deg]
FM
Open rotorLR06?D00??0.6
LR09?D00??0.6LR13?D00??0.6
LR09?D20??0.5LR13?D20??0.6
0 10 20 30 40
0
0.1
0.2
0.3
0.4
?0 [deg]
FM
Open rotorLR13?D10??0.1?L31
LR13?D10??0.8?L31LR13?D10??1.6?L31
LR13?D10??0.1?L50
0 10 20 30 40
0
0.1
0.2
0.3
0.4
?0 [deg]
FM
Open rotorLR06?D10??0.1
LR09?D10??0.6LR13?D10??0.1
LR13?D10??0.6
0 10 20 30 40
0
0.1
0.2
0.3
0.4
?0 [deg]
FM
Open rotorLR06?D20??0.1
LR09?D20??0.1LR13?D20??0.1
LR13?D20??0.8
Figure 3.10: Figure of merit vs. collective angle
have maximum values of FM??d of 0.47 compared to 0.18 for the open rotor. This
corresponds to a 62% reduction in power for the same thrust, or a 90% increase in
thrust at the same power, when the rotor disk areas are held the same (Eq. 3.3).
These results, which take into account the real power consumption of the rotors,
show improvements in performance that are even better that those predicted by
momentum theory (see Table 2.2 on page 113), and clearly indicate the benefits of
shrouding a rotor at the MAV scale.
To determine whether similar improvements in performance would be obtained
with a more efficient rotor, which is better suited to the low-Reynolds number flight
regime, the LR13-D10-L31 shroud was tested with a rotor with cambered blades, as
150
0 0.01 0.02 0.03
0
0.1
0.2
0.3
0.4
CT
FM
Open rotorLR06?D00??0.6
LR09?D00??0.6LR13?D00??0.6
LR09?D20??0.5LR13?D20??0.6
0 0.01 0.02 0.03
0
0.1
0.2
0.3
0.4
CT
FM
Open rotorLR13?D10??0.1?L31
LR13?D10??0.8?L31LR13?D10??1.6?L31
LR13?D10??0.1?L50
0 0.01 0.02 0.03
0
0.1
0.2
0.3
0.4
CT
FM
Open rotorLR06?D10??0.1
LR09?D10??0.6LR13?D10??0.1
LR13?D10??0.6
0 0.01 0.02 0.03
0
0.1
0.2
0.3
0.4
CT
FM
Open rotorLR06?D20??0.1
LR09?D20??0.1LR13?D20??0.1
LR13?D20??0.8
Figure 3.11: Figure of merit vs. thrust coefficient
151
0 0.01 0.02 0.03
0
0.1
0.2
0.3
0.4
0.5
CT
FM ?
  s d 1/2
Open rotorLR06?D00?d 0.6
LR09?D00?d 0.6LR13?D00?d 0.6
LR09?D20?d 0.5LR13?D20?d 0.6
0 0.01 0.02 0.03
0
0.1
0.2
0.3
0.4
0.5
CT
FM ?
  s d 1/2
Open rotorLR13?D10?d 0.1?L31
LR13?D10?d 0.8?L31LR13?D10?d 1.6?L31
LR13?D10?d 0.1?L50
0 0.01 0.02 0.03
0
0.1
0.2
0.3
0.4
0.5
CT
FM ?
  s d 1/2
Open rotorLR06?D10?d 0.1
LR09?D10?d 0.6LR13?D10?d 0.1
LR13?D10?d 0.6
0 0.01 0.02 0.03
0
0.1
0.2
0.3
0.4
0.5
CT
FM ?
  s d 1/2
Open rotorLR06?D20?d 0.1
LR09?D20?d 0.1LR13?D20?d 0.1
LR13?D20?d 0.8
Figure 3.12: Generalized figure of merit vs. thrust coefficient
152
described in Chapter 2. This ?cambered? shrouded-rotor model was tested at a single
collective of 20?, and, with a tip clearance of 0.3%Dt, produced a thrust coefficient
of 0.040 and a power coefficient of 0.0075. This is equivalent to a figure of merit
of 0.507, and a generalized figure of merit of 0.543. These cambered blades were
the same blades used on the rotor that was tested by Sirohi et al. [23] during the
design of their ?TiFlyer? MAV. In those tests, the open rotor achieved a maximum
figure of merit of 0.45, equivalent to FM? = 0.32, at a collective of 18?. Although
the rotor hub, and hence the root cut-out, for the rotor in those tests was different
from that used for this shrouded rotor (R0 = 0.22R vs. 0.18R), the rotor diameters
were almost the same (6.35 in vs. 6.3 in), and the two sets of performance data
can therefore be justifiably compared. The ratio of the generalized figures of merit
thus indicates that, for this ?cambered? rotor, use of a LR13-D10-?0.3-L31 shroud
can result in a 42% increase in thrust at the same power, or a 41% reduction in
power at the same thrust. These performance improvements are not as great as
those achieved with the baseline rotor?s uncambered blades, which suggests that as
the design of the rotor itself is improved3, less of a performance benefit would be
obtained from shrouding the rotor. However, further tests would be necessary to
determine whether this is true.
The thrust coefficients referred to in the preceding discussions have been the
?total? thrust coefficient, as measured by the thrust transducer of the hover test-
stand. The shroud surface pressure measurements, described in greater detail in
3Figures of merit as high as 0.622 have been achived by MAV-scale open rotors that incorporated
taper, twist, camber and sharpened leading edges in their design [20].
153
Section 3.4 of this chapter, were integrated to obtain the thrust produced by the
shrouds themselves, and these were subtracted from the total thrust measurements
to obtain the thrust of the rotor alone. The variations in CTrotor and CTshroud thus
obtained are shown in Figures 3.13?3.15, plotted as functions of blade collective
angle. Figure 3.13 indicates that, at a given collective, the rotor does generate
less thrust when it is placed inside a shroud, compared to when it is in open air.
This would be due to the increased inflow through the rotor, which reduces the
blade section angles of attack at a fixed collective. At higher collectives, however,
the difference between CTrotor and CTOR becomes smaller, with the two quantities
becoming almost equal in some cases. Note that the shrouded-rotor coefficients
plotted in this figure were obtained by normalizing the rotor thrust by the shroud
throat cross-sectional area, while the open rotor thrust coefficients were obtained by
normalizing by the smaller area of the actual rotor disk. For a more fair comparison,
Figure 3.14 shows the values of CTrotor obtained when the shrouded-rotor thrusts
are normalized by the actual rotor disk area, too. As expected, the coefficient
values obtained are now slightly higher than those in Figure 3.13 ? more so for the
configurations with the larger tip gaps ? but are still mostly lower than those for
the open rotor. The highest thrust coefficients achieved by the rotors, whether open
or shrouded, are around 0.02; this corresponds to a blade loading coefficient (CT/?)
of 0.167, or a mean rotor lift coefficient ( ?CL = 6CT/?) of 1.0, which would be at
the upper stall limits of the blunt, uncambered, flat-plate airfoils (Table 2.3, p. 115)
at these low Reynolds numbers, and is clearly evidenced by the the flattening-out
of the curves at the highest collectives. Figure 3.15 shows that most of the shroud
154
0 10 20 30 40
0
0.01
0.02
?0 [deg]
C T rotor
Open rotorLR06?D00??0.6
LR09?D00??0.6LR13?D20??0.6
0 10 20 30 40
0
0.01
0.02
?0 [deg]
C T rotor
Open rotorLR13?D10??0.1?L31
LR13?D10??0.8?L31LR13?D10??1.6?L31
LR13?D10??0.1?L50
0 10 20 30 40
0
0.01
0.02
?0 [deg]
C T rotor
Open rotorLR13?D10??0.1
LR13?D10??0.6
0 10 20 30 40
0
0.01
0.02
?0 [deg]
C T rotor
Open rotorLR06?D20??0.1
LR13?D20??0.1LR13?D20??0.8
Figure 3.13: Rotor thrust coefficient, as normalized by shroud thoat area (piD2t/4),
vs. collective angle
models achieve their maximum thrust at a collective angle between 25? and 30?, and
exhibit a gentle stall-like characteristic at higher collectives, while some ? the LR13
models with tip clearances of 0.6% and 0.8%Dt ? initially level off or decrease in
that particular collective range, but then increase to higher values at ?0 = 40?.
3.3.2 Effects of changing shroud parameter values
The effects of varying each of the shroud geometric parameters ? blade tip
clearance, inlet lip radius, diffuser included angle, diffuser length and diffuser ex-
pansion ratio (Figure 3.16) ? on the performance of the shrouded rotors are shown
155
0 10 20 30 40
0
0.01
0.02
?0 [deg]
C T rotor
Open rotorLR06?D00??0.6
LR09?D00??0.6LR13?D20??0.6
0 10 20 30 40
0
0.01
0.02
?0 [deg]
C T rotor
Open rotorLR13?D10??0.1?L31
LR13?D10??0.8?L31LR13?D10??1.6?L31
LR13?D10??0.1?L50
0 10 20 30 40
0
0.01
0.02
?0 [deg]
C T rotor
Open rotorLR13?D10??0.1
LR13?D10??0.6
0 10 20 30 40
0
0.01
0.02
?0 [deg]
C T rotor
Open rotorLR06?D20??0.1
LR13?D20??0.1LR13?D20??0.8
Figure 3.14: Rotor thrust coefficient, as normalized by rotor disk area (piD2/4), vs.
collective angle
156
0 10 20 30 400
0.004
0.008
0.012
0.016
?0 [deg]
C T shroud
LR06?D00??0.6LR09?D00??0.6
LR13?D20??0.6
0 10 20 30 400
0.004
0.008
0.012
0.016
?0 [deg]
C T shroud
LR13?D10??0.1?L31LR13?D10??0.8?L31
LR13?D10??1.6?L31LR13?D10??0.1?L50
0 10 20 30 400
0.004
0.008
0.012
0.016
?0 [deg]
C T shroud
LR13?D10??0.1LR13?D10??0.6
0 10 20 30 400
0.004
0.008
0.012
0.016
?0 [deg]
C T shroud
LR06?D20??0.1LR13?D20??0.1
LR13?D20??0.8
Figure 3.15: Shroud thrust coefficient vs. collective angle
157
in Figures 3.17?3.20. As mentioned earlier in Chapter 2, the matrix of shrouded-
rotor models (Table 2.1, p. 112) enabled comparisons of data from multiple series of
models for each of the shroud parameters. These series are listed below in Table 3.1.
For each parameter, four different metrics have been used to illustrate the improve-
ment in performance of the shrouded rotors over the open rotor, and to compare
the performances of the different shrouded rotors. These are:
1. The ratio of the maximum value of CT/CP of the shrouded rotor to that of
the open rotor.
2. The ratio of the maximum value of the generalized figure of merit, FM? =
FM??d, of the shrouded rotor to that of the open rotor.
3. The ratio of the shrouded rotor?s thrust coefficient (CTSR) to that of the open
rotor (CTOR), with the two configurations at the same power coefficient, and
evaluated at four different values of CP: 0.002, 0.004, 0.007 and 0.010.
4. The ratio of the shrouded rotor?s power coefficient (CPSR) to that of the open
rotor (CPOR), at the same thrust coefficient, and evaluated at three different
values of CT: 0.010, 0.015 and 0.020.
The figures show the values of each of these metrics plotted as functions of the shroud
parameters, for each of the different comparison series. The values of CP and CT at
which the ratios of the thrust and power coefficients, respectively, were evaluated
were chosen such that they spanned the range of measured power and thrust for all
of the models. Although thrust coefficients higher than 0.020 were measured for the
158
Figure 3.16: Principal shroud parameters affecting shrouded-rotor performance:
diffuser included angle (?d), diffuser length (Ld), inlet lip radius (rlip), blade tip
clearance (?tip)
shrouded rotors, this was the highest value of CT achieved by the open rotor, and
hence the highest value at which a comparison of the power consumption could be
made.
Additionally, to illustrate the physics of the operation of the shrouded ro-
tors, the effects of the shroud parameters on the ratios CTSR/CTOR, CPSR/CPOR,
CTrotor/CTOR (rotor thrust) andCTrotor/CTtotal (rotor thrust fraction) have been shown,
with the quantities for the open and shrouded rotors compared at the same collec-
tive angle and evaluated at four different values of ?0: 10?, 20?, 30? and 40?. Since
the performance measurements were made at these specific collective angles, cal-
culation of the corresponding ratios was straightforward. For comparisons at the
same power or thrust coefficients, however, interpolation between the measured val-
ues was required; this was performed using piecewise-cubic Hermite interpolating
polynomials. Note that, unlike the ratios comparing the open and shrouded rotors,
159
Table 3.1: Shrouded-rotor model comparison series for analyzing effects of the
shroud geometric parameters on performance characteristics
Model series for blade tip clearance ?tip values [%Dt]
LR09-D20 0.09, 0.50
LR13-D20 0.09, 0.63, 0.82
LR13-D10 0.09, 0.63
LR13-D10-L31 0.09, 0.82, 1.57
Model series for inlet lip radius rlip values [%Dt]
D10-?0.1 6.48, 13.1
D20-?0.1 6.48, 8.99, 13.1
D00-?0.6 6.48, 8.99, 13.1
D10-?0.6 8.99, 13.1
D20-?0.6 8.99, 13.1
Model series for diffuser included angle ?d values [?]
LR06-?0.1 9.70, 20.22
LR13-?0.1 9.70, 20.22
LR09-?0.6 0, 9.70, 20.22
LR13-?0.6 0, 9.70, 20.22
Model series for diffuser length Ld values [%Dt]
LR13-D10-?0.1 31, 50, 72
160
Table 3.1: Shrouded-rotor model comparison series (contd.)
Model series for diffuser expansion ratio ?d values
LR06-?0.1-L72 1.31, 1.63
LR13-?0.1-L72 1.31, 1.63
LR13-?0.1-D10 1.14, 1.22, 1.31
LR09-?0.6-L72 1.03, 1.31, 1.63
LR13-?0.6-L72 1.03, 1.31, 1.63
LR13-?0.8 1.15, 1.63
the rotor thrust fraction of a shrouded rotor, Trotor/Ttotal, is, according to the mo-
mentum theory/actuator disk model, independent of thrust or power level ? or,
equivalently, of rotor collective angle ? and depends only on the wake expansion
ratio (Equation 1.5 on page 17).
Some of the observations that can be made from these figures are the same,
no matter which shroud parameter is being considered, and these are therefore de-
scribed first, before discussing the specific effects of changing the shroud parameter
values. When compared at the same power coefficient, the ratio of total thrust,
CTSR/CTOR, shows values between 1.3 and almost 2.0, with the ratio increasing
slightly from CP = 0.002 to 0.004 and then decreasing as CP is further increased
to 0.010 (Figs. 3.17c, 3.18c, 3.19c and 3.20c). For all of the shrouded rotors, the
CP value of 0.004 corresponds most closely to a collective angle of 20?, which is
the collective at which they all achieved their highest value of [generalized] figure
of merit. It is therefore not entirely surprising that highest values of CTSR/CTOR are
also achieved at this value of power coefficient. When compared at the same collec-
161
tive angle, the ratio increases with increasing collective, with most of the increase
occurring between ?0 = 10? and 20? (Figs. 3.17e, 3.18e, 3.19e and 3.20e). At ?0 =
10?, the ratio has values of 0.8?1.0, while at the higher collectives, this increases to
between 1.2 and 1.6.
Similarly, when compared at the same thrust coefficient, the ratio of the power
coefficients, CPSR/CPOR, steadily decreases with increasing CT, with values between
0.3 and 0.7 (Figs. 3.17d, 3.18d, 3.19d and 3.20d), but, when compared at the same
collective (Figs. 3.17f, 3.18f, 3.19f and 3.20f), steadily increases with increasing ?0,
showing values between 0.5 and 1.0. The reason for these opposite trends, seen in
both CTSR/CTOR and CPSR/CPOR, is simply due to the differences in the ways that
CT and CP increase with increasing collective for the open and shrouded rotors
(Figures 3.4 and 3.5). Considering, for example, the behavior of CTSR/CTOR: at
low collectives, the shrouded-rotor thrust is higher than that of the open rotor,
while the power consumed is lower, thereby exaggerating the differences in thrust
when compared at the same power. At higher collectives, the difference between the
shrouded-rotor thrust and the open-rotor thrust becomes much larger, soCTSR/CTOR
at the same collective increases, but the relative difference in power consumption
becomes smaller, so CTSR/CTOR at the same power level decreases. The behavior of
the ratio of power coefficients can be explained similarly.
Looking at the thrust produced by the rotor itself, in the shrouded condition
versus in the open condition, CTrotor/CTOR ? when compared at the same collective
? is seen to steadily increase with increasing collective, with values between 0.3 and
1.0 (Figs. 3.17g, 3.18g, 3.19g and 3.20g). The rotor thrust fraction, Trotor/Ttotal, also
162
mostly increases with increasing ?0, with values between 0.3 and 0.75, depending
on the shroud model (Figs. 3.17h, 3.18h, 3.19h and 3.20h). Both of these results
suggest that at the higher collectives, the shroud is not able to off-load the rotor as
much as at lower collectives, and the rotor is, so to speak, ?left to fend for itself.?
This conclusion is supported by the stall-like behavior seen in CTshroud for many of
the shroud models (Figure 3.15), while CTrotor continues to increase at the higher
collectives, albeit at a decreased rate (Figure 3.13).
3.3.2.1 Blade tip clearance
The effects of varying the blade tip clearance, ?tip, on the shrouded-rotor per-
formance are shown in Figures 3.17a?h. For this parameter, data were available
from four series of shrouded-rotor models, LR09-D20, LR13-D20, LR13-D10 and
LR13-D10-L31, in each of which lip radius, diffuser angle and diffuser length were
held constant, and the tip clearance allowed to change from one model to the next
(Table 3.1).
Increasing the tip clearance is clearly seen to have adverse effects on perfor-
mance, causing decreases in maximum CT/CP, maximum FM? and CTSR/CTOR at
the same CP, and increases in CPSR/CPOR at the same CT. Of the four series, the
LR09-D20 models show the most rapid degradations. The LR13-D10 models, while
exhibiting the best overall performance, show the least change in performance with
changing tip clearance. In fact, maximum CT/CP increases in this series, as ?tip is
increased from 0.1 to 0.6%Dt; however, this is because measurements were made
163
  0 0.4 0.8 1.2 1.61.3
1.5
1.7
1.9
?tip [%Dt]
(C T/C
P) max, SR
 / (C
T/C
P) max, OR
LR09?D20LR13?D20
LR13?D10LR13?D10?L31
(a) Effect on CT/CP
  0 0.4 0.8 1.2 1.61.4
1.8
2.2
2.6
?tip [%Dt]
(FM 
? d 1/2
) max, SR
 / (FM 
? d 1/2
) max, OR
LR09?D20LR13?D20
LR13?D10LR13?D10?L31
(b) Effect on FM??d
Figure 3.17: Effect of blade tip clearance
at ?0 = 15? for the ?0.6 model but not for the ?0.1 model, as explained earlier.
Comparing the LR09-D20 series and the LR13-D20 series shows that changing ?tip
has a stronger effect on the shroud with the smaller lip radius. Similarly, comparing
the LR13-D10 and LR13-D20 series shows more pronounced changes for the D20
models, and comparing the LR13-D10 and LR13-D10-L31 series shows stronger ef-
fects for the shorter diffuser. As will be shown in subsequent sections, these are the
cases in which the shrouded rotors generally show the worst performance, and it is
in these that changing the tip clearance seems to have the greatest effect.
When compared at the same collective, CTSR/CTOR is seen to decrease with in-
creasing tip clearance, but the LR13-D20 series of models shows anomalous behavior.
No consistent trends can be seen in the behavior of CPSR/CPOR or of CTrotor/CTOR.
The rotor thrust fraction, Trotor/Ttotal, generally seems to increase with increasing
tip clearance, but here, too, some of the data go against this trend.
164
  0 0.4 0.8 1.2 1.61.2
1.4
1.6
1.8
2
?tip [%Dt]
C T SR
 / C T
OR
CP = 0.002
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
CP = 0.004
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
CP = 0.007
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
CP = 0.01
LR09?D20LR13?D20
LR13?D10LR13?D10?L31
(c) Effect on total thrust, compared at the same power
  0 0.4 0.8 1.2 1.60.3
0.4
0.5
0.6
0.7
?tip [%Dt]
C P SR
 / C P
OR
CT = 0.01
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
CT = 0.015
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
CT = 0.02
LR09?D20LR13?D20
LR13?D10LR13?D10?L31
(d) Effect on power, compared at the same total thrust
  0 0.4 0.8 1.2 1.6
0.8
1
1.2
1.4
1.6
?tip [%Dt]
C T SR
 / C T
OR
?0 = 10
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 20
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 30
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 40
LR09?D20LR13?D20
LR13?D10LR13?D10?L31
(e) Effect on total thrust, compared at the same collective angle
Figure 3.17: Effect of blade tip clearance (contd.)
165
  0 0.4 0.8 1.2 1.60.5
0.6
0.7
0.8
0.9
1
?tip [%Dt]
C P SR
 / C P
OR
?0 = 10
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 20
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 30
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 40
LR09?D20LR13?D20
LR13?D10LR13?D10?L31
(f) Effect on power, compared at the same collective angle
  0 0.4 0.8 1.2 1.60.3
0.4
0.5
0.6
0.7
0.8
0.9
1
?tip [%Dt]
C T rotor
 / C T
OR
?0 = 10
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 20
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 30
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 40
LR13?D20LR13?D10
LR13?D10?L31
(g) Effect on rotor thrust, compared at the same collective angle
  0 0.4 0.8 1.2 1.60.3
0.35
0.4
0.45
0.5
0.55
0.6
0.65
?tip [%Dt]
T rotor
 / T total
?0 = 10
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 20
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 30
  0 0.4 0.8 1.2 1.6?
tip [%Dt]
?0 = 40
LR13?D20LR13?D10
LR13?D10?L31
(h) Effect on rotor thrust fraction, compared at the same collective angle
Figure 3.17: Effect of blade tip clearance (contd.)
166
6.5   9  131.3
1.5
1.7
1.9
rlip [%Dt]
(C T/C
P) max, SR
 / (C
T/C
P) max, OR
D00??0.6D10??0.1
D10??0.6D20??0.1
D20??0.6
(a) Effect on CT/CP
6.5   9  131.4
1.8
2.2
2.6
rlip [%Dt]
(FM 
? d 1/2
) max, SR
 / (FM 
? d 1/2
) max, OR
D00??0.6D10??0.1
D10??0.6D20??0.1
D20??0.6
(b) Effect on FM??d
Figure 3.18: Effect of lip radius
3.3.2.2 Inlet lip radius
The effects of varying the inlet lip radius, rlip, on the shrouded-rotor perfor-
mance are shown in Figures 3.18a?h. For this parameter, data were available from
five series of shrouded-rotor models: D10-?0.1, D20-?0.1, D00-?0.6, D10-?0.6 and
D20-?0.6 (Table 3.1).
As lip radius is increased, all five series of models show increases in maximum
CT/CP, maximum FM? andCTSR/CTOR at the sameCP. CTSR/CTOR when compared
at the same collective, also shows increases with increasing rlip at collectives of 20?
and higher, but these are not as pronounced as those seen when compared at the
same power. Except for the D00-?06 series of models, and anomalous data points
corresponding to the LR09-D20-?0.1 model, the rates of increase with increasing
lip radius are about the same for all the series. The ratio of power coefficients,
CPSR/CPOR at the same CT, decreases with increasing lip radius, with the effects
being seen more clearly at the two higher thrust coefficient values than at the lowest
167
6.5   9  131.2
1.4
1.6
1.8
2
rlip [%Dt]
C T SR
 / C T
OR
CP = 0.002
6.5   9  13r
lip [%Dt]
CP = 0.004
6.5   9  13r
lip [%Dt]
CP = 0.007
6.5   9  13r
lip [%Dt]
CP = 0.01
D00??0.6D10??0.1
D10??0.6D20??0.1
D20??0.6
(c) Effect on total thrust, compared at the same power
6.5   9  130.3
0.4
0.5
0.6
0.7
rlip [%Dt]
C P SR
 / C P
OR
CT = 0.01
6.5   9  13r
lip [%Dt]
CT = 0.015
6.5   9  13r
lip [%Dt]
CT = 0.02
D00??0.6D10??0.1
D10??0.6D20??0.1
D20??0.6
(d) Effect on power, compared at the same total thrust
6.5   9  13
0.8
1
1.2
1.4
1.6
rlip [%Dt]
C T SR
 / C T
OR
?0 = 10
6.5   9  13r
lip [%Dt]
?0 = 20
6.5   9  13r
lip [%Dt]
?0 = 30
6.5   9  13r
lip [%Dt]
?0 = 40
D00??0.6D10??0.1
D10??0.6D20??0.1
D20??0.6
(e) Effect on total thrust, compared at the same collective angle
Figure 3.18: Effect of lip radius (contd.)
168
6.5   9  13
0.5
0.6
0.7
0.8
0.9
1
rlip [%Dt]
C P SR
 / C P
OR
?0 = 10
6.5   9  13r
lip [%Dt]
?0 = 20
6.5   9  13r
lip [%Dt]
?0 = 30
6.5   9  13r
lip [%Dt]
?0 = 40
D00??0.6D10??0.1
D10??0.6D20??0.1
D20??0.6
(f) Effect on power, compared at the same collective angle
6.5   9  130.5
0.6
0.7
0.8
0.9
1
rlip [%Dt]
C T rotor
 / C T
OR
?0 = 10
6.5   9  13r
lip [%Dt]
?0 = 20
6.5   9  13r
lip [%Dt]
?0 = 30
6.5   9  13r
lip [%Dt]
?0 = 40
D00??0.6D20??0.1
(g) Effect on rotor thrust, compared at the same collective angle
6.5   9  130.45
0.5
0.55
0.6
0.65
0.7
0.75
rlip [%Dt]
T rotor
 / T total
?0 = 10
6.5   9  13r
lip [%Dt]
?0 = 20
6.5   9  13r
lip [%Dt]
?0 = 30
6.5   9  13r
lip [%Dt]
?0 = 40
D00??0.6D20??0.1
(h) Effect on rotor thrust fraction, compared at the same collective angle
Figure 3.18: Effect of lip radius (contd.)
169
 0 10 201.3
1.5
1.7
1.9
?d [deg]
(C T/C
P) max, SR
 / (C
T/C
P) max, OR
LR06??0.1LR09??0.6
LR13??0.1LR13??0.6
(a) Effect on CT/CP
 0 10 201.4
1.8
2.2
2.6
?d [deg]
(FM 
? d 1/2
) max, SR
 / (FM 
? d 1/2
) max, OR
LR06??0.1LR09??0.6
LR13??0.1LR13??0.6
(b) Effect on FM??d
Figure 3.19: Effect of diffuser angle
one.
Increasingrlip alsocausessharpreductionsintherotorthrustfraction,Trotor/Ttotal,
and generally decreasing trends in CTrotor/CTOR indicating increasing off-loading of
the rotor by the shroud. As with the blade tip clearance, changing the lip radius
does not appear to have a consistent effect on CPSR/CPOR when compared at the
same collective.
3.3.2.3 Diffuser angle
Theeffectsofvaryingthediffuserincludedangle,?d, areshowninFigures3.19a?
h. For this parameter, data were available from four series of shrouded-rotor models:
LR06-?0.1, LR13-?0.1, LR09-?0.6 and LR13-?0.6 (Table 3.1).
The figures show that, of the three diffuser angles tested, the ten-degree dif-
fusers showed the best performance, whether the measure of performance improve-
ment over the open rotor is maximum CT/CP, maximum FM? or CTSR/CTOR at the
170
 0 10 201.2
1.4
1.6
1.8
2
?d [deg]
C T SR
 / C T
OR
CP = 0.002
 0 10 20?
d [deg]
CP = 0.004
 0 10 20?
d [deg]
CP = 0.007
 0 10 20?
d [deg]
CP = 0.01
LR06??0.1LR09??0.6
LR13??0.1LR13??0.6
(c) Effect on total thrust, compared at the same power
 0 10 200.3
0.4
0.5
0.6
0.7
?d [deg]
C P SR
 / C P
OR
CT = 0.01
 0 10 20?
d [deg]
CT = 0.015
 0 10 20?
d [deg]
CT = 0.02
LR06??0.1LR09??0.6
LR13??0.1LR13??0.6
(d) Effect on power, compared at the same total thrust
 0 10 20
0.8
1
1.2
1.4
1.6
?d [deg]
C T SR
 / C T
OR
?0 = 10
 0 10 20?
d [deg]
?0 = 20
 0 10 20?
d [deg]
?0 = 30
 0 10 20?
d [deg]
?0 = 40
LR06??0.1LR09??0.6
LR13??0.1LR13??0.6
(e) Effect on total thrust, compared at the same collective angle
Figure 3.19: Effect of diffuser angle (contd.)
171
 0 10 20
0.5
0.6
0.7
0.8
0.9
1
?d [deg]
C P SR
 / C P
OR
?0 = 10
 0 10 20?
d [deg]
?0 = 20
 0 10 20?
d [deg]
?0 = 30
 0 10 20?
d [deg]
?0 = 40
LR06??0.1LR09??0.6
LR13??0.1LR13??0.6
(f) Effect on power, compared at the same collective angle
 0 10 20
0.5
0.6
0.7
0.8
0.9
1
?d [deg]
C T rotor
 / C T
OR
?0 = 10
 0 10 20?
d [deg]
?0 = 20
 0 10 20?
d [deg]
?0 = 30
 0 10 20?
d [deg]
?0 = 40
LR13??0.1LR13??0.6
(g) Effect on rotor thrust, compared at the same collective angle
 0 10 200.45
0.5
0.55
0.6
0.65
0.7
?d [deg]
T rotor
 / T total
?0 = 10
 0 10 20?
d [deg]
?0 = 20
 0 10 20?
d [deg]
?0 = 30
 0 10 20?
d [deg]
?0 = 40
LR13??0.1LR13??0.6
(h) Effect on rotor thrust fraction, compared at the same collective angle
Figure 3.19: Effect of diffuser angle (contd.)
172
same CP. Reducing the diffuser angle to 0? causes a deterioration in performance,
while increasing it to 20? causes an even greater deterioration. The same conclusions
can be drawn from the variations seen in CPSR/CPOR at the same CT, which show a
minimum at ?d = 10?; here too, as in the case with changing lip radius, the effect is
seen more clearly at the higher thrust coefficients.
Compared at the same collective angle, the ?0.6 models again show minimum
power consumption at ?d = 10?, while the ?0.1 models show slightly decreasing
or unchanging power consumption when ?d is increased from 10? to 20?. On the
other hand, the ?0.1 models show a decrease in CTSR/CTOR (at the same ?0) when
?d is increased from 10? to 20?, while the ?0.6 models show erratic behavior for this
quantity. Increasing ?d from 10? to 20? also causes CTrotor/CTOR and Trotor/Ttotal to
generally increase, indicating a decrease in off-loading by the shroud.
3.3.2.4 Diffuser length
The effects of varying the diffuser length, Ld, are shown in Figures 3.20a?h.
For this parameter, data were available from one series of shrouded-rotor models:
the LR13-D10-?0.1 models, with diffuser lengths of 31%, 50% and 72% of the shroud
throat diameter.
The data for this parameter show fairly clear improvements in performance
? increases in thrust or decreases in power consumption ? when diffuser length
is increased from 31%Dt to 50%Dt, but less clear results with the further increase
to 72%Dt: power consumption either stays roughly the same or increases slightly,
173
31 50 721.76
1.8
1.84
1.88
Ld [%Dt]
(C T/C
P) max, SR
 / (C
T/C
P) max, OR
LR13?D10??0.1
(a) Effect on CT/CP
31 50 722.4
2.5
2.6
Ld [%Dt]
(FM 
? d 1/2
) max, SR
 / (FM 
? d 1/2
) max, OR
LR13?D10??0.1
(b) Effect on FM??d
Figure 3.20: Effect of diffuser length
while the thrust produced stays about the same, increases slightly, or even decreases
slightly in some cases. The behavior of maximumCT/CP seems anomalous, showing
a steady decrease as Ld is increased from 31%Dt to 72%Dt; however, this is only
because the L31 model had measurements made at ?0 = 15? while the the L50 and
L72 models did not. If, instead, the value of CT/CP at a collective of 10? is used
for the L31 model in the comparison (Figure 3.8), as is the case for the other two
models, then the variation obtained in ?maximum? CT/CP with increasing diffuser
length is similar to that seen for maximum FM?. On the other hand, performance
improvments are generally accompanied by an increase in off-loading of the rotor
by the shroud; for this shroud parameter, CTrotor/CTOR and Trotor/Ttotal are seen
to generally increase with increasing diffuser length, indicating a decrease in the
off-loading. The reasons for this behavior is unknown.
174
31 50 721.4
1.6
1.8
2
Ld [%Dt]
C T SR
 / C T
OR
CP = 0.002
31 50 72L
d [%Dt]
CP = 0.004
31 50 72L
d [%Dt]
CP = 0.007
31 50 72L
d [%Dt]
CP = 0.01
LR13?D10??0.1
(c) Effect on total thrust, compared at the same power
31 50 720.3
0.35
0.4
0.45
0.5
Ld [%Dt]
C P SR
 / C P
OR
CT = 0.01
31 50 72L
d [%Dt]
CT = 0.015
31 50 72L
d [%Dt]
CT = 0.02
LR13?D10??0.1
(d) Effect on power, compared at the same total thrust
31 50 721
1.2
1.4
1.6
Ld [%Dt]
C T SR
 / C T
OR
?0 = 10
31 50 72L
d [%Dt]
?0 = 20
31 50 72L
d [%Dt]
?0 = 30
31 50 72L
d [%Dt]
?0 = 40
LR13?D10??0.1
(e) Effect on total thrust, compared at the same collective angle
Figure 3.20: Effect of diffuser length (contd.)
175
31 50 720.5
0.6
0.7
0.8
Ld [%Dt]
C P SR
 / C P
OR
?0 = 10
31 50 72L
d [%Dt]
?0 = 20
31 50 72L
d [%Dt]
?0 = 30
31 50 72L
d [%Dt]
?0 = 40
LR13?D10??0.1
(f) Effect on power, compared at the same collective angle
31 50 720.3
0.4
0.5
0.6
0.7
0.8
0.9
Ld [%Dt]
C T rotor
 / C T
OR
?0 = 10
31 50 72L
d [%Dt]
?0 = 20
31 50 72L
d [%Dt]
?0 = 30
31 50 72L
d [%Dt]
?0 = 40
LR13?D10??0.1
(g) Effect on rotor thrust, compared at the same collective angle
31 50 720.3
0.35
0.4
0.45
0.5
0.55
Ld [%Dt]
T rotor
 / T total
?0 = 10
31 50 72L
d [%Dt]
?0 = 20
31 50 72L
d [%Dt]
?0 = 30
31 50 72L
d [%Dt]
?0 = 40
LR13?D10??0.1
(h) Effect on rotor thrust fraction, compared at the same collective angle
Figure 3.20: Effect of diffuser length (contd.)
176
  1 1.2 1.4 1.6 1.81.3
1.5
1.7
1.9
?d
(C T/C
P) max, SR
 / (C
T/C
P) max, OR
LR06??0.1?L72LR09??0.6?L72
LR13??0.1?L72LR13??0.6?L72
LR13??0.1?D10LR13??0.8
(a) Effect on CT/CP
  1 1.2 1.4 1.6 1.81.4
1.8
2.2
2.6
?d
(FM 
? d 1/2
) max, SR
 / (FM 
? d 1/2
) max, OR
LR06??0.1?L72LR09??0.6?L72
LR13??0.1?L72LR13??0.6?L72
LR13??0.1?D10LR13??0.8
(b) Effect on FM??d
Figure 3.21: Effect of diffuser expansion ratio
3.3.2.5 Diffuser expansion ratio
The diffuser expansion ratio, ?d, is a function of both the diffuser length and
the included angle ? ??d =
bracketleftBig
1 + 2LdDt tanparenleftbig?d2parenrightbig
bracketrightBig2
? and its effects have therefore been
described in the previous two sections devoted to these two parameters. However, in
order to compare the relative effects of Ld and ?d, the data from those two sections
are shown together in Figures 3.21a?h, this time explicitly plotted as functions of?d.
The data shown are therefore from four series of models in which the diffuser angle
is varied, one series in which the diffuser length is varied, and one additional series
? the LR13-?0.8 models ? in which both angle and length change (Table 3.1).
The superior performance of the 10-degree diffusers, compared to the 0-degree
and 20-degree diffusers of the same length, is seen in these figures as well, at the
corresponding expansion ratio of 1.31. However, by now considering the effect of
changing ?d by changing the diffuser length, it appears that better performance
may be obtained at the slightly lower ratio of 1.22. Further reductions in ?d caused
177
  1 1.2 1.4 1.6 1.81.2
1.4
1.6
1.8
2
?d
C T SR
 / C T
OR
CP = 0.002
  1 1.2 1.4 1.6 1.8?
d
CP = 0.004
  1 1.2 1.4 1.6 1.8?
d
CP = 0.007
  1 1.2 1.4 1.6 1.8?
d
CP = 0.01
LR06??0.1?L72LR09??0.6?L72
LR13??0.1?L72LR13??0.6?L72
LR13??0.1?D10LR13??0.8
Momentum Theoryprediction (ratio at 
same ideal power) 
(c) Effect on total thrust, compared at the same power
  1 1.2 1.4 1.6 1.80.3
0.4
0.5
0.6
0.7
?d
C P SR
 / C P
OR
CT = 0.01
  1 1.2 1.4 1.6 1.8?
d
CT = 0.015
  1 1.2 1.4 1.6 1.8?
d
CT = 0.02
LR06??0.1?L72LR09??0.6?L72
LR13??0.1?L72LR13??0.6?L72
LR13??0.1?D10LR13??0.8
Momentum Theory prediction(ratio of ideal power) 
(d) Effect on power, compared at the same total thrust
  1 1.2 1.4 1.6 1.8
0.8
1
1.2
1.4
1.6
?d
C T SR
 / C T
OR
?0 = 10
  1 1.2 1.4 1.6 1.8?
d
?0 = 20
  1 1.2 1.4 1.6 1.8?
d
?0 = 30
  1 1.2 1.4 1.6 1.8?
d
?0 = 40
LR06??0.1?L72LR09??0.6?L72
LR13??0.1?L72LR13??0.6?L72
LR13??0.1?D10LR13??0.8
(e) Effect on total thrust, compared at the same collective angle
Figure 3.21: Effect of diffuser expansion ratio (contd.)
178
  1 1.2 1.4 1.6 1.8
0.5
0.6
0.7
0.8
0.9
1
?d
C P SR
 / C P
OR
?0 = 10
  1 1.2 1.4 1.6 1.8?
d
?0 = 20
  1 1.2 1.4 1.6 1.8?
d
?0 = 30
  1 1.2 1.4 1.6 1.8?
d
?0 = 40
LR06??0.1?L72LR09??0.6?L72
LR13??0.1?L72LR13??0.6?L72
LR13??0.1?D10LR13??0.8
(f) Effect on power, compared at the same collective angle
  1 1.2 1.4 1.6 1.80.3
0.4
0.5
0.6
0.7
0.8
0.9
1
?d
C T rotor
 / C T
OR
?0 = 10
  1 1.2 1.4 1.6 1.8?
d
?0 = 20
  1 1.2 1.4 1.6 1.8?
d
?0 = 30
  1 1.2 1.4 1.6 1.8?
d
?0 = 40
LR13??0.1?L72LR13??0.6?L72
LR13??0.1?D10LR13??0.8
(g) Effect on rotor thrust, compared at the same collective angle
  1 1.2 1.4 1.6 1.80.3
0.35
0.4
0.45
0.5
0.55
0.6
0.65
?d
T rotor
 / T total
?0 = 10
  1 1.2 1.4 1.6 1.8?
d
?0 = 20
  1 1.2 1.4 1.6 1.8?
d
?0 = 30
  1 1.2 1.4 1.6 1.8?
d
?0 = 40
LR13??0.1?L72LR13??0.6?L72
LR13??0.1?D10LR13??0.8
Momentum Theoryprediction 
(h) Effect on rotor thrust fraction, compared at the same collective angle
Figure 3.21: Effect of diffuser expansion ratio (contd.)
179
performance to degrade, with the deterioration due to reducing Ld occuring at the
greater rate than that due to reducing ?d. At higher power coefficients, the shrouds
appear to be slightly better able to further expand the flow to higher ?d; however,
the improvement is marginal, and occurs at collectives higher than those at which
maximum power loading is achieved.
Superimposed on Figures 3.21c and 3.21d are the predictions of momentum
theory for the thrust and power ratios as functions of ?d (Equations A.27 and A.28),
but considering only the ideal power consumption. It is clearly seen that, by consid-
ering the actual power requirements of the open and shrouded rotors, the improve-
ments obtained by shrouding are substantially greater than those predicted by the
theory. Only at the highest expansion ratios tested, at which the diffusers were not
able to fully explan the flow, did the performance deteriorate to worse than that
predicted by the theory.
Figure3.21h shows the observed variations with expansion ratio of the rotor
thrust fraction, Trotor/Ttotal, also compared with the momentum-theory prediction.
Unlike the total thrust and power, which, at low values of ?d, do initially show the
same increasing or decreasing trends with increasing expansion ratio that the theory
predicts, the observed thrust fraction always increases with increasing ?d, over the
entire range of expansion ratios tested, in complete contrast to the prediction that
it should decrease with increasing ?d (Equation A.17). The reason for this behavior
is not known. Of the two diffuser parameters, changing the length seems to have a
stronger effect on Trotor/Ttotal than does changing the diffuser angle.
180
3.3.2.6 Relative effects of the different parameters
The preceding discussions have shown that, in general, the shrouded-rotor
models showed deteriorating performance when the inlet lip radius was decreased,
the blade tip clearance increased, or the diffuser expansion ratio increased or de-
creased from a certain optimum value. It is of interest to the MAV designer to
know, quantitatively, how much the performance of a shrouded rotor will be affected
by changing the value of one or more of these shroud parameters, and which of the
parameters have the strongest influence on the performance. Therefore, linear fits
were applied to the variations in three of the four performance-improvement metrics
? the shrouded-rotor-to-open-rotor ratios of maximum CT/CP, of maximum FM?,
and of thrust coefficient at a power coefficient of 0.004, which was the CP at which
the greatest gains in thrust over the open rotor were seen ? for each of the four
shrouded-rotor parameters. The results are tabulated in Table 3.2, in the form of
the amount of increase (?+?) or decrease (???) in each of the metrics per unit change
in the value of each shroud parameter.
The numbers showthattheeffectofchangingbladetip clearance becomes more
pronounced as the other parameters are changed to degrade the shrouded-rotor?s
performance, i.e., decreasing diffuser length, increasing diffuser angle (from 10? to
20?) or decreasing inlet lip radius. The worsening effect of increasing ?d from 10? to
20? is exacerbated by larger tip clearances, but is less affected by changing the lip
radius. On the other hand, decreasing?d from 10? to 0? has a more pronounced effect
on the LR13 models than on the LR09 models. ChangingLd seems to have a stronger
181
effect than that of changing ?d for decreasing ?d, but a weaker effect for increasing
?d. In contrast to the trends seen for ?tip and ?d, the effect of changing the inlet lip
radius seems to be fairly independent of the influence of the other shroud parameters
? with the sole exception of the D00-?0.6 models. Thus, these data suggest that,
as the shroud geometry is changed from some optimal configuration, the effects of
changing any one of the geometric parameters becomes more pronounced. In other
words, it is not possible to categorically state that any one particular parameter has
a greater effect on performance than any other, since the amount of influence of any
parameter depends on the values of the other parameters. The exception to this
seems to be the lip radius, changes in which have the almost same effect no matter
what the values of the other parameters are.
182
Ta
ble
3.2:
Qua
ntita
tiv
eev
aluatio
no
fsensitivit
yo
fshro
ude
d-ro
tor
performa
nce
to
cha
nges
in
par
ameter
va
lues
Par
ameter
Se
rie
s
?
(C
TSR
/C
TOR
)CP
=0
.004
?
(C
T/
CP
)ma
x,SR
/(C
T/
CP
)ma
x,OR
?
(FM
?)
ma
x,SR
/(FM
?)
ma
x,OR
Chang
ep
er
1%
Dt
increase
in
?tip
Blade
tip
clear
ance
LR1
3-D10
-0.0
1
+0.3
1
+0
.03
LR
13-D1
0-L31
-0.25
-0.1
9
-0.50
LR
13-D2
0
-0.35
-0.3
5
-0.77
LR
09-D2
0
-0.92
-1.0
6
-1.83
Chang
ep
er
1%
Dt
increase
in
rlip
,?
10
Inlet
lip
radius
D00
-?0.6
+0.0
05
-0.08
-0.0
1
D1
0-?
0.1
+0.4
55
+0.3
4
+0
.90
D1
0-?
0.6
+0.4
56
+0.1
7
+1
.09
D2
0-?
0.1
+0.4
58
+0.4
3
+0
.91
D2
0-?
0.6
+0.4
85
+0.4
5
+0
.83
Chang
ep
er
unit
cha
nge
in
?d
aw
ay
from
optim
um
Decr.
?d
Incr.
?d
Decr.
?d
Incr.
?d
Decr.
?d
Incr.
?d
Diffuser
angle
LR0
9-?
0.6
-L
72
-0.13
-1.1
2
-1.03
-1.74
-0.1
9
-1.7
9
LR
13-
?0.6
-L
72
-0.73
-1.0
8
-1.49
-1.39
-1.6
0
-2.1
3
LR
06-
?0.1
-L
72
?
-0.4
2
?
-0.32
?
-0.6
4
LR
13-
?0.1
-L
72
?
-0.4
2
?
-0.13
?
-0.6
3
Diffuser
leng
th
LR1
3-?
0.1
-D
10
-1.20
-0.1
5
?
-0.21
-2.5
9
-0.2
4
183
3.4 Shroud surface pressure measurements
This section presents the measurements of the static pressure distributions
over the shroud surfaces. The general characteristics of the pressure distributions,
common to all the shroud models, and of the consequent forces exerted on the inlet,
the diffuser, and the shroud as a whole, are discussed first, followed by more detailed
analyses of the effects of varying the individual shroud geometric parameters.
3.4.1 General characteristics
The pressure distributions for the individual shroud models, for each of four
different rotor collective angles, ?0 = 10?, 20?, 30? and 40?, are shown in Fig-
ures 3.22b?n, with the (suction) pressure coefficient, ?Cp = (patm ?p)/qtip, plotted
as a function of non-dimensional distance along the shroud surface. The coordinate
system for the normalized surface distance coordinate, s, is shown in Figure 3.22a.
For points downstream of the rotor, i.e., the shroud diffuser, s is equal to the actual
surface distance divided by the diffuser slant length, S = Ld/cos(?d/2), while for
points upstream of the rotor, i.e., the shroud inlet, the distances are normalized by
the inlet lip circumference, pi ?rlip. Therefore, s is equal to 0 at the rotor plane,
+1 at the diffuser exit plane, and ?1 at the outer edge of the inlet lip. An angular
coordinate, ?lip, is also convenient for discussion of the pressure distribution over
the shroud inlet, where ?lip = 0? corresponds to the rotor plane (s = 0) and ?lip =
180? corresponds to the outer edge of the inlet lip (s = ?1).
To better depict the actual, two-dimensional shape of the pressure distribution,
184
(a) Coordinate system for
pressure distribution plots
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(b) Model LR13-D20-?0.1
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(c) Model LR13-D20-?0.6
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(d) Model LR13-D20-?0.8
Figure 3.22: Shroud surface pressure distributions in hover: LR13-D20 models
185
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(e) Model LR13-D10-?0.1
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(f) Model LR13-D10-?0.6
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(g) Model LR13-D10-?0.1-L50
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(h) Model LR13-D10-?0.1-L31
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(i) Model LR13-D10-?0.8-L31
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(j) Model LR13-D10-?1.6-L31
Figure 3.22: Shroud surface pressure distributions in hover (contd.): LR13-D10
models
186
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(k) Model LR09-D20-?0.1
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(l) Model LR09-D00-?0.6
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(m) Model LR06-D20-?0.1
?1 ?0.8?0.6?0.4?0.2 0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
s
?C p
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(n) Model LR06-D00-?0.6
Figure 3.22: Shroud surface pressure distributions in hover (contd.): LR06 and
LR09 models
187
each figure also contains an inset diagram that shows the distributions wrapped
around a sectional, non-normalized view of the shroud?s profile shape, along with
the locations of the pressure taps on the shroud surface. In both representations,
the two-dimensional and the ?unwrapped?, the actual measured values are shown
by the symbols, connected by piecewise cubic Hermite interpolating polynomials
that were used to estimate the pressure values at the intermediate locations. These
interpolated distributions were integrated over the length of the shroud profile (s
= ?1 to +1), and around the azimuth of the shroud (? = 0??360?) to obtain the
estimates of the shroud thrust, Tshroud. For the purpose of integration, the pressure
at the boundary points ? the oute edges of the diffuser and inlet lip ? was assumed
to be equal to ambient atmospheric pressure.
The most distinctive feature of all the pressure distributions is a region of very
low pressures ? a ?suction peak? ? on the inner surface of the shroud inlets, just
above the plane of the rotor disk and immediately preceding the rise in pressure
through the rotor. All of the pressure profiles show an increase in suction from
?lip = 180? to about 125? (s = ?1.0 to ?0.7), at which point the rate of increase
decreases slightly. This decrease could, however, be an artifact of the imposed end-
condition mentioned above. A steeper increase is then seen, to a suction peak at
?lip = 50??40?, after which occurs a second leveling-off or slight drop, and then,
except at a rotor collective of 10?, a second, stronger suction peak just above the
rotor plane, at ?lip ? 20??0?. In some cases, a distinction is not seen between
these ?primary? and ?secondary? peaks, with the two merging into a single suction
peak. However, the secondary peak is always distinctly seen at ?0 = 40?, for all the
188
shrouded-rotor models. This secondary peak on the inlet of a shrouded rotor has
not been experimentally observed before; however, its existence has been predicted
(Fig. 3.23) by the CFD analysis performed in 1993 by Rajagopalan and Keys [87]:
Their analysis showed that because of the finite gap between the blade tips and the
shroud wall, air flowing through this gap from the higher-pressure region below the
rotor to the lower-pressure region above, opposite to the main flow direction, results
in the formation of a small eddy ? a rudimentary blade tip vortex, essentially ? and
that the high velocities of this vortical flow would create such a low-pressure region.
Unfortunately, because of the location of the rotor at the shroud throat, the forces
resulting from these high suction pressure act principally radially inwards, rather
than in the axial direction, and therefore do not directly contribute much useful
thrust. However, by suitably tailoring the shape of the shroud inlet profile in this
region near the rotor blade tips, in such a way that the resulting forces are inclined
more towards the axial direction, it may be possible to harness the potential of the
strong secondary suction peak to produce even greater amounts of thrust. This may
be a fruitful area for further research.
After the inlet suction peak(s), a rise in pressure occurs as the flow passes
through the rotor, the increase being most abrupt at a rotor collective of 20?. In the
diffusers, after the initial rapid increase in pressure below the rotor disk, pressure
generally increases gradually to ambient levels at the diffuser exit. However, in
the D00 diffusers (Figs. 3.22l, n), the wake expands to ambient pressure by about
halfway through the diffuser, and then actually shows a very slight over-pressure
before decreasing back again to ambient levels.
189
Figure 3.23: CFD prediction of secondary suction peak by Rajagopalan and
Keys [87].
With increasing collective, the suction pressures increase over the entire inlet,
the secondary peak becomes sharper and stronger, possibly corresponding to an
increase in strength of the blade tip vortex, and the plateau region in between the
peaks becomes smaller and less distinct. In most cases, the strength of the peak
increases monotonically with increasing collective, with a large increase seen between
?0 = 10? and 20?, and smaller increases from 20? to 40?. However, in some cases,
the strength is seen to decrease from ?0 = 20? to 30?, and then increase slightly at
40?. Diffuser suction pressures also increase as collective is increased from 10? to
30?, with little change thereafter as ?0 is increased to 40?. The exceptions to this
trend are the D10-L31 models with large blade tip clearances (Figs. 3.22i, j).
190
3.4.1.1 Inlet thrust distribution
Because of the dominant nature of the suction peak on the shroud pressure
distributions, the inlet pressure distribution has a strong influence on the overall
shroud thrust; it will be seen in the following section that the magnitude of the
positive thrust from the inlet is four or more times the magnitude of the negative
thrust from the diffuser. It is therefore worthwhile to analyze in more detail the
characteristics of the inlet pressure and thrust distributions. It is immediately ob-
vious that, because the normal vectors to the inlet surface do not all point in the
thrust (axial) direction, the thrust distributions will not be the same as that of
the inlet pressure distributions shown in Figures 3.22b?n. In addition, due to the
compound curvature of the inlet, the outer portions of the two-dimensional inlet
profile contribute greater amounts of area to the complete, three-dimensional inlet
surface than do the inner portions, so that the amount of force produced by the
outer portions of the inlet lip are greater than would be suggested by the pressure
distributions. For this reason, the resultant two-dimensional, sectional force vec-
tor, Fprimeinlet (Figure 3.24), obtained by integration of the inlet pressure distribution
over a single azimuthal section of the inlet4, does not yield the value of the net inlet
thrust when simply summed around the azimuth of the shroud inlet. In other words,
Finlet negationslash= Tinlet. Even though, due to the radial symmetry of the flow-field in hover,
the summation results in the radial (cos(negationslash Fprimeinlet)) components of the sectional forces
canceling out, leaving a three-dimensional force that is purely axial, still, the bias
due to the outer portions of the inlet result in an effective axial force component of
4The prime denotes that this is a two-dimensional force.
191
Figure 3.24: Sectional inlet resultant force vector
the sectional force that is greater than that given by |Fprimeinlet|?sin(negationslash Fprimeinlet).
Thus, there are two ways of looking at the inlet thrust distribution: first,
the two-dimensional distribution on a single, azimuthal section of the inlet, which
can be represented by the magnitude and direction of the resultant sectional force
vector, Fprimeinlet; and second, a ?three-dimensional? distribution that takes into consid-
eration the additional area of the outer portions of the inlet when integrated around
the shroud azimuth. Such a distribution is shown in Figure 3.25a, which plots the
variation in the differential inlet thrust, dCTinlet/d?lip, with inlet angular coordinate,
?lip, for one of the models, LR13-D10-?0.1, at four different rotor collectives. The
relative magnitudes of the thrust distributions at the different collectives are simi-
lar to the relative magnitudes seen in the pressure distributions; therefore, in this
figure, to better contrast the shapes of the distributions at the different collectives,
the magnitudes are shown scaled to the total inlet thrust at the corresponding col-
192
03060901201501800
20
40
60
80
dC T inlet
 / d?
lip [% C
T inlet
]
?lip [deg]
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(a) Variation in differential inlet thrust
03060901201501800
20
40
60
80
100
% of C
T inlet
 produced between 
? lip and 
? lip = 0
o
?lip [deg]
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(b) Fractional contribution to total inlet thrust
Figure 3.25: Inlet thrust distributions for model LR13-D10-?0.1
lective. Comparing these distributions with those of the inlet pressures shown in
Figure 3.22e, it can be seen that the peak thrust contribution occurs at ?lip between
45? and 60?, in contrast to the pressure distributions which peak at ?lip between 10?
and 20?. As rotor collective is increased, the location of the peak thrust contribution
moves outwards, to slightly higher values of ?lip. The other shroud models showed
similar distributions, and this trend was noticed, to varying amounts, in each of
them.
Figure 3.25b plots the fraction of the total inlet thrust produced between ?lip
= 0? and any other value of ?lip. This graph shows that about 80% of the total
inlet thrust is produced by the inner half of the inlet lip (?lip = 0? to 90?). This
suggests that the weight of the shroud can be reduced by reducing the stiffness of
the structure forming the outer portion of the inlet lip, since it does not bear a
substantial load. It is also seen that even though the location of the peak thrust
shifts by approximately 15? as rotor collective is increased (Figure 3.25a), the extent
of lip surface required for a given amount of inlet thrust changes by less than 10?
193
(Figure 3.25b).
3.4.2 Effects of changing shroud parameter values
The effects of varying the four shroud geometric parameters ? ?tip, rlip, ?d
and Ld ? on the pressure distributions and the consequent shroud forces are shown
in Figures 3.26?3.29 and described in the following subsections. Just as for the
performance analyses, multiple series of model comparisons are presented for each
shroud parameter (Table 3.3). Each sub-figure shows the comparisons of the differ-
ent models in a single series. The plots on the left-hand side of each figure show the
two-dimensional (2-D) pressure distributions of the different models being compared,
at four different collective angles. In these plots, unlike in the inset diagrams of Fig-
ures 3.22b?n, the different linear and angular dimensions of the different shrouds
have been scaled to the same values, to enable easier comparison of the pressure
profiles. The right-hand side of each figure shows graphs of the variations with
rotor collective of shroud thrust (CTshroud), shroud inlet thrust (CTinlet), shroud dif-
fuser thrust (CTdiffuser), and the magnitude (|CFprimeinlet|) and direction (negationslash CFprimeinlet) of the
resultant sectional inlet force vector, for each of the different models in the series.
Despite the slightly misleading nature of this quantity, discussed earlier, specifying
the magnitude and direction of Fprimeinletremains a useful technique for comparing the
inlet thrust distributions of the different shroud models, and has therefore been
included in these analyses.
From the figures, it can be seen that the force coefficients all generally increase
194
Table 3.3: Shrouded-rotor model comparison series for analyzing effects of the
shroud geometric parameters on shroud pressure distributions
Model series for blade tip clearance ?tip values [%Dt]
LR13-D20 0.09, 0.63, 0.82
LR13-D10 0.09, 0.63
LR13-D10-L31 0.09, 0.82, 1.57
Model series for inlet lip radius rlip values [%Dt]
D20-?0.1 6.48, 8.99, 13.1
D00-?0.6 6.48, 8.99
Model series for diffuser included angle ?d values [?]
LR13-?0.1 9.70, 20.22
LR13-?0.6 9.70, 20.22
Model series for diffuser length Ld values [%Dt]
LR13-D10-?0.1 31, 50, 72
195
in magnitude with increasing collective, up to around ?0 = 30?, and then level off or
decrease at higher collectives, exhibiting a stall-like behavior. As would be expected,
the variations in|CFprimeinlet|andCTinlet are very similar, and follow the trends of the inlet
pressure distributions described earlier. negationslash CFprimeinlet generally increases with increasing
collective, because of the increasing strength and merging of the two suction peaks,
and has values mostly between 40? and 60?. Note that this is 5?12? lower than the
location of the peak thrust contribution when 3-D effects are included ? compare
Figures 3.25a and 3.26b for model LR13-D10-?0.1. The magnitudes of the diffuser
thrust are at most only about 26% those of the inlet thrust, so the net shroud
thrust more closely follows the trends of the inlet thrust, both with respect to rotor
collective and with respect to changes in the shroud geometry.
3.4.2.1 Blade tip clearance
The effects of changing the blade tip clearance, ?tip, are shown in Figure 3.26,
with data available from three series of models: LR13-D10-L31, LR13-D10 and
LR13-D20 (Table 3.3). From the measurements, increasing the tip clearance was
seen to have the following effects:
(i) The suction pressures on the inlet lip decrease. This effect is generally more
evident at?0 = 20? and 30? than at 10? and 40?. Because of the drop in suction
pressure, the magnitudes of CFprimeinlet, CTinlet and CTshroud decrease, the slopes of
the initial part of the graphs of these coefficients with respect to ?0 decrease,
and the ?stall? behavior of these quantities changes from a ?sharper? stall to a
196
q 0 = 10o q 0 = 20o
q 0 = 30o q 0 = 40o
0
0.01
0.02 CT (shroud)
0
0.01
0.02 CT (inlet)
?1
?0.5
0x 10?3 CT (diffuser)
10 20 30 400
0.005
0.01
q 0 [deg]
|CF?inlet|
10 20 30 4045
50
55
q 0 [deg]
?  CF?inlet [deg]
LR13?D10?d 0.1?L31LR13?D10?d 0.8?L31
LR13?D10?d 1.6?L31
(a) Series LR13-D10-L31
q 0 = 10o q 0 = 20o
q 0 = 30o q 0 = 40o
0
0.005
0.01
0.015 CT (shroud)
0
0.01
0.02 CT (inlet)
?2
?1
0x 10?3 CT (diffuser)
10 20 30 400
0.005
0.01
q 0 [deg]
|CF?inlet|
10 20 30 4040
45
50
q 0 [deg]
?  CF?inlet [deg]
LR13?D10?d 0.1LR13?D10?d 0.6
(b) Series LR13-D10
q 0 = 10o q 0 = 20o
q 0 = 30o q 0 = 40o
0
0.005
0.01
0.015 CT (shroud)
0
0.01
0.02 CT (inlet)
?4
?2
0x 10?3 CT (diffuser)
10 20 30 400
0.005
0.01
q 0 [deg]
|CF?inlet|
10 20 30 4035
40
45
50
q 0 [deg]
?  CF?inlet [deg]
LR13?D20?d 0.1LR13?D20?d 0.6
LR13?D20?d 0.8
(c) Series LR13-D20
Figure 3.26: Effects of blade tip clearance on pressure distribution
197
more gentle one. At small tip clearances, the magnitudes of the forces increase
up to a collective of about 30? and then decrease. At the larger tip clearances
tested, the forces continue to increase over the full range of collectives, albeit
with a lower slope. At the intermediate values of ?tip, the forces decrease in
magnitude at collectives between 20??25?, after which the magnitude increases
again from ?0 = 30? to 40?.
(ii) The location of the secondary suction peak moves very gradually towards
the rotor plane (to lower values of ?lip). This is most clearly seen for the
D10 models. The behavior of negationslash CFprimeinlet changes from increasing with increasing
collective (with a stall pattern at higher collectives), to a more gentle increase,
to remaining roughly constant with some fluctuations, to actually decreasing
slightly.
(iii) Diffuser suction pressures and |CTdiffuser| increase at collectives of 10? and 20?.
A crossover seems to take place near ?0 = 30?, so that the opposite trend is
seen at the higher collectives. This effect of changing ?tip is most pronounced
in the D10-L31 models.
(iv) At collectives of 10? and 20?, the pressure jump through the rotor plane be-
comes less pronounced. This effect is not so consistent at the higher collectives.
3.4.2.2 Inlet lip radius
The effects of changing the inlet lip radius, rlip, are shown in Figure 3.27,
with data from two series of models: D00-?0.6 and D20-?0.1 (Table 3.3). From the
198
q 0 = 10o q 0 = 20o
q 0 = 30o q 0 = 40o
0
0.005
0.01 CT (shroud)
0
0.005
0.01 CT (inlet)
?1
0
1 CT (diffuser)
10 20 30 400
2
4
6x 10?3
q 0 [deg]
|CF?inlet|
10 20 30 4045
50
55
60
q 0 [deg]
?  CF?inlet [deg]
LR06?D00?d 0.6LR09?D00?d 0.6
(a) Series D00-?0.6
q 0 = 10o q 0 = 20o
q 0 = 30o q 0 = 40o
0
0.005
0.01
0.015 CT (shroud)
0
0.01
0.02 CT (inlet)
?6
?4
?2
0x 10?3 CT (diffuser)
10 20 30 400
0.005
0.01
q 0 [deg]
|CF?inlet|
10 20 30 4020
40
60
80
q 0 [deg]
?  CF?inlet [deg]
LR06?D20?d 0.1LR09?D20?d 0.1
LR13?D20?d 0.1
(b) Series D20-?0.1
Figure 3.27: Effects of inlet lip radius on pressure distribution
measurements, increasing the lip radius was seen to have the following effects:
(i) The inlet pressure profile becomes less uniform and more skewed towards the
inner portion of the inlet, generally decreasing over most of the inlet lip (?lip
= 45??180?) and increasing over only the innermost portion (?lip = 0??45?),
although a lot of small-scale variation is seen. However, the LR09 inlet has
46% more surface area than the LR06 inlet, and the LR13 inlet an additional
80% more, and this increase in total inlet surface area with increasing lip radius
more than compensates for the loss in suction, resulting in a net increase in
199
|CFprimeinlet|.
(ii) The location of the secondary peak shows a very slight tendency to move fur-
ther upstream away from the rotor plane, i.e., to larger values of?lip. However,
the range of this movement is very small, within ?lip = 0? to 30?, so the in-
creasing strength of the secondary peak, relative to the decreasing suction on
the outer portions of the inlet, still causes negationslash CFprimeinlet to decrease, from ?60? for
the LR06 models to ?45?for the LR13.
(iii) The inlet thrust is a function of both the magnitude and direction of Fprimeinlet,
and is seen to increase from LR06 to LR09, and then decrease slightly from
LR09 to LR13.
(iv) In the diffusers, too, the suction pressures generally decrease. In the D00-
?0.6 models, this has no effect on the net thrust because ?d = 0?, but in the
D20-?0.1 models it causes a reduction in |CTdiffuser|.
(v) Except at ?0 = 10?, the net shroud thrust increases with increasing lip radius.
3.4.2.3 Diffuser angle
The effects of changing the diffuser included angle,?d, are shown in Figure 3.28,
with data from two series of models: LR13-?0.1 and LR13-?0.6 (Table 3.3). From
the measurements, the following observations were made:
(i) Although small-scale variations are seen in the inlet pressure distributions,
changing the diffuser angle does not appear to significantly modify the inflow
200
q 0 = 10o q 0 = 20o
q 0 = 30o q 0 = 40o
0
0.005
0.01
0.015 CT (shroud)
0
0.01
0.02 CT (inlet)
?4
?2
0x 10?3 CT (diffuser)
10 20 30 400
0.005
0.01
q 0 [deg]
|CF?inlet|
10 20 30 4035
40
45
50
q 0 [deg]
?  CF?inlet [deg]
LR13?D10?d 0.1LR13?D20?d 0.1
(a) Series LR13-?0.1
q 0 = 10o q 0 = 20o
q 0 = 30o q 0 = 40o
0
0.005
0.01
0.015 CT (shroud)
0
0.01
0.02 CT (inlet)
?4
?2
0x 10?3 CT (diffuser)
10 20 30 400
0.005
0.01
q 0 [deg]
|CF?inlet|
10 20 30 4040
45
50
q 0 [deg]
?  CF?inlet [deg]
LR13?D10?d 0.6LR13?D20?d 0.6
(b) Series LR13-?0.6
Figure 3.28: Effects of diffuser included angle on pressure distribution
201
pattern of the shrouded rotors. Similarly, negligible effect is seen on the vari-
ations of CFprimeinlet and CTinlet with collective. Below ?0 = 25?, the D20 models
show slightly higher forces, while above 25?, the forces on the D10 models are
higher; however, the differences are very small. These differences are slightly
greater in the ?0.6 series than in the?0.1 series.
(ii) For both series, negationslash CFprimeinlet lies around 45?. In the ?0.1 series, negationslash CFprimeinlet is slightly
higher for the D10 model than for the D20 model. In the ?0.6 series, apart
from the fluctuations in the values for the D10 model, little difference is seen
in the mean trends for the two models.
(iii) In the diffuser, on the other hand, the D20 models consistently show higher
suction pressures. Because of this, and also because of the greater inclination
of the their diffuser walls, the D20 models produce up to four times as much
negative thrust from the diffuser as the D10 models.
(iv) The net effect is that the D10 models produce up to 25% to 40% more net
shroud thrust than the D20 models, with the improvement seen mostly at
collectives above 20?.
3.4.2.4 Diffuser length
The effects of changing the diffuser length, Ld, are shown in Figure 3.29, with
data from a single model series: the LR13-D10-?0.1 models (Table 3.3). From the
measurements, increasing the diffuser length was seen to have the following effects:
202
q 0 = 10o q 0 = 20o
q 0 = 30o q 0 = 40o
0
0.01
0.02 CT (shroud)
0
0.01
0.02 CT (inlet)
?1.5
?1
?0.5
0x 10?3 CT (diffuser)
10 20 30 400
0.005
0.01
q 0 [deg]
|CF?inlet|
10 20 30 4040
45
50
55
q 0 [deg]
?  CF?inlet [deg]
LR13?D10?d 0.1?L31LR13?D10?d 0.1?L50
LR13?D10?d 0.1?L72
(a) Series LR13-D10-?0.1
?0 = 10o ?0 = 20o
?0 = 30o ?0 = 40o
(b) Series LR13-D10-?0.1, but with correct
scaling of diffuser lengths
Figure 3.29: Effects of diffuser length on pressure distribution
203
(i) No significant effects are seen on the inlet pressure distributions, and on the
variations of |CFprimeinlet|, CTinlet and CTshroud with rotor collective. The magnitudes
of the forces decrease slightly, more so from L31 to L50 than from L50 to L72,
but the reduction is marginal. Similarly, negationslash CFprimeinlet decreases, but the overall
decrease, while being more clear than the decrease in the forces, is still only
by 5? or less.
(ii) When the diffuser pressure profiles are compared with all the diffusers scaled to
the same length (Figure 3.29a), the shorter diffusers appear to expand the flow
to ambient pressure more slowly, and so have higher suction pressures exerted
on their surfaces. However, for this shroud parameter, this representation
? with the diffusers all scaled to the same length ? is misleading. When
the pressure distributions are compared with the diffuser lengths shown in
correct relation to each other (Figure 3.29b), it can be seen that the initial
rate of expansion, immediately downstream of the rotor, is fairly independent
of diffuser length. Further downstream from the rotor, though, the longer
diffusers experience higher suction pressures. This, coupled with the larger
surface area over which they act, causes |CTdiffuser| to increase significantly for
the longer diffusers. However, since it is an order of magnitude smaller than
the inlet thrust coefficient, the net shroud thrust is not much affected.
204
0
20
40
60
80
100
% C
T inlet
?0 = 10o ?0 = 20o
045901351800
20
40
60
80
100
?lip [deg]
% C
T inlet
?0 = 30o
04590135180 ?
lip [deg]
?0 = 40o
LR06?D00??0.6LR09?D00??0.6
(a) Series D00-?0.6
0
20
40
60
80
100
% C
T inlet
?0 = 10o ?0 = 20o
045901351800
20
40
60
80
100
?lip [deg]
% C
T inlet
?0 = 30o
04590135180 ?
lip [deg]
?0 = 40o
LR06?D20??0.1LR09?D20??0.1
LR13?D20??0.1
(b) Series D20-?0.1
Figure 3.30: Effect of lip radius on inlet thrust distributions
3.4.2.5 Inlet thrust distribution
In contrast to the movement of the secondary suction peak, it was found that
increasing the inlet lip radius caused the peak thrust location to move to lower
values of ?lip (Fig. 3.30). This implies that for shrouds with large inlet lip radii,
a smaller extent of the inlet lip is required to produce a certain percentage of the
maximum possible inlet thrust, while for shrouds with sharper inlet lips, the inlet
needs to be extended to larger values of ?lip to attain the same percentage of the
maximum possible inlet thrust.
The effects of the other shroud parameters on the inlet thrust distribution were
found to be negligible. Variations were seen between the distributions of the models
with different tip clearances, but no consistent trends could be deduced. As diffuser
length was increased, a smaller extent of the inlet lip was required for the same
percentage of inlet thrust, just as for increasing lip radius, but the differences were
minimal. When diffuser angle was changed, the LR13-?0.6 models showed virtually
205
no change in the inlet thrust distribution, while the LR13-?0.1 series showed that
the D20 model required a smaller lip extent than the D10 model ? but, again, by
a very small amount.
3.5 Wake axial velocity measurements
This section discusses the measurements made of the axial flow velocities in
the wakes of the open and shrouded rotors in hover. The velocity data are shown
normalized by the ideal induced velocity at the rotor disk,vi (Equation 1.1, page 16),
while length dimensions have been normalized by the rotor radius, R.
3.5.1 General characteristics
For the open rotor, at a collective angle of 20?, the flow-field was mapped out
to a depth (z) of 5 in (1.6R) below the plane of the rotor. This was also done for one
of the shrouded rotors, model LR13-D10-?0.6, at the same collective and to a similar
depth below the diffuser exit plane. These measurements are shown in Figure 3.31,
both in the form of two-dimensional maps of the entire wake, with superimposed
dots indicating the locations at which the measurements were made, as well as in
the form of radial profiles of the velocity at different axial stations. Limitations of
the measurement apparatus prevented measurements from being made immediately
below the plane of the rotor disk in the case of the shrouded rotor; however, the
two sets of data do overlap between z/R = 1.12 and 1.60. Comparing the two
sets of measurements within this range, it is readily apparent that the shroud is
206
very effective at improving the uniformity of the flow in the wake, which translates
directly to lower induced power for the shrouded rotor, even though it is at a higher
thrust coefficient than that of the open rotor (0.0211 vs. 0.0147). This improvement
in uniformity is critical, considering the significant velocity deficit seen in the core
region of the flow, caused by the lack of twist in the rotor blades and by the blockage
and viscous losses at the rotor centerbody.5
By observing the inclination of the countour lines in the figures, it appears
that the diameter of the shrouded rotor?s wake remains fairly constant after exiting
from the diffuser; this supports the fundamental assumption made in the momentum
theory model that the flow is fully expanded at the diffuser exit plane. In contrast,
the countours in the wake of the open rotor show the characteristic contraction that
is expected for this configuration. However, when compared with the prediction ?
superimposed on the 2-D map ? for the location of the wake boundary as given
by the Landgrebe model, which is an empirical method that is commonly used for
conventional, large-scale rotors [14, p. 450], these data seem to suggest that the
wake of this micro-scale open rotor contracts significantly more than that of larger
rotors. This hypothesis is supported by further comparisons, shown in Figure 3.32,
of measurements in the wakes of hovering MAV-scale and conventional-scale rotors,
using data from other researchers. Figure 3.32a shows the positions of the blade
tip vortex cores at different wake ages, obtained by Hein and Chopra from flow-
visualization experiments on a MAV-scale rotor [20], while Figure 3.32b shows data
5See also the flow visualization studies conducted by Hein and Chopra on a MAV-scale rotor
in hover [20].
207
0
1
2
3
v / v
i
z/R = 0.04
0
1
2
3
v / v
i
z/R = 0.1
0
1
2
3
v / v
i
z/R = 0.2
0
1
2
3
v / v
i
z/R = 0.4
0 0.5 1
0
1
2
3
v / v
i
Radial coordinate [r/R]
z/R = 0.6
z/R = 0.8
z/R = 1
z/R = 1.2
z/R = 1.4
0 0.5 1Radial coordinate [r/R]
z/R = 1.6
(a) Open rotor, CT = 0.0147
0
1
2
3
v / v
i
z/R = 1.12
0
1
2
3
v / v
i
z/R = 1.44
0
1
2
3
v / v
i
z/R = 1.76
0 0.5 1
0
1
2
3
v / v
i
Radial coordinate [r/R]
z/R = 2.08
z/R = 2.4
z/R = 2.72
0 0.5 1Radial coordinate [r/R]
z/R = 3.04
(b) Shrouded-rotor LR13-D10-?0.6, CT = 0.0211
Figure 3.31: Rotor wake development at ?0 = 20?
208
from measurements by Bohorquez and Pines [168], mapping out the wake of another
micro-rotor, similar to those done in this study. In both figures, the prediction by
the Landgrebe wake-boundary model has been superimposed. Figure 3.32c shows a
comparison of the wake profiles at different axial stations, obtained by interpolation
from the data shown in Figure 3.31a, with measurements made by Boatright on a
full-scale rotor [174]. All three comparisons indicate a greater contraction of the
wake of the micro-scale rotors compared to that of larger ones.
Figure 3.31a also shows that, for the open rotor, the peak inflow velocity at
the plane of the rotor disk takes place at a radial location of about 0.7R, which is
much further inboard compared to the inflow profiles of conventional-scale rotors.
This suggests that the blade tip vortices of MAV-scale rotors have much larger
cores than do conventional-scale rotors. The flow-visualization studies conducted
on similarly-sized micro-rotors by Hein and Chopra [20] and by Ramasamy et al.
[24, 171] resulted in similar observations.
Fortheremainingshrouded-rotormodelsintheLR13-D10?family?, wakeveloc-
ity measurements were only made at the diffuser exit plane, but at additional rotor
collective angles as well. Figure 3.33 shows these radial wake profiles (w/vi) for four
of the models, at collectives of 10?, 20?, 30? and 40?, along with the momentum-
theory prediction, equal to 1/?d. It can be seen that with increasing collective, the
wake profile becomes increasingly non-uniform, leading to increased induced power
losses [14, p. 89]. In the case of the shortest-diffuser models (L31: Figs. 3.33a, b),
the pitot-probe readings became negative at the higher collectives (30? and 40?), in-
dicating large deviations in the flow direction away from the axial direction. A few
209
0.6 0.7 0.8 0.9 1
0
0.2
0.4
0.6
0.8 r / R
z / R
Landgrebe modelprediction 
Mean tip vortexcore trajectory 
(a) Tip vortex trajectory, MAV-scale rotor: CT
= 0.0241, ? = 0.162, Nb = 2, no twist. Data
from [20]
(b) Wake structure, MAV-scale rotor: CT =
0.0157, ? = 0.110, Nb = 2, no twist, ?0 = 15?.
Data from [168]
0
1
2
v / v
i
z/R = 0.1
z/R = 0.3 z/R = 0.5
0 0.2 0.4 0.6 0.8 1
0
1
2
v / v
i
z/R = 0.7
r / R 0 0.2 0.4 0.6 0.8 1
z/R = 1
r / R 0 0.2 0.4 0.6 0.8 1
z/R = 1.5
r / R
Large scaleMAV scale
(c) Wake development: Comparison with data from tests on a large-scale rotor with ?4? of twist
[174]
Figure 3.32: Comparison of rotor wake development of conventional-scale and
MAV-scale rotors
210
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
Momentum Theory prediction 
Diffuser wall location 
(a) LR13-D10-?0.8-L31
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
Momentum Theory prediction 
Diffuser wall location 
(b) LR13-D10-?0.1-L31
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
Momentum Theory prediction 
Diffuser wall location 
(c) LR13-D10-?0.1-L50
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
Momentum Theory prediction 
Diffuser wall location 
(d) LR13-D10-?0.1-L72
Figure 3.33: Wake profiles at diffuser exit plane of shrouded rotors
investigatory measurements were made with the pitot probe aligned in the tangen-
tial (azimuthal) direction, and these indicated the presence of large swirl velocities
near the diffuser wall. These observations suggest that the rotor may have stalled at
these higher collectives. The longer-diffuser models (L50, L72) also showed the trend
of decreasing axial velocities near the shroud wall with increasing rotor collective,
but in these cases the probe measurements did not become negative.
The radial variations in axial velocity can be integrated to obtain the value
for an average velocity that produces the same mass flow through the diffuser exit
211
plane. These average values ? again, normalized by vi ? are shown in Fig. 3.34a,
and appear to be fairly close to the theoretical value of 1/?d. However, it is clear
that the profiles are far from uniform, and the penalty of this non-uniformity on
the induced power expenditure of the rotor can be quantified by integrating the
radial distributions of axial kinetic energy, and dividing the result by the ideal
power requirement of the shrouded rotor (CPi = C3/2T /2??d) to obtain an estimate
for the induced power correction factor, ?. The effect of increasing collective on
the estimates of ? thus obtained are shown in Fig. 3.34b. With the exception of
the LR13-D10-?0.1-L72 model, the lowest value of ? occurs at a rotor collective of
20?. At ?0 = 10?, the LR13-D10-?0.1-L72 model shows a value of ? that is less
than 1.0, which is not a physically realistic solution. The wake measurements for
this model, at this collective, were therefore probably in error, a conclusion that is
also suggested by the observation that the wake profile for this model at ?0 = 10?
does not show quite the same behavior shown by those of the other models at that
collective angle (Figs. 3.33d and also 3.37a).
It should be noted that this derivation of the estimates for ? did not include
measurements of the swirl velocity, and hence did not account for energy lost in
imparting rotational momentum to the flow due to the rearward inclination of the
lift vectors on the rotor blade elements. Even though a portion of the swirl in
the wake is due to the profile drag of the blades, which would not be insignificant
at low Reynolds numbers, still, shrouding a rotor has the effect of increasing the
induced inflow angle (hence, the tilt of the lift vector) and reducing the aerodynamic
angle of attack (hence, the profile drag) of the blade elements, compared to an open
212
10 20 30 40
0.8
0.9
1
1.1
?0 [deg]
(w avg
/v i) / (w/v
i) theory
LR13?D10??0.8?L31LR13?D10??0.1?L31
LR13?D10??0.1?L50LR13?D10??0.1?L72
(a) Average wake velocity
10 20 30 400.4
0.8
1.2
1.6
2
2.4
?0 [deg]
Induced power correction factor, 
?
LR13?D10??0.8?L31LR13?D10??0.1?L31
LR13?D10??0.1?L50LR13?D10??0.1?L72
(b) Induced power correction factor
Figure 3.34: Effects of collective angle on induced power
rotor producing the same thrust (Eq. 3.1 on page 141). The true values of ? would
therefore be somewhat higher than the values shown in these figures.
3.5.2 Effects of changing shroud parameter values
3.5.2.1 Blade tip clearance
Figure 3.35 shows the effect of changing blade tip clearance, ?tip, on the wake
profile at the diffuser exit plane of the shrouded rotors. Figures 3.35a?d show the
wake profiles of the three LR13-D10-L31 models at each of four collective angles,
while Figures 3.35e and 3.35f compare the exit wake profiles of the LR13-D10-L31
and LR13-D10-L72 models, at a collective of 20?, with the wake profile of the open
rotor at the same collective and at the corresponding distance from the rotor plane
(z/R = 0.62 and 1.44, respectively). It can be seen that increasing ?tip causes the
shrouded rotor?s wake profile becomes less uniform, with increasing axial velocity
deficit at the diffuser wall, and become more like that of the open rotor, with the
213
consequent increases in induced power losses and the induced power correction factor
(Fig. 3.36b). The average wake velocity generally decreases (Fig. 3.36a), and the
radiallocationofpeakdownwashmovesinboard, possiblycorrelatedwithanincrease
in size of the tip vortex cores.
3.5.2.2 Diffuser length
Figure 3.37 shows the effect of increasing the expansion ratio by changing dif-
fuser length,Ld, on the wake profile at the diffuser exit plane of the three LR13-D10-
?0.1 shrouded-rotor models. Figures 3.37a?d show the wake profiles of the models
at each of four collective angles, while Figure 3.37e compares the exit wake profiles,
at a collective of 20?, with the wake profiles of the open rotor at the same collective
and at the corresponding distances from the rotor plane ? z/R = 0.62, 1.00 and
1.44. From these figures, increasing diffuser length is seen to have two effects. The
first is that the entire wake profile stretches radially outwards, corresponding to the
increasing expansion of the wake. The second is that although the velocity deficit
in the inner, core region of the wake increases, it gets reduced in the outer regions,
at the diffuser wall. This is probably because, with the expansion of the flow, the
higher-velocity flow at the mid-span locations of the wake energizes and entrains
the flow at the diffuser walls. The net effect is that the induced power correction
factor generally remains about the same (Fig. 3.38b). Although the peak downwash
velocity decreases in magnitude, the average velocity also generally remains about
the same (Fig. 3.38a). In contrast, in the case of the open rotor (Figure 3.37e),
214
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
LR13?D10??0.1?L31LR13?D10??0.8?L31
Momentum Theory prediction 
Diffuser wall location 
(a) Series LR13-D10-L31, ?0 = 10?
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
LR13?D10??0.1?L31LR13?D10??0.8?L31
LR13?D10??1.6?L31
Momentum Theory prediction 
Diffuser wall location 
(b) Series LR13-D10-L31, ?0 = 20?
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
LR13?D10??0.1?L31LR13?D10??0.8?L31
Momentum Theory prediction 
Diffuser wall location 
(c) Series LR13-D10-L31, ?0 = 30?
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
LR13?D10??0.1?L31LR13?D10??0.8?L31
Momentum Theory prediction 
Diffuser wall location 
(d) Series LR13-D10-L31, ?0 = 40?
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
2
2.4 Radial coordinate [r/R]
w / v
i
LR13?D10??0.1?L31LR13?D10??0.8?L31
LR13?D10??1.6?L31 Open Rotor,z/R = 0.62 
MT prediction for Shrouded Rotors 
Diffuser wall location 
(e) Series LR13-D10-L31, ?0 = 20?, compared
with open rotor
0 0.2 0.4 0.6 0.8 1
0
0.5
1
1.5
2
Radial coordinate [r/R]
w / v
i
LR13?D10??0.1LR13?D10??0.6 Open Rotor,z/R = 1.44 
MT prediction for Shrouded Rotors 
Diffuser wall location 
(f) Series LR13-D10-L72, ?0 = 20?, compared
with open rotor
Figure 3.35: Effect of blade tip clearance on shrouded-rotor exit-plane wake profiles
215
0 0.4 0.8 1.2 1.60.8
0.85
0.9
0.95
1
1.05
?tip [%Dt]
(w avg
/v i) / (w/v
i) theory
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(a) Average wake velocity
0 0.4 0.8 1.2 1.61.2
1.6
2
2.4
?tip [%Dt]
Induced power correction factor, 
? ?0 = 10o?0 = 20o?
0 = 30o?0 = 40o
(b) Induced power correction factor
Figure 3.36: Effects of changing blade tip clearance on induced power
with increasing distance from the rotor plane, the radial location of peak downwash
moves inboard, corresponding to the contraction of the wake, and while the profile
also becomes more uniform in this case, it does so by the action of the higher-velocity
mid-span flow entraining both the outer and the inner regions of the wake.
216
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
LR13?D10??0.1?L31LR13?D10??0.1?L50
LR13?D10??0.1?L72
(a) Series LR13-D10-?.1, ?0 = 10?
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
LR13?D10??0.1?L31LR13?D10??0.1?L50
LR13?D10??0.1?L72
(b) Series LR13-D10-?.1, ?0 = 20?
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
LR13?D10??0.1?L31LR13?D10??0.1?L50
LR13?D10??0.1?L72
(c) Series LR13-D10-?.1, ?0 = 30?
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
Radial coordinate [r/R]
w / v
i
LR13?D10??0.1?L31LR13?D10??0.1?L50
LR13?D10??0.1?L72
(d) Series LR13-D10-?.1, ?0 = 40?
0 0.2 0.4 0.6 0.8 1
0
0.4
0.8
1.2
1.6
2
2.4 Radial coordinate [r/R]
w / v
i
LR13?D10??0.1?L31LR13?D10??0.1?L50
LR13?D10??0.1?L72Open rotor, z/R = 0.62
Open rotor, z/R = 1.00Open rotor, z/R = 1.44
(e) Series LR13-D10-?.1, ?0 = 20?, compared
with open rotor
Figure 3.37: Effect of diffuser length on shrouded-rotor exit-plane wake profiles
217
30 40 50 60 70
0.8
0.9
1
1.1
Ld [%Dt]
(w avg
/v i) / (w/v
i) theory
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(a) Average wake velocity
30 40 50 60 700.5
1
1.5
2
Ld [%Dt]
Induced power correction factor, 
?
?0 = 10o?
0 = 20o?0 = 30o
?0 = 40o
(b) Induced power correction factor
Figure 3.38: Effects of changing diffuser length on induced power
218
Chapter 4
Experimental Results: Wind-Tunnel Tests
4.1 Introduction
This chapter presents the measurements that were made in the translational-
flight tests of the open rotor and the shrouded-rotor model LR13-D10-?0.1-L31 at a
rotor collective angle of 20?. With a shroud chord (equal torlip +Ld) of 44%Dt, this
shrouded-rotor model falls under the category of what are known as ?short-chord?
shrouds. The data are presented in the form of the effects of changing free-stream
velocity (airspeed) and angle of attack on the shroud surface pressure distribution
and the on the seven performance variables: thrust, normal force, lift, drag, location
of center of pressure, pitch moment and shaft power (Fig. 4.1). All quantities are
presented in non-dimensional forms, either as aerodynamic coefficients or, in the
case of the center-of-pressure location (xcp), as normalized by the rotor radius (R).
For presenting the effects of airspeed, a non-dimensional ?airspeed ratio? parameter,
?prime, has been used, defined as:
?prime = v?/?R (4.1)
219
(a) Forces and moments
(b) Angle of attack (c) Center of pressure
Figure 4.1: Translational flight: variable definitions
220
Table 4.1: Airspeed ratio (?prime) values tested
Rotational speed Blade tip speed Wind tunnel speed
10 ft/s 15 ft/s 20 ft/s
2000 rpm 54.5 ft/s 0.18 0.28 0.37
3000 rpm 81.8 ft/s 0.12 0.18 0.24
4000 rpm 109.1 ft/s 0.09 0.14 0.18
This parameter differs from the conventional rotor advance ratio, ?, as follows:
?prime = ?/cos(?rotor)
= ?/cos(90? ??) (4.2)
and from the conventional propeller advance ratio, J, by the constant factor 1/pi.
At the higher angles of attack (? ? 90?), then, ?prime ? ?, and at ? = 0?, ?prime = J/pi.
The use of this parameter makes it convenient to present, on a single graph, the
characteristics of the rotor at all angles of attack, from the ?propeller-mode? or
climbing-flight condition (? = 0?) to the edgewise-flight condition (? = 90?), at all
rotational speeds and airspeeds tested. In edgewise flow, the maximum value of ?prime
at which the models were tested was 0.37. In axial flow, the models were tested
up to ?prime = 0.28 for the open rotor and 0.18 for the shrouded rotor, corresponding
respectively to maximum propeller advance ratios of 0.88 and 0.57. Table 4.1 gives
the values of ?prime for each rotational speed and free-stream velocity tested.
221
4.2 Shroud surface pressure measurements
The pressure-distribution data are shown in Figures 4.3?4.5. The effect of
changing airspeed on the pressure distributions is first shown in Figures 4.3a?d, at
each of four angles of attack: 0?, 30?, 60? and 90?. These figures also include, for
comparison, the pressure measurements taken on this model during the hover tests
(?prime = 0). The effects of changing angle of attack are then presented in Figures 4.5a?
c, for each of the three airspeeds tested ? 10, 15 and 20 ft/s, corresponding to
airspeed ratios of 0.12, 0.18 and 0.24. In each figure, the measured values are shown
by the symbols, joined by piecewise-cubic Hermite interpolating polynomials. The
data for the windward side of the shroud (? = 180?) are shown on the left side
of each figure, and the data for the leeward side (? = 0?) on the right side. The
data are plotted in coefficient form (Cp) as functions of the non-dimensional distance
coordinate, s, along the shroud surface. Thes-coordinate system used in these plots
is the same one used for the pressure-distribution plots in Chapter 3, and is shown
again for convenience in Figure 4.2. Its origin lies at the blade passage region, where
the inlet and diffuser surfaces meet. The coordinate s is positive for points below
the rotor (the diffuser), and negative for points above the rotor (the inlet). Surface
distances in the diffuser are normalized by the diffuser slant length, S, and in the
inlet by the inlet lip circumference, pi ?rlip. Therefore, s = 0 at the rotor plane,
s = +1 at the diffuser exit plane, and s = ?1 at the outer edge of the inlet lip. The
angular coordinate, ?lip, is also convenient for locating points on the surface of the
shroud inlet; ?lip = 0? corresponds to the rotor plane and ?lip = 180? corresponds to
222
Figure 4.2: Coordinate system for pressure distribution plots
the outer edge of the inlet lip.
For better visualization of the actual shape of the pressure distributions, the
data from Figures 4.3a?d are also plotted diagrammatically in a two-dimensional
(?2-D?) representation, superimposed on a sectional view of the shroud, as viewed
from the ? = 270? azimuth, in Figures 4.4a?d. In these 2-D plots, the ?empty?
symbols represent pressures lower than ambient atmospheric (suction), while the
solid (filled) symbols represent pressures greater than ambient (over-pressure).
Figures 4.3a and 4.4a show the surface pressure measurements in the axial-
flight condition (? = 0?). Figure 4.3a shows the data plotted versus the non-
dimensional surface coordinate s, while Figure 4.4a shows the two-dimensional rep-
resentation of the pressure distributions. As expected for this flight condition, the
pressure distributions are symmetric between ? = 180? and ? = 0?. In hover (?prime =
0), suction pressures (Cp < 0) are seen over the entire inner surface of the shroud,
with the flow clearly expanding completely to ambient atmospheric pressure at the
223
?1 ?0.75?0.5?0.250 0.250.5 0.75 1?0.04
?0.02
0
0.02
0.04
0.06
0.08
s
?C p
? = 180o
?? = 0.12?? = 0.18
?? = 0
?1?0.75?0.5?0.2500.250.50.751?0.04
?0.02
0
0.02
0.04
0.06
0.08
s
?C p
? = 0o
?? = 0.12?? = 0.18
?? = 0
(a) ? = 0?(axial flow)
?1 ?0.75?0.5?0.250 0.250.5 0.75 1
?0.04
0
0.04
0.08
0.12
s
?C p
? = 180o
?? = 0.12?? = 0.18
?? = 0.24?? = 0
?1?0.75?0.5?0.2500.250.50.751
?0.04
0
0.04
0.08
0.12
s
?C p
? = 0o
?? = 0.12?? = 0.18
?? = 0.24?? = 0
(b) ? = 30?
?1 ?0.75?0.5?0.250 0.250.5 0.75 1
?0.04
0
0.04
0.08
0.12
s
?C p
? = 180o
?? = 0.12?? = 0.18
?? = 0.24?? = 0
?1?0.75?0.5?0.2500.250.50.751
?0.04
0
0.04
0.08
0.12
s
?C p
? = 0o
?? = 0.12?? = 0.18
?? = 0.24?? = 0
(c) ? = 60?
Figure 4.3: Effects of changing airspeed on pressure distributions, at fixed angle of
attack
224
?1 ?0.75?0.5?0.250 0.250.5 0.75 1
?0.04
0
0.04
0.08
s
?C p
? = 180o
?? = 0.12?? = 0.18
?? = 0.24?? = 0
?1?0.75?0.5?0.2500.250.50.751
?0.04
0
0.04
0.08
s
?C p
? = 0o
?? = 0.12?? = 0.18
?? = 0.24?? = 0
(d) ? = 90?(edgewise flow)
Figure 4.3: Effects of changing airspeed on pressure distributions, at fixed angle of
attack
diffuser exit plane (s = 1). The greatest suction is seen just slightly upstream of the
blade passage region, and this is due to high-velocity air leaking through the gap
between the blade tips and the shroud wall, from the higher-pressure region below
the rotor to above it. The suction pressures over the shroud inlet are greater than
those on the diffuser wall, and are the reason for the significant thrust contribution
of the shroud in hovering flight.
In translational flight, due to the impingement of the free-stream air on the
shroud, stagnation points are formed on the most forward portion of the shroud
inlet (s = ?0.5, ?lip = 90?), which faces directly into the oncoming flow. The
stagnation points, which separate the internal and external flows over the shroud,
lie at the center of regions of pressure greater than ambient. These over-pressure
regions become stronger and larger as airspeed increases, and replace the suction
pressures seen over the forward portions of inlet in hover. As will be seen in the
performance data, this results in a severe reduction in the thrust produced by the
225
? = 180o
?? = 0.12?? = 0.18
?? = 0   
? = 0o
(a) ? = 0?(axial flow)
? = 180o
?? = 0.12?? = 0.18
?? = 0.24?? = 0   
? = 0o
(b) ? = 30?
? = 180o
?? = 0.12?? = 0.18
?? = 0.24?? = 0   
? = 0o
(c) ? = 60?
Figure 4.4: Effects of changing airspeed on pressure distributions, at fixed angle of
attack: 2-D depictions
226
? = 180o
?? = 0.12?? = 0.18
?? = 0.24?? = 0   
? = 0o
(d) ? = 90?(edgewise flow)
Figure 4.4: Effects of changing airspeed on pressure distributions, at fixed angle of
attack: 2-D depictions
shroud (Figure 4.6a). Suction pressures are still seen near the blade passage region,
in the diffuser and on the outermost parts of the shroud inlet (s < ?0.7) where
the external flow accelerates past the outer surface of the shroud; however, these
surfaces are inclined with their normal vectors pointing primarily radially inwards
or outwards, and do not contribute much to the vehicle thrust.
As in hover, maximum suction is seen just above the blade passage region, at
s ? ?0.05 (?lip ? 10?). With increasing airspeed, both suction and over-pressures
increase in magnitude, and the location of the over-pressure region on the inlet
moves slightly inwards. At the diffuser exit plane, the pressure approaches the
ambient atmospheric value, but, unlike in hover, does not actually reach it. The
difference increases with increasing airspeed, and indicates that further expansion of
the rotor downwash occurs beyond the diffuser exit plane. Note that the assumption
of complete expansion of the flow at the diffuser exit plane is the cornerstone of
the simple Momentum Theory model for shrouded-rotor analysis. For short-chord
227
?1 ?0.75?0.5?0.250 0.250.5 0.75 1?0.04
?0.02
0
0.02
0.04
0.06
0.08
0.1
s
?C p
? = 180o
? = 0o? = 30o
? = 60o? = 90o
?1?0.75?0.5?0.2500.250.50.751?0.04
?0.02
0
0.02
0.04
0.06
0.08
0.1
s
?C p
? = 0o
? = 0o? = 30o
? = 60o? = 90o
(a) v? = 10 ft/s (?prime = 0.12)
?1 ?0.75?0.5?0.250 0.250.5 0.75 1
?0.04
0
0.04
0.08
0.12
s
?C p
? = 180o
? = 0o? = 30o
? = 60o? = 90o
?1?0.75?0.5?0.2500.250.50.751
?0.04
0
0.04
0.08
0.12
s
?C p
? = 0o
? = 0o? = 30o
? = 60o? = 90o
(b) v? = 15 ft/s (?prime = 0.18)
?1 ?0.75?0.5?0.250 0.250.5 0.75 1
?0.04
0
0.04
0.08
0.12
s
?C p
? = 180o
? = 30o? = 60o
? = 90o 
?1?0.75?0.5?0.2500.250.50.751
?0.04
0
0.04
0.08
0.12
s
?C p
? = 0o
? = 30o? = 60o
? = 90o 
(c) v? = 20 ft/s (?prime = 0.24)
Figure 4.5: Effects of changing angle of attack on pressure distributions, at fixed
airspeed
228
shrouds like the one tested in this investigation, the measurements show that such
an assumption may not be valid.
When the angle of attack is increased to 30? (Figures 4.3b, 4.4b), the region
of suction pressure expands on the windward side of the shroud (? = 180?) and
shrinks on the downwind side (? = 0?), on both the inlet and diffuser surfaces.
The stagnation points on the inlet lie at approximately s = ?0.67 to ?0.85 (?lip =
120??153?) on the windward side, and at approximately at s = ?0.3 to ?0.4 (?lip =
54??72?), which coincides with the direction of the oncoming free-stream flow. The
stagnation points move inwards as the airspeed increases, indicating a decrease in
the capture area of the rotor. The suction pressures on the windward side of the inlet
are seen to be much greater than those in hover; this is probably due to the increased
turning of the flow from the free-stream direction to the rotor axial direction. On
the downwind side, suction is only seen in small regions near the blade tips, over
the outermost portion of the inlet, and near the diffuser exit. Over-pressures are
seen in the diffuser just below the rotor disk due to the pressure jump through the
rotor. The extreme asymmetry of the pressure distribution, compared to the hover
condition, results in a strong nose-up pitch moment on the shrouded rotor, as will be
seen later when the performance measurements are discussed (Figures 4.11d, e, g).
As in the axial-flow case, the magnitudes of the suction and over-pressures increase
with increasing airspeed. At the diffuser exit plane, the flow approaches ambient
pressure levels more closely on the downwind surface of the diffuser, while remaining
considerably lower on the windward side. In general, suction pressures are seen on
the surfaces of the shroud that are oriented away from the oncoming flow, while
229
overpressures are seen on the surfaces facing the flow. Similar pressure distributions
would be expected on the outer surfaces of the shroud, where measurements were
not made. This aspect of the pressure dissymmetry would manifest itself as a force
on the vehicle perpendicular to the rotor axis (Figures 4.7c, e).
As the angle of attack is increased to 60? (Figures 4.3c, 4.4c) and 90? (edgewise
flow, Figures 4.3d and 4.4d), the trends of increasing suction on the windward side of
the shroud and increasing over-pressure on the downwind side are seen to continue.
At these angles, no stagnation point is seen to lie within the range of pressure taps
on the windward side. On the downwind side, the stagnation point lies between
s = ?0.2 and ?0.3 (?lip = 36??54?) at ? = 60?, and near s = ?0.2 at ? = 90?. At
? = 90?, suction pressures cover the entire windward surface of the inlet, just like
in hover; however, the magnitude of these suction pressures are now much greater,
and are spread more evenly over the entire surface instead of being concentrated
near the rotor disk. Although large over-pressures are seen on the downwind side
of the inlet, the net effect is to increase the thrust produced by the shroud over
that produced in hover. Once again, this will be clearly seen in the performance
data (Figure 4.6b). From ? = 60? to 90?, the trend of increasing magnitude in the
surface pressures with increasing airspeed is seen to reverse over the windward side
of the shroud inlet. Over the remaining regions of the shroud surface, however, the
trend remains consistent.
The effect of changing angle of attack on the pressure distributions is more
clearly illustrated in Figures 4.5a?c, at each of the three airspeeds tested. In each
of the figures, the same trends are seen: with increasing angle of attack, the region
230
of suction pressures on the windward side of the inlet expands, while shrinking on
the inner portion of the downwind side. The maximum suction on the windward
side occurs between ? = 30? and 60?, and thereafter decreases in magnitude. In the
diffuser, pressure increases with increasing angle of attack (i.e., decreasing suction
or increasing over-pressure). On the windward side of the diffuser, the internal flow
recovers most closely to ambient pressure at?= 90?, while on the downwind side, the
departure from ambient levels at the diffuser exit plane is approximately the same
for all angles of attack. At ? = 180?, the magnitude of the pressure jump through
the rotor disk seems to be unaffected by the magnitude of the suction pressure above
the rotor. At ? = 0?, however, the magnitude of the pressure jump decreases with
increasing airspeed and with increasing ?, i.e., with increasing over-pressure above
the rotor.
4.3 Performance measurements
The performance data are plotted in Figures 4.6?4.12 as functions of the air-
speed ratio, ?prime, at constant angle of attack, for analyzing the effects of free-stream
velocity, and as functions of ? at constant airspeed ratio for analyzing the effects
of angle of attack. In all the graphs, the data for 4000 rpm are shown by the
solid (filled) symbols, those for 3000 rpm by the shaded-gray symbols, and those for
2000 rpm by the ?empty? symbols. When the coefficients are plotted versus?prime, it was
seen that, for a given angle of attack, the values are nearly independent of the rotor
rotational speed, and vary close to quadratically or cubically with the airspeed ratio.
231
Therefore, to more easily discern the trends followed by the variables with changing
airspeed, polynomial fits have been applied to the data for each angle of attack, and
these are shown in the figures superimposed on the measured values. For the force
coefficients (thrust, normal force, lift and drag), quadratic polynomials were found
to fit the data well, while for the power coefficient, cubic polynomials gave the best
fit. For the pitch moment coefficient, quadratic polynomials fit the data well at low
angles of attack (Figures 4.11b, d). However, at the highest angles of attack (? ?
75?), cubic polynomials were required to match the data well (Figures 4.11c, e). For
the location of the center of pressure, cubic polynomials generally gave the best fit
(Figures 4.10b?c). As would be expected, the polynomial fits diverge widely beyond
the limits of the measured data, in most cases, and cannot be considered as reliable
estimates for extrapolation. When the coefficients are plotted versus ?, no such
polynomial variation can be assumed, so the measured values are shown connected
by straight-line segments.
For each of the seven performance variables, the discussion of the measure-
ments follows the following order: the effect of increasing airspeed on the open and
shrouded rotors is compared, in the climb or axial-flow condition (? = 0?) and the
edgewise-flow condition (? = 90?), then the effects of increasing airspeed at the
other, intermediate angles of attack are discussed, and finally the effects of changing
angle of attack at constant airspeed. In the figures showing the data for the shrouded
rotor, the lines showing the trends for the open rotor have been superimposed for
comparison of the shrouded and unshrouded cases.
232
4.3.1 Thrust
Figure 4.6a shows the thrust of the open and shrouded rotors in the axial-flow
condition (?= 0?). With increasing airspeed, the thrust of the open rotor decreases,
as would be expected because of the increasing inflow while the rotor remains at
fixed pitch, thereby reducing the blade section angles of attack. The shrouded rotor
produces greater thrust than the open rotor in hover, but exhibits a much faster
drop-off in thrust with increasing airspeed. By an airspeed ratio of about 0.075 (J =
0.24), the shrouded rotor?s thrust has dropped below that of the open rotor. As was
seen from the pressure data, a significant contribution to this reduction in thrust
comes from the collapse of the suction forces on the shroud inlet and the increase in
over-pressure with increasing airspeed. Essentially, the shroud behaves like a bluff
body in the flow. As can be seen from Figures 2.1 (p. 108) and 2.5 (p. 119), the
shroud model tested has a section profile that is not a streamlined airfoil-like shape.
Flow separation off the trailing ?edge? of the diffuser may be causing substantial base
drag on the model. This loss of shroud thrust, and the actual creation of additional
drag by the shroud, is the most probably cause for the sharper reduction in thrust
of the shrouded rotor, as compared to the open rotor.
Figure 4.6b shows the open and shrouded rotors compared in the edgewise-
flow condition (? = 90?). As airspeed increases, the in-plane component of the local
velocity at the blades of the open rotor increases, leading to an increase in thrust.
The thrust of the shrouded rotor also increases with increasing?prime, but at a seemingly
higher rate than that of the open rotor, so that the difference between the open-
233
0 0.1 0.2 0.3 0.4?0.03
?0.02
?0.01
0
0.01
0.02
0.03
??
C T
Shrouded Rotor, 4000 RPMShrouded Rotor, 3000 RPM
Shrouded Rotor, 2000 RPMShrouded Rotor, Polynomial fit
Free Rotor, 3000 RPMFree Rotor, 2000 RPM
Free Rotor, Polynomial Fit
(a) Variation with airspeed: axial flow (? = 0?)
0 0.1 0.2 0.3 0.40
0.01
0.02
0.03
0.04
0.05
0.06
??
C T
Shrouded Rotor, 3000 RPMShrouded Rotor, 2000 RPM
Shrouded Rotor, Polynomial fitFree Rotor, 3000 RPM
Free Rotor, 2000 RPMFree Rotor, Polynomial Fit
(b) Variation with airspeed: edgewise flow (? =
90?)
0 0.1 0.2 0.3 0.4?0.01
0
0.01
0.02
0.03
C T
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(c) Variation with airspeed: open rotor
0 0.1 0.2 0.3 0.4
?0.02
0
0.02
0.04
0.06
C T
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(d) Variation with airspeed: shrouded rotor
0 30 60 90?0.01
0
0.01
0.02
0.03
C T
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
(e) Variation with angle of attack: open rotor
0 30 60 90
?0.02
0
0.02
0.04
0.06
C T
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
?? = 0.14?? = 0.18
(f) Variation with angle of attack: shrouded
rotor
Figure 4.6: Translational flight: variations in thrust
234
and shrouded-rotor thrusts actually increases with increasing airspeed ratio. This,
as was mentioned earlier, is probably because of the greatly increased suction on
the windward side of the shroud inlet.
Figure 4.6c shows the variation of thrust with airspeed for the open rotor at all
angles of attack. It is seen that the change from decreasing thrust (with increasing
?prime) to increasing thrust occurs at an angle of attack between 60? and 75?. Figure 4.6d
presents the data for the shrouded rotor in the same way, with the trendlines for the
open rotor from Figure 4.6c superimposed for comparison. For the shrouded rotor,
the change from decreasing to increasing thrust occurs at an angle of attack between
75? and 90?. This implies that at a certain value of ?, which lies between 60? and
75? for the open rotor, and near 75? for the shrouded rotor, the thrust coefficient is
independent of airspeed; this is confirmed in Figures 4.6e and 4.6f, which plot the
thrust data for the open and shrouded rotors versus angle of attack, for the different
airspeed ratios tested. These figures also illustrate how the thrust decreases as the
rotor pitches forward from hover into the wind, with a more rapid drop-off as the
speed of the relative wind increases. The shrouded rotor is seen to experience a
more rapid drop-off than the open rotor, at the same airspeed ratio.
4.3.2 Normal force
The normal force is the force perpendicular to the thrust axis, in the body
frame of reference. Figure 4.7a shows the effect of increasing airspeed on the normal
force coefficient of the open and shrouded rotors in the edgewise-flow condition. For
235
both configurations, CN increases with increasing airspeed, but much more so for
the shrouded rotor. This is because of the substantial bluff-body drag of the shroud
when broadside to the oncoming flow, and also due to the ?momentum drag? that
arises from the turning of the flow from the free-stream direction to the rotor axial
direction [125].
Figures 4.7b and 4.7c show the variation in CN with ?prime for the open and
shrouded rotors at the other angles of attack, and Figures 4.7d and 4.7e cross-plot
the data as functions of ? for different airspeed ratios. At ? = 0? (axial flight),
the normal force is zero. It increases with increasing angle of attack, substantially
more for the shrouded rotor than for the open rotor, before reaching a maximum
at around ? = 75? for the open rotor, and between 60? and 75? for the shrouded
rotor, before decreasing again at ? = 90? (edgewise flow). In general, the effect of
changing?is more pronounced at higher?prime, and for the shrouded rotor as compared
to the open one.
4.3.3 Lift
Equation 2.1 (p. 118) shows that at low ?, the lift coefficient is more greatly
influenced by the variations in CN, while at high ? it is more strongly affected by
the variations in CT. At ? = 0? (axial flight), CL = CN = 0, while at ? = 90?
(edgewise flow), CL = CT, and hence the variations of CL with airspeed for the
open and shrouded rotors are given by Figure 4.6b.
For the other angles of attack, Figures 4.8a and 4.8b show that, for both the
236
0 0.1 0.2 0.3 0.4
0
0.02
0.04
0.06
??
C N
Shrouded Rotor, 3000 RPMShrouded Rotor, 2000 RPM
Shrouded Rotor, Polynomial fitFree Rotor, 3000 RPM
Free Rotor, 2000 RPMFree Rotor, Polynomial Fit
(a) Variation with airspeed: edgewise flow (? =
90?)
0 0.1 0.2 0.3 0.4
0
4
8
12
16x 10?3
C N
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(b) Variation with airspeed: open rotor
0 0.1 0.2 0.3 0.4
0
0.02
0.04
0.06
0.08
0.1
C N
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(c) Variation with airspeed: shrouded rotor
0 30 60 90
0
4
8
12
16x 10?3
C N
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
(d) Variation with angle of attack: open rotor
0 30 60 90
0
0.02
0.04
0.06
0.08
0.1
C N
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
?? = 0.14?? = 0.18
(e) Variation with angle of attack: shrouded
rotor
Figure 4.7: Translational flight: variations in normal force
237
0 0.1 0.2 0.3 0.4
0
0.01
0.02
0.03
C L
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(a) Variation with airspeed: open rotor
0 0.1 0.2 0.3 0.4
0
0.02
0.04
0.06
C L
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(b) Variation with airspeed: shrouded rotor
0 30 60 90
0
0.01
0.02
0.03
C L
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
(c) Variation with angle of attack: open rotor
0 30 60 90
0
0.02
0.04
0.06
C L
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
?? = 0.14?? = 0.18
(d) Variation with angle of attack: shrouded
rotor
Figure 4.8: Translational flight: variations in lift
238
shrouded and open rotors,CL decreases with increasing?prime for??45?, and increases
with increasing ?prime for ? ? 60?. Figures 4.8c and 4.8d show that lift increases with
increasing ? at fixed airspeed, with more rapid rises as the airspeed is increased. At
any given airspeed and angle of attack, the shrouded rotor produces more lift than
the open rotor.
4.3.4 Drag
At ? = 0? (axial flight), CD = -CT, and the variations of drag coefficient with
airspeed for the open and shrouded rotors can be inferred from Figure 4.6a. In
climb, the shrouded rotor initially has lower drag because of its greater thrust, but
as airspeed increases, its drag becomes far greater than that of the open rotor. At
? = 90? (edgewise flow), CD = CN, and the comparison of the two configurations is
given by Figure 4.7a. Figures 4.9b and 4.9d show that, at almost all flight conditions,
the drag of the shrouded rotor is much higher than that of the open rotor.
Figures 4.9a and 4.9b show that, for both the open and shrouded rotors, CD
increases with increasing ?prime, more rapidly for the shrouded rotor, and also more
rapidly as ? is reduced. For the open rotor, this results in the minimum CD for
a given airspeed occurring at ? = 0? for low ?prime, but at higher ? for higher ?prime
(Figure 4.9c). For the shrouded rotor, at high ? and high ?prime, the drag behavior is
seen to change, from increasing with increasing ? to decreasing with increasing ?
(Figure 4.9d). Although such behavior is unexpected for a shrouded rotor, wind-
tunnel tests on annular airfoils [26, 55] have shown a similar maximum in drag
239
0 0.1 0.2 0.3 0.4?0.02
?0.01
0
0.01
0.02
C D
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(a) Variation with airspeed: open rotor
0 0.1 0.2 0.3 0.4?0.04
0
0.04
0.08
C D
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(b) Variation with airspeed: shrouded rotor
0 30 60 90?0.02
?0.01
0
0.01
0.02
C D
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
(c) Variation with angle of attack: open rotor
0 30 60 90?0.04
0
0.04
0.08
C D
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
?? = 0.14?? = 0.18
(d) Variation with angle of attack: shrouded
rotor
Figure 4.9: Translational flight: variations in drag
240
coefficient at an angle of attack between 55? and 75? followed by a drop-off towards
higher values of ?. This suggests that the thrust of the model used in these tests
might have been low enough that the annular-airfoil behavior of the shroud itself
could be seen at the higher angles of attack.
4.3.5 Location of center of pressure
Figure 4.10a shows that, in the edgewise flow condition (? = 90?), the center
of pressure moves from above the rotor disk (xcp> 0) to below it (xcp< 0) with
increasing airspeed. This, as is explained in Reference [125], is due to the greater
increase in bluff-body drag, which scales with the square of the free-stream velocity,
compared to the momentum drag, which increases linearly with v?. Figure 4.10a
also shows that, at any given airspeed, the center of pressure of the shrouded rotor
lies about 0.75R further above the corresponding location for the open rotor.
The decrease in xcp with increasing airspeed is also seen at the other angles
of attack, for both the open rotor (Figures 4.10b, c) and the shrouded rotor (Fig-
ures 4.10d, e). At fixed ?prime, with increasing ?, xcp shows a tendency to move slightly
downstream for the open rotor, up to ? ? 30?, before moving upstream again (Fig-
ure 4.10f). For all but the lowest airspeed ratios, the center of pressure lies below the
rotor disk, which would lead to a nose-down pitch moment for the open rotor. For
the shrouded rotor (Figure 4.10g), the center of pressure almost always lies above
the plane of the rotor disk, leading to a nose-up pitch moment, and always moves
upstream with increasing ?.
241
0 0.1 0.2 0.3 0.4?0.5
0
0.5
1
1.5
??
x cp / R
Shrouded Rotor, 3000 rpmShrouded Rotor, 2000 rpm
Shrouded Rotor, 3rd order fitShrouded Rotor, 2nd order fit
Open Rotor, 3000 rpmOpen Rotor, 2000 rpm
Open Rotor, 3rd order fitOpen Rotor, 2nd order fit
(a) Variation with airspeed: edgewise flow (? =
90?)
0 0.1 0.2 0.3 0.4?0.8
?0.4
0
0.4
x cp / R
??
? = 15o? = 30o
? = 45o? = 60o
? = 75o? = 90o
(b) Variation with airspeed: open rotor,
2nd-order fit
0 0.1 0.2 0.3 0.4?0.8
?0.4
0
0.4
x cp / R
??
? = 15o? = 30o
? = 45o? = 60o
? = 75o? = 90o
(c) Variation with airspeed: open rotor,
3rd-order fit
0 0.1 0.2 0.3 0.4
0
0.4
0.8
1.2
x cp / R
??
? = 10o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(d) Variation with airspeed: shrouded rotor,
2nd-order fit
0 0.1 0.2 0.3 0.4
0
0.4
0.8
1.2
x cp / R
??
? = 10o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(e) Variation with airspeed: shrouded rotor,
3rd-order fit
Figure 4.10: Translational flight: variations in location of center of pressure
242
0 30 60 90?0.8
?0.4
0
0.4
x cp / R
? [deg]
?? = 0.18?? = 0.28
?? = 0.37?? = 0.12
?? = 0.18?? = 0.24
(f) Variation with angle of attack: open rotor
0 30 60 90?0.8
?0.4
0
0.4
0.8
1.2
x cp / R
? [deg]
?? = 0.18?? = 0.28
?? = 0.37?? = 0.12
?? = 0.18?? = 0.24
?? = 0.14?? = 0.18
(g) Variation with angle of attack: shrouded
rotor
Figure 4.10: Translational flight: variations in location of center of pressure
4.3.6 Pitch moment
The pitch moment, calculated at the rotor hub, is the product of the normal
force and the distance of the center of pressure from the plane of the rotor disk.
Figure 4.11a compares the effect of increasing airspeed on CM for the open and
shrouded rotors, in the edgewise flow condition (in axial flow, CM = CN = 0). For
the open rotor, as airspeed increases, there is initially a slight nose-up moment (CM
> 0), which eventually decreases and becomes a strong nose-down moment (CM <
0). This is because, at fixed?,CN increases with?prime whilexcp decreases from positive
values to increasingly negative ones (Figures 26, 39), so the maximum product lies
at some intermediate value of ?prime. For the shrouded rotor, on the other hand, the
increase in CN is roughly matched by the decrease in xcp, which remains positive,
so CM rapidly increases to large positive values and remains roughly steady there
as ?prime is increased further. This very large nose-up pitch moment is what makes it
difficult for a shrouded rotor to pitch forward into the wind to maintain hovering
243
position in a gust, and also to maintain a certain vehicle attitude to provide the
required thrust for forward flight.
For the other angles of attack, Figures 4.11b?e show the variation in CM with
?prime for the open and shrouded rotors. For the open rotor, as ? decreases, the initial
nose-up moment becomes less strong, and CM remains near zero for longer before
becoming negative at higher airspeed ratios (Figures 4.11b, c). For the shrouded
rotor, too, as ? decreases, the maximum nose-up moment becomes less, and CM
starts to decrease back towards zero as ?prime increases (Figures 4.11d, e).
At constant ?prime, CM for the open rotor becomes increasingly negative with
increasing ?, at higher values of ?prime, but at lower ?prime this trend starts to reverse and
the pitch moment becomes nose-up again (Figure 4.11f). Equivalently, when the
rotor in hover experiences a side-gust, at low gust velocities it will experience a
small nose-up moment, while at high gust velocities, it will experience a large nose-
down moment, which becomes less strong as the rotor pitches into the wind and ?
decreases. For the shrouded rotor (Figure 4.11g), CM is seen to increase strongly
with increasing ?.
4.3.7 Power
Figure 4.12a compares the effect of increasing airspeed on the power con-
sumption of the open and shrouded rotors in the climb condition (? = 0?), while
Figure 4.12b shows the open and shrouded rotors compared in the edgewise-flow con-
dition (? = 90?). In axial flow, at fixed collective, CP for both configurations drops
244
0 0.1 0.2 0.3 0.4?0.01
0
0.01
0.02
??
C M
Shrouded Rotor, 3000 rpmShrouded Rotor, 2000 rpm
Shrouded Rotor, 3rd order fitShrouded Rotor, 2nd order fit
Open Rotor, 3000 rpmOpen Rotor, 2000 rpm
Open Rotor, 3rd order fitOpen Rotor, 2nd order fit
(a) Variation with airspeed: edgewise flow (? =
90?)
0 0.1 0.2 0.3 0.4?0.006
?0.004
?0.002
0 
0.002 
C M
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(b) Variation with airspeed: open rotor,
2nd-order fit
0 0.1 0.2 0.3 0.4?0.006
?0.004
?0.002
0 
0.002 
C M
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(c) Variation with airspeed: open rotor,
3rd-order fit
0 0.1 0.2 0.3 0.4?0.005
0
0.005
0.010
0.015
0.020
0.025
C M
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(d) Variation with airspeed: shrouded rotor,
2nd-order fit
0 0.1 0.2 0.3 0.4?0.005
0
0.005
0.010
0.015
0.020
0.025
C M
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(e) Variation with airspeed: shrouded rotor,
3rd-order fit
Figure 4.11: Translational flight: variations in pitch moment
245
0 30 60 90?0.006
?0.004
?0.002
0 
0.002 
C M
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
(f) Variation with angle of attack: open rotor
0 30 60 90?0.01
0
0.01
0.02
C M
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
?? = 0.14?? = 0.18
(g) Variation with angle of attack: shrouded
rotor
Figure 4.11: Translational flight: variations in pitch moment
rapidly with increasing ?prime. In edgewise flow, the power consumption of the open
rotor continuously increases with increasing airspeed, while that of the shrouded
rotor remains constant up till about ?prime = 0.3, and then appears to decrease. In both
cases, the shrouded rotor always consumes less power than the open rotor.
Figures 4.12c and 4.12d show the variations of CP with increasing ?prime for the
other angles of attack. For the shrouded rotor, CP decreases at higher rates as ?
is decreased. For the open rotor, similar behavior is seen: as ? is decreased, CP
increases at lower rates with increasing ?prime, and then begins to decrease at higher
rates. Figures 4.12e and 4.12f show the data as functions of angle of attack. Both
open and shrouded rotors show an increase in power as ? is increased, with the rate
of increase being larger for higher values of ?prime. However, the increase in CP is much
more rapid for the open rotor than for the shrouded rotor. Additionally, just as
for the thrust coefficient, the power of the open rotor appears to be independent of
airspeed at an angle of attack between 60? and 75?. For the shrouded rotor, the
246
0 0.1 0.2 0.3 0.40
1
2
3
4
5x 10?3
??
C P
Shrouded Rotor, 4000 RPMShrouded Rotor, 3000 RPM
Shrouded Rotor, 2000 RPMShrouded Rotor, Polynomial fit
Open Rotor, 3000 RPMOpen Rotor, 2000 RPM
Open Rotor, Polynomial Fit
(a) Variation with airspeed: axial flow (? = 0?)
0 0.1 0.2 0.3 0.41
2
3
4
5
6
7x 10?3
??
C P
Shrouded Rotor, 3000 RPMShrouded Rotor, 2000 RPM
Shrouded Rotor, Polynomial FitOpen Rotor, 3000 RPM
Open Rotor, 2000 RPMOpen Rotor, Polynomial fit
(b) Variation with airspeed: edgewise flow (? =
90?)
0 0.1 0.2 0.3 0.4
0
2
4
6
x 10?3
C P
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(c) Variation with airspeed: open rotor
0 0.1 0.2 0.3 0.4
0
2
4
6
x 10?3
C P
??
? = 0o? = 15o
? = 30o? = 45o
? = 60o? = 75o
? = 90o
(d) Variation with airspeed: shrouded rotor
0 30 60 90
0
2
4
6
x 10?3
C P
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
(e) Variation with angle of attack: open rotor
0 30 60 90
0
2
4
6
x 10?3
C P
? [deg]
?? = 0?? = 0.18
?? = 0.28?? = 0.37
?? = 0?? = 0.12
?? = 0.18?? = 0.24
?? = 0.14?? = 0.18
(f) Variation with angle of attack: shrouded
rotor
Figure 4.12: Translational flight: variations in power
247
data indicate that a similar independence from ?prime occurs at ? = 90?.
248
Chapter 5
Vehicle Configuration Trade Studies
5.1 Introduction
The previous chapters have shown that, especially in the hovering-flight con-
dition, significant improvements in performance are attainable by enclosing a MAV-
scale rotor within a shroud. However, a serious drawback of the shrouded-rotor
configuration is the the weight of the shroud itself, and it is imperative to take this
factor into consideration when deciding between the open-rotor and the shrouded-
rotorconfigurationsduringthedesignofamicroairvehicle. Inaddition, theaddition
of a shroud increases the size of the vehicle ? with an LR13 shroud, for example,
the outer diameter of the rotor tested in this investigation increased from 6.3 in to
9.6 in. A trade study should also therefore be conducted between using a shrouded
rotor with a smaller rotor diameter versus using an open rotor with a larger rotor
diameter. Based on the data that were acquired in this investigation, this chap-
ter presents such trade studies to determine whether it is worthwhile to choose
the shrouded-rotor configuration for the design of a MAV for missions consisting
primarily of hovering and low-speed flight.
249
5.2 Shroud weights
In order to conduct these trade studies, estimates are needed for the weights of
the shrouds. The weights of the shroud models used in this investigation ? 260 g for
the LR13-D10-L72, and 130 g for the LR13-D10-L31 shroud, for example ? cannot
be used, since the material for their fabrication was chosen from considerations
of ease of manufacture and rapid-prototyping, and these weights are therefore not
representative of the shroud on an actual flying MAV. Better estimates can be
obtained from the weights of the shrouds used on the ?TiFlyer? shrouded-rotor MAVs
tested and flown by Sirohi et al. [23] and Hrishikeshavan and Chopra [25]. Unlike
the ?solid-block? shrouds used in this research, these shrouds had a ?shell? structure
made from plies of carbon-fiber composite material (Graphite/Epoxy). From the
geometry of these shrouds, the area density of the material/construction can be
calculated, and applied to the geometries of the shrouds tested in this investigation.
Two of these ?shell? shrouds were formed from continuous layers of carbon composite
and had area densities of 0.51 kg/m2 and 0.70 kg/m2, the difference arising from
the greater number of plies used to increase the stiffness of the second shroud.1
Two other shrouds used a different construction method: that of stretching mylar
or cellophane fabric over a carbon-composite frame, to obtain a stiff structure while
reducing weight. These shrouds had area densities of 0.59 kg/m2 and 0.16 kg/m2,
respectively. For the trade studies, the average of these different area-density values
? 0.49 kg/m2 ? has been chosen as the representative value.
1The former shroud suffered from ?severe vibration problems ... traced to the low stiffness of
the [shroud] in an out-of-plane torsion mode? and consequent ?interference with the rotor? [23].
250
5.3 Trade studies
The three trade studies described below have been conducted making use of the
concept of the generalized figure of merit, FM? = FM??d, which was introduced in
Chapter 3 of this dissertation. Equation (3.3) (p. 145) from that chapter is repeated
here, after manipulating it so as to directly express the generalized figure of merit
in terms of the rotor diameter instead of the rotor disk area:
FM? = T
3/2
P?pi?DR (5.1)
Comparisons at stipulated values of thrust, powerorrotor diameter can thusberead-
ily obtained from the ratios of the FM? values of different rotors, open or shrouded.
As discussed in Chapter 3 and in Appendix B, the optimal operating point for any
rotor is that at which it achieves its maximum value of FM?, and the comparisons in
these trade studies have been made for the different rotors operating at that point.
5.3.1 Shrouded vs. open, using the baseline 6.3-inch rotor
The first trade study compares the baseline, 6.3-inch-diameter open rotor
tested in this investigation, which had a maximum FM? of 0.177, against the base-
line rotor when coupled with the LR13-D10-L50 shroud, which had a maximum
FM? of 0.470, the highest reached by any of the shrouds with this rotor. Based
on a shroud area density of 0.49 kg/m2, the weight of the shroud works out to be
251
0 4 8 12 16 20?50
0
50
100
150
200
250
Power [W]
Thrust [g]
Open rotorShrouded rotor
Shrouded rotor, after subtracting weight of shroud
(a) Shrouded vs. open, using the baseline 6.3-inch
rotor.
0 4 8 12 16 20?50
0
50
100
150
200
250
Power [W]
Thrust [g]
Open rotorShrouded rotor
Shrouded rotor, after subtracting weight of shroud
(b) Shrouded vs. open, using the cambered
6.3-inch rotor.
0 4 8 12 16 20?50
0
50
100
150
200
250
300
Power [W]
Thrust [g]
9.5?inch open rotor6.3?inch shrouded rotor
Shrouded rotor, after subtracting weight of shroud
(c) 6.3-inch shrouded rotor vs. 9.5-inch open
rotor.
Figure 5.1: Trade studies comparing open- and shrouded-rotor configurations for
MAVs.
252
41 g. Figure 5.1a shows plots of the resulting thrust-power characteristics2, calcu-
lated using Eq.(5.1), of the open rotor, the shrouded rotor, and the shrouded rotor
after subtracting the weight of the shroud, to better illustrate the net gains from
shrouding. It can be seen that for rotor system (including also the transmission,
attitude control devices, electronics, batteries and support structures) + payload
weights greater than approximately 45 g, the use of a shroud provides performance
gains, either in the form of decreased power requirements for the same payload, or
increased payload capacity at the same power.
5.3.2 Shrouded vs. open, using the cambered 6.3-inch rotor
Thesecondtradestudycomparesthemore-efficient, cambered6.3-inch-diameter
open rotor, which had a maximum FM? of 0.318, against the cambered rotor when
coupled with the LR13-D10-L31 shroud, which had a maximum FM? of 0.543. How-
ever, this shrouded-rotor model, as tested, had a blade tip clearance of 0.3%Dt.
Referring to the trends for the LR13-D10-L31 models given in Table 3.2 (p. 183),
reducing the tip clearance to 0.1%Dt could possibly increase the maximum FM?
of the shrouded rotor by 0.1?FM?OR,max, from 0.543 to 0.575. This higher value
has therefore been considered instead, for this trade study. Based on a shroud area
density of 0.49 kg/m2, the weight of the shroud works out to be 33 g. As before, the
thrust-power characteristics of the two configurations have been plotted, in Fig. 5.1b.
With this rotor and this shroud, the benefits from shrouding are obtained at rotor
2These plots are similar in appearance to the CT-versus-CP rotor polars that were shown in
Chapter 3; however, in a rotor polar, it is the collective angle that varies as a parameter along
the curve, whereas in these figures shown here, the collective angle is constant (at the point for
maximum FM?) and the rotor tip speed varies along the curve.
253
system + payload weights greater than approximately 70 g. As was realized earlier,
in Chapter 3, the amount of benefit that is obtained with this, more efficient rotor
is not as great as that obtained when the less efficient, baseline rotor is used.
5.3.3 6.3-inch shrouded rotor vs. 9.5-inch open rotor
The final trade study compares a shrouded rotor with a smaller rotor diameter
against an open rotor with a larger rotor diameter. For the 9.5-inch-diameter open
rotor, a FM? value of 0.460 has been used, corresponding to the highest figure of
merit of 0.65 achieved by an open rotor of this size [16], which utilized blades with
camber as well as taper and twist. Similarly, for the smaller rotor, the highest value
of FM that has been achieved at that size is 0.62 [20], corresponding to a FM? value
of 0.438. With the addition of a shroud to such a rotor, it can be expected that a
generalized figure of merit of 0.60 could be attained. This value has therefore been
used for the 6.3-inch-rotor-diameter shrouded rotor in this trade study. The thrust-
power characteristics for these two configurations are shown in Fig. 5.1c. Based on
these curves, it is clear that, from the standpoint of performance alone, for a given
overall vehicle size limit, it is always preferrable to use a larger-diameter open rotor
rather than a smaller-diameter shrouded rotor. In the end, however, it may be that,
as in almost all other aircraft that have utilized shrouded rotors, it will be the safety
benefits of the shroud, rather than the potential performance benefits, that will be
the deciding reason for choosing this configuration over an open rotor.
254
Chapter 6
Concluding Remarks
Micro unmanned aerial vehicles (MAVs) have been predicted to be the next,
emerging generation of aerospace systems, with many potential applications in both
civilian and military missions. For hover-capable aircraft, reasons of improved safety,
improved performance, and potentially lower noise emissions ? compared to a tradi-
tional ?open?-rotor design ? have led to the shrouded-rotor configurations becoming
one of the most popular choices in the design of vertical take-off and landing (VTOL)
MAVs. Although much work has been performed on larger-scale ducted propellers
and shrouded rotors, there existed scarce data on the performance of shrouded ro-
tors in the low-Reynolds-number regime of MAV-scale aircraft. In response to this
dearth of reliable information, a systematic, experimental investigation has been
conducted of the flight characteristics of MAV-scale aircraft of this configuration.
The investigation had a two-fold goal: first, to investigate the improvements in per-
formance that are achievable from using the shrouded-rotor configuration for micro
air vehicles, and second, to aquire a body of experimental data that could be used
for the validation of analytical, predictive tools, and as guidelines in the design of
more efficient, more capable MAVs.
The first phase of the research consisted of an experimental investigation of
the effects of varying the shroud profile shape on the hovering performance of MAV-
255
scale shrouded rotors, using seventeen models with a throat diameter (Dt) of 16 cm
and with various values of diffuser expansion angle (?d), diffuser length (Ld), inlet
lip radius (rlip) and blade tip clearance (?tip). The values of these shrouded-rotor
parameters that were tested were: from 0? to 20? for the diffuser angle, from 31%
to 72%Dt for the diffuser length, from 6.5% to 13%Dt for the inlet lip radius, and
from 0.1% to 1.6%Dt for the blade tip clearance. Measurements were made of
the thrust produced and mechanical power consumed, and, with the objective of
obtaining a better understanding of the physics of the flow-field around this aircraft
configuration, the distributions of pressure along the shroud surface and of the axial
velocities in the wakes of the models. Integration of the shroud surface pressure
distributions yielded an estimate of the fraction of the total thrust borne by the
shroud itself, while integration of the wake velocity distributions yielded an estimate
for the induced power losses of the shrouded rotors. The results of the various tests
were compared to the predictions of the momentum-theory analytical model for a
shrouded rotor.
This was followed by a wind-tunnel investigation ? the second phase of the
research ? of the performance of the unshrouded (?open?) rotor and of one of the
shrouded-rotor models in translational flight. The fixed-collective, rigid-rotor mod-
els were tested at tunnel speeds of up to 20 ft/s, corresponding to a maximum
airspeed ratio, ?prime = v?/?R, of 0.37 and angles of attack from 0? (axial climb or
?propeller mode?) to 90? (hovering flight in a cross-wind). Measurements were made
of thrust, shaft power, normal force, pitching moment and the pressure distribution
over the shroud surface. The most noteworthy observations that resulted from these
256
investigations are summarized below.
6.1 Conclusions
6.1.1 Hover: Performance measurements
1. At rotor collectives above 10?, all shrouded-rotor models showed improvements
in performance over the isolated (?open?) rotor ? that is, higher thrusts and
lower power requirements. At the same power consumption, increases in thrust
over the open rotor by up to 94% were observed, or, conversely, up to 49%
reductions in power at the same thrust, for the condition where the open and
shrouded rotors are required to have the same tip speed. If, instead, the rotor
disk areas are required to be the same, then up to 90% increases in thrust at
the same power are obtainable, or up to 62% reductions in power at the same
thrust.
2. The performance improvements of the shrouded rotors over the open rotor
obtained in these tests surpassed those predicted by momentum theory. This
is because, in comparing the two configurations, the total power requirements
of the rotors were considered, whereas momentum theory only considers the
ideal power requirements. In other words, shrouding the rotors had the effect
of reducing the non-ideal losses by a greater extent than the reductions in
the ideal losses. This is an important benefit for MAV-scale rotors, which
experience much greater profile-power losses than do larger-scale rotors.
257
3. Tests on a different micro-rotor, which had cambered blades and better open-
rotor performance than the baseline rotor, revealed performance improvements
from shrouding that, although also surpassing the predictions of momentum
theory, were not nearly as remarkable as the improvements observed with the
baseline rotor.
4. Within the range of values of the shrouded-rotor parameters tested, the opti-
mal configuration was that with ?tip = 0.1%Dt, rlip = 13%Dt, ?d = 10? and
Ld = 50% to 72%Dt. Increasing blade tip clearance, decreasing the inlet lip
radius, decreasing the diffuser length and increasing or decreasing the diffuser
angle from these optimum values led to degradations in performance. The im-
provement in performance with increasing expansion ratio (?d), as predicted
by momentum theory, was seen to occur up to an expansion ratio of around
1.2. Increasing ?d beyond this value, whether by increasing diffuser length or
diffuser angle, caused performance to deteriorate. These trends are similar,
qualitatively, to those seen in past research on larger-scale shrouded rotors.
5. The effects of changing any one of the four parameters (?tip, rlip, ?d, Ld) on
the shrouded-rotors? performance became more pronounced as the other three
parameters were changed to degrade the performance. It is therefore not pos-
sible to categorically state that any one particular parameter has a greater
effect on performance than any other, since the amount of influence of any
parameter depends on the values of the other parameters. The exception to
this seems to be the lip radius, changes in which had almost same effect no
258
matter what the values of the other parameters are.
Thus, in general, it would seem that changing the value of any shroud param-
eter so as to improve performance results in less sensitivity of the performance
to changes in any of the other parameters; similarly, improving the efficiency
of the rotor itself would seem to result in less of a performance benefit from
the presence of a shroud.
6. The traditional formulation of the figure of merit (FM) has been shown to
be inappropriate for comparing the efficiencies of open and shrouded rotors of
different expansion ratios. In its stead, an alternative, ?generalized? form of
the figure of merit, equal to FM???d or C3/2T /2CP, has been proposed, which
is applicable to any rotor, shrouded or open.
6.1.2 Hover: Flow-field measurements
1. Distinctprimaryandsecondarysuction-pressurepeakswereseenontheshroud
inlet; this phenomenon has been predicted by CFD analyses, but has not been
previously seen experimentally. The stronger secondary peak occurs due to
high-velocity vortical flow that forms at the gap between the blade tips and the
shroud wall. Tailoring the inlet shape in this region such that the resulting
suction forces are inclined more towards the axial direction can potentially
result in significant thrust increases.
2. Up to 80% of the total inlet thrust was produced by the inner half of the inlet
lips, which were semi-circular in profile. The outer portion of the lip bears
259
little load, and therefore can be constructed to be less stiff, resulting in weight
savings.
3. The negative thrust from the diffuser had magnitudes up to only about one-
fourth those of the positive thrust from the inlet, and therefore had less of an
effect on the net shroud thrust. This implies that maintaining attached flow
and favorable pressure distributions on the inlet would be more important
than doing so in the diffuser.
4. With increasing collective, the ability of the shroud to successfully off-load the
rotor decreased.
5. Increasing the blade tip clearance from 0.1%Dt to 1.6%Dt resulted in a reduc-
tion in inlet suction and the net shroud thrust.
6. Increasing the inlet lip radius from 6%Dt to 13%Dt reduced the strength of the
inlet suction, but, because of the simultaneous increase in inlet surface area,
resulted in a net increase in inlet and shroud thrust, and in greater off-loading
of the rotor.
7. Increasing the diffuser expansion ratio increased the suction pressures in the
diffuser, but had little effect on the inlet pressure distributions. For the same
diffuser length, the 20-degree diffuser produced up to four times as much
negative thrust as the 10-degree diffuser, with a consequent 20?30% reduction
in net shroud thrust. The increase in length of the 10-degree diffuser from
31%Dt to 72%Dt, on the other hand, while nearly doubling the negative thrust
260
from the diffuser, had little effect on the net shroud thrust.
8. The shrouds were effective in impoving the uniformity of the wake velocity dis-
tributions, compared to that of the open rotor, thus reducing both the ideal as
well as the non-ideal components of the induced power expenditure. Reduc-
ing the blade tip clearance resulted in greater uniformity and therefore lower
values of the induced power correction factor, ?. Increasing the expansion
ratio by increasing the diffuser length, at fixed diffuser angle, did not have a
significant effect on ?.
9. The diameter of the wake appeared to remain fairly constant after exiting the
diffuser, thereby validating the fundamental assumption of the momentum
theory model of a shrouded rotor: that the flow is fully expanded at the
diffuser exit plane. However, because of the axial velocity deficit in the core
of the wake due to the blockage from the centerbody, the actual achieved
expansion will be less than expected. A smooth aft fairing for the centerbody
is therefore of great importance.
10. The wake of the open rotor appeared to contract to a greater extent than that
usually seen for conventional-scale rotors.
6.1.3 Translational flight
1. In axial flight (? = 0?), the net thrust of the shrouded rotor was initially
greater than that of the open rotor, but rapidly decreased to being lower than
the open-rotor thrust as airspeed was increased, possibly because of the drag
261
of the shroud. The power consumption of both configurations also decreased
with increasing airspeed, but that of the shrouded rotor always remained less
than that of the open rotor.
2. Inedgewise-flowconditions(?=90?), thethrustofboththeopenandshrouded
rotors increased with increasing airspeed, with the thrust of the shrouded ro-
tor always remaining higher than that of the open rotor. This was probably
due to the large increases in suction pressure seen on the windward side of
the shroud inlet. The power consumption of the open rotor increased with
increasing airspeed, while that of the shrouded rotor remained roughly con-
stant and then decreased. The normal force on the shrouded rotor increased
much more rapidly with increasing airspeed than did that of the open rotor,
again probably because of the drag of the shroud. The location of the center
of pressure at which the normal force acts generally moved from above the
plane of the rotor disk to below it, for both configurations. However, for the
shrouded rotor, the center of pressure always lay approximately 0.75R further
above the location for the open rotor. Both of these factors contributed to
much larger nose-up pitching moments for the shrouded rotor than for the
open rotor. The physical basis of these pitching moments was clearly seen in
the extreme asymmetry of the flow-field, with large suction pressures on the
windward side of the shroud inlet and large over-pressures on the leeward side.
3. At an angle of attack between 60? and 75? for the open rotor, and between 75?
and 90? for the shrouded rotor, the data indicate that the thrust and power
262
coefficients of the rotors are independent of the free-stream velocity. At higher
angles of attack, CT and CP increased with increasing airspeed, while at lower
angles they decreased with increasing ?prime.
4. The normal force coefficient displayed a stall-like behavior with increasing
angle of attack, with a maximum value near ? = 75?. The location of the
center of pressure generally moved from below the rotor plane to above the
rotor plane as ? increased. The resultant effect was an increase in nose-up
pitch moment or a decrease in nose-down moment with increasing ?.
5. In general, the addition of the shroud caused the effects of changing airspeed
or angle of attack on the rotor forces and moments to become stronger. The
variations in power consumption were an exception to this rule.
6. Polynomial curve-fits showed the force and power coefficients to vary with
the square and cube, respectively, of airspeed ratio. The variations of the
location of the center of pressure and of the pitch moment coefficients were
best approximated with cubic polynomials.
7. At high tunnel speeds and high angles of attack, the shrouded rotor displayed
a decrease in drag coefficient with increasing?that is characteristic of annular
airfoils. This was possibly due to the relatively low thrust of the shrouded-
rotor model tested.
8. The pressure in the rotor wake at the diffuser exit plane more closely ap-
proaches ambient atmospheric levels on the downwind side of the diffuser,
263
compared to the windward side. At high airspeeds and at low angles of attack
(near-axial flow), the static pressure at the diffuser exit remains significantly
lower than ambient, indicating that further expansion of the wake takes place
beyond the diffuser. This observation challenges the validity of the fundamen-
tal assumption of the momentum-theory model for shrouded rotors, namely,
that the flow has fully expanded at the diffuser exit.
6.2 Contributions to state of the art
The contributions of this research to the state of the art lie in the value of
the extensive, systematic body of experimental data that has been generated, com-
prising both performance as well as flow-field measurements. The latter, especially,
have been instrumental in providing a better understanding of the physics of the
flow around a MAV-scale shrouded rotor. Unlike most previous research programs,
this data set contains valuable measurements of the mechanical power consumption
of the shrouded-rotor models, clarification of the effects of increasing power input
by increasing rotor collective rather than by increasing tip speed, data on the effects
of interference (or coupling) between the various shrouded-rotor parameters, and
comparisons with the performance of the open rotor over a collective range that
spans its entire operating envelope, thereby enabling comparisons of the open and
shrouded rotors at their respective points of optimal operation, for a true estimate
of the benefits that can be obtained from shrouding. The research program demon-
strated that significant improvements in performance can be obtained by shrouding
264
a rotor at the MAV scale; however, the amount of improvement becomes less when
the rotor itself is designed to be more efficient. In either case, though, the measured
improvements were greater than those predicted by momentum theory, due both to
substantial reductions in the rotor profile drag as well as increases in the unifor-
mity of the wake profiles ? non-ideal power losses that are not considered by the
momentum theory. This is a significant result, since the predictions of momentum
theory are usually considered to be upper limits of attainable performance. The
measurements of the shroud surface pressure distributions and wake axial velocity
distributions provide physical explanations for the performance variations; a notable
contribution of the pressure measurements is the first known experimental demon-
stration of the secondary suction peak on the inlet of the shroud, which has, so far,
only been predicted by CFD analyses. The changes in the pressure distributions
with changing airspeed and angle of attack in translational flight clearly illustrated
the sources of the high drag and nose-up pitch moment exerted on the shroud in this
flight condition, and also the reason for the increased thrust of the shrouded rotor
in edgewise flight. Finally, a new standard metric, the ?generalized figure of merit,?
has been proposed for correctly comparing the performances of open and shrouded
rotors of different expansion ratios.
In the end, however, the air-vehicle design trade studies that were performed
showed that, due to the due to the increases in weight and aircraft size that result
from the incorporation of a shroud, the performance benefits that would be obtained
for an actual aircraft might only be marginal, and that the safety benefits of a
shroud might be the principal reason for choosing this configuration rather than
265
a conventional open-rotor design. Still, the results of this research program are
useful in that they show how the shroud should be designed so as to minimize any
detrimental consequences of incorporating a shroud, and indeed, how to obtain the
best possible performance if other design requirements result in the choice of the
shrouded-rotor configuration.
6.3 Recommendations for future work
There are several avenues along which research could be conducted to extend
and improve upon the knowledge gained from this present investigation. The most
obvious would be to test a greater number of values of each of the different shrouded-
rotor parameters; in the hover tests, although data were obtained from multiple
series of models, each series had a maximum of three values of any of the parameters
of interest ? this is not sufficient for obtaining a true understanding of the effects
of varying those parameters on performance, and where the performance limits
lie. This is especially true in the case of the diffuser parameters ? the diffuser
length and expansion angle ? for which the data showed that an extreme point
lay somewhere within the range of values tested. As was done for the 10-degree
diffuser, the 20-degree diffuser model should be cut into sections to obtain a shorter
diffuser with the same expansion ratio as the longest 10-degree diffuser; this would
yield valuable information on the relative effects of expansion angle and expansion
ratio. Such a test was planned in this investigation, but was not carried out due
to time limitations. A follow-up investigation to this would be to test the effects of
266
varying the values of the shroud parameters on the translational-flight performance
of shrouded rotors. For those tests, it would also be worthwhile to measure the
pressure distributions over the external surfaces of the shrouds. In addition, for the
translational-flight tests, shrouds with tapered diffuser-exit ?trailing edges? should be
tested, to avoid the separation drag that is sure to have occurred on the models that
were tested in this investigation. These tests should also be conducted with a rotor
that has been optimized for the low-Reynolds-number flight regime, with blades
that incorporate taper and twist, and with thin, cambered airfoils with sharpened
leading and trailing edges.
Another research avenue that could be pursued is the performance of coaxial,
contra-rotating rotors when shrouded. Although most current shrouded-rotor UAVs
use vanes in the downwash of the rotor for yaw control and to nullify the reaction
torque of the rotor, it is possible that use of coaxial rotors, as were used in the
Sikorsky Cypher aircraft, could result in aircraft with greater capabilities and better
flight characteristics. For further innovation, a study of different inlet lip profile
shapes should be conducted, not remaining restricted to simple semi-circular lips,
with one goal being to harness the potential of the inlet suction-pressure peaks for
greater thrust production. Measurements should be made of the inflow velocity
distributions, to supplement those that have been made in the wakes of the rotors,
and thus to obtain a better understanding of the flow conditions at the rotor disk.
Finally, efforts should be made to develop techniques for predicting the perfor-
mance of shrouded-rotor MAVs, to relieve the amount of experimental testing that
would otherwise need to be conducted. This research direction would, of course,
267
rely heavily upon computational-fluid-dynamic (CFD) methods, but a fair amount
of insight can also be obtained by the more elementary potential-flow and blade-
element methods. For the latter, a model for the effect of the blade tip clearance
on the aerodynamics is essential, as well as reliable data on the performance of the
blade airfoils at these low Reynolds numbers.
268
Appendix A
Theoretical Derivations
This appendix provides the detailed derivations of the theoretical model used
for predicting the performance of a shrouded rotor. The Momentum Theory model
for the hover and climb (axial-flight) conditions is first given, followed by the Com-
bined Blade-Element-Momentum Theory (?BEMT?) model for the hover condition.
The derivations shown here have been presented for the principal purpose of com-
paring the aerodynamics of open and shrouded rotors. For a more in-depth analysis
and discussion of the aerodynamics of an open rotor, a standard textbook such as
Ref. [14] is recommended.
A.1 Momentum Theory
The ?simple? Momentum Theory model of a rotor is based on the conservation
laws of fluid dynamics, and uses the simplifying assumptions of inviscid, incompress-
ible, steady, and quasi-one-dimensional flow. The rotor is assumed to impart only
axial momentum to the flow, i.e., there is no swirl in the wake. Additionally, in the
case of a shrouded rotor, the flow is assumed to have expanded back to ambient
atmospheric pressure at the diffuser exit plane.
269
(a) Open rotor
(b) Shrouded rotor
Figure A.1: Flow-field models for open and shrouded rotors in hover.
270
A.1.1 Hover
Figure A.1 shows the flow-fields of an open (unshrouded) and a shrouded
rotor in the hovering-flight condition. Station 0 represents the free-stream, infinitely
upstream of the rotor. Stations 1 and 2, respectively, lie just above and below the
plane of the rotor disk. Station 3 is at the exit plane of the shroud?s diffuser or
the far wake, infinitely downstream, of an open rotor. The following identities and
definitions therefore arise:
Velocities: v0 = 0
v1 = v2 = vi
v3 = w
Areas: A1 = A2 = A
A3 = Ae = ?dA
Pressures: p0 = p3 = patm
?p = p2 ?p1
A.1.1.1 Open rotor
In the case of an open rotor (?OR?), the development of the model is as follows:
Conservation of Mass: ?m = ?Avi (A.1)
Conservation of Momentum: TOR = ?m(v3 ?v0) = ?mw (A.2)
Conservation of Energy: PiOR = KE3 ?KE0 = 12 ?mw2 (A.3)
Actuator-disk model of the rotor: TOR = ?pA (A.4)
271
PiOR = TORvi (A.5)
Actuator-diskmodelofrotorandConservationofEnergy(applicationoftheBernoulli
equation between stations 0 and 1, and between stations 2 and 3):
p0 + 12?v20 = p1 + 12?v21
p2 + 12?v22 = p3 + 12?v23
? ?p = 12?w2 (A.6)
By combining either equations A.2, A.3 and A.5, or equations A.1, A.2, A.4
and A.6, the relation between the induced velocity at the rotor disk plane and in
the far wake is obtained:
w = 2vi (A.7)
Together with equation A.2, this gives
vi =
radicalBigg
TOR
2?A (A.8)
and from either equation A.3 or A.5
PiOR = T
3/2
OR?
2?A (A.9)
272
A.1.1.2 Shrouded rotor
In the case of a shrouded rotor (?SR?), the development is as follows:
Conservation of Mass: ?m = ?Avi = ?Aew (A.10)
? w = vi?
d
(A.11)
Conservation of Momentum: Ttotal = ?mw = ?Av
2
i
?d (A.12)
? vi =
radicalBigg
?dTtotal
?A (A.13)
Conservation of Energy: Pi = 12 ?mw2 = 12?Av
3
i
?2d (A.14)
? Pi = T
3/2
total?
4?d?A (A.15)
Thus, the relation A.11 has replaced the result A.7 derived for the open rotor, and
the actuator-disk model of the rotor was not required in order to derive expressions
for the induced velocity (eq. A.13) and induced power (eq. A.15).
The actuator-disk model is only required to derive expressions for the the
thrust provided by the rotor alone:
Trotor = ?p?A or Piv
i
= 12?Aw2 (A.16)
? TrotorT
total
= 1/2?Aw
2
(?Avi)w =
1
2?d (A.17)
In non-dimensional coefficient form, the various quantities are expressed as
273
follows:
CTtotal = Ttotal?A(?R)2
CTrotor = Trotor?A(?R)2 = 12?
d
CTtotal
?i = vi?R =radicalbig?dCTtotal
= ?dradicalbig2CTrotor
(A.18)
CPi = Pi?A(?R)3 = C
3/2
Ttotal
2??d
= ?2?dC3/2Trotor
= ?iCTrotor
(A.19)
Fortheshroudedrotor, Ttotal ?TSR. Inthecaseofanopenrotor, Trotor =Ttotal ?TOR,
so, from equation A.17, ?d = 1/2, which is exactly the value of the ?expansion? that
momentum theory predicts for an open rotor (eq. A.7). Equations A.10?A.17, as
well as the coefficient forms of the quantities shown above, therefore apply equally to
both shrouded and unshrouded (open) rotors, with the effective value of 1/2 being
used for the expansion ratio in the case of an open rotor.
The contribution to the total shrouded-rotor thrust provided by the diffuser
alone can be found by applying the conservation laws to a control volume consisting
of the fluid within the diffuser only. The control volume is therefore bounded by
the diffuser walls, the rotor disk plane and the diffuser exit plane. Applying the
momentum equation and summing the forces in the z-direction, where +z is an
axial coordinate in the direction of the flow (opposite to the direction of positive
274
thrust on the rotor), we have:
summationdisplay
Fz =
summationdisplay
vz(?vectorA?vectorv)
?Fdiff +p2A2 ?p3A3 = ?v2?A2v2 +v3?A3v3
where Fdiff is the force exerted by the diffuser on the control volume
?Fdiff +p2A?patm?dA = ??Av2i +??dAv
2
i
?2d
?Fdiff = ?Av2i
parenleftbigg1
?d ?1
parenrightbigg
+A(?dpatm ?p2) (A.20)
Applying the Bernoulli equation (conservation of energy):
p2 + 12?v22 = p3 + 12?v23
= patm + 12?v
2
i
?2d
?p2 = patm + 12?v2i
parenleftbigg 1
?2d ?1
parenrightbigg
(A.21)
Substituting eq. A.21 into eq. A.20 and simplifying, we obtain:
Fdiff = ??Av
2
i (?d ?1)
2
2?2d +A(?d ?1)patm (A.22)
Note that ?Fdiff is the force exerted by the control volume on the diffuser, in the +z
direction. Therefore, the thrust (which is in the ?z direction) on the diffuser due to
the control volume is equal to ?(?Fdiff ). The net thrust on the diffuser is equal to
275
the sum of the forces (in the ?z direction) due to both the internal control volume
and the external atmosphere. This latter quantity is equal to ?patm(A3 ?A2) =
?patmA(?d ?1). Adding this to Fdiff from eq. A.22, we have:
Net thrust on diffuser = Tdiff = ??Av
2
i (?d ?1)
2
2?2d
Making use of eq. A.12, we obtain the final result:
Tdiffuser
Ttotal = ?
(?d ?1)2
2?d (A.23)
Given expressions for the thrust fractions from the rotor (eq. A.17) and the diffuser
(eq. A.23), the thrust fraction from the inlet is directly obtained:
Tinlet
Ttotal = 1?
Trotor
Ttotal ?
Tdiffuser
Ttotal
= ?d2 (A.24)
From eq. A.17, the thrust-sharing between the shroud and the rotor is:
Tshroud
Trotor =
1? 12?
d
1
2?d
= (2?d ?1) (A.25)
which shows that as the expansion ratio is increased, the shroud increasingly off-
loads the rotor.
276
A.1.1.3 Comparisons of shrouded-rotor and open-rotor performance
Expressions A.9 and A.15 describe the relationship between four fundamental
characteristics of open or shrouded rotors: thrust, power requirement, rotor disk
area and (effective) expansion ratio. By taking the ratio of these expressions:
PiSR
PiOR =
1?
2?d
parenleftbiggT
SR
TOR
parenrightbigg3/2parenleftbiggA
OR
ASR
parenrightbigg1/2
(A.26)
comparisons can be made between the characteristics of open and shrouded rotors,
as a function of the shrouded rotor?s expansion ratio. By holding any two variables
fixed between the two configurations, the behavior of the other variables can be
analysed. For example, if the open and shrouded rotors are compared, subject to
the condition that they have the same rotor disk area (ASR = AOR) and consume
the same amount of (ideal) power (PSR = POR), then the following relations result:
TSR
TOR = (2?d)
1/3
Trotor
TOR =
1
(2?d)2/3
viSR
viOR =
?mSR
?mOR = (2?d)
2/3
wSR
wOR =
1
(2?d)1/3 (A.27)
Similarly, if the configurations are compared when they are producing the same total
amount of thrust (TSR = TOR) and are of the same disk area, then:
PiSR
PiOR =
1?
2?d
277
Trotor
TOR =
1
2?d
viSR
viOR =
?mSR
?mOR =
?2?
d
wSR
wOR =
1?
2?d (A.28)
If the configurations are compared when they are producing the same total amount
of thrust and are consuming the same amount of ideal power, then:
ASR
AOR =
1
2?d ?
DSR
DOR =
1?
2?d
Trotor
TOR =
1
2?d
viSR
viOR = 2?d
?mSR
?mOR = 1
wSR
wOR = 1 (A.29)
And, if the configurations are compared when the rotors are producing the same
amount of thrust (Trotor = TOR) and are of the same disk area, then:
TSR
TOR = 2?d
PiSR
PiOR = 2?d
viSR
viOR =
?mSR
?mOR = 2?d
wSR
wOR = 1 (A.30)
278
A.1.1.4 Pressure distributions
The momentum-theory model can be used to obtain a prediction of the vari-
ations in air pressure through the flow-field of an open or shrouded rotor. For a
hovering rotor, the total pressure upstream of the rotor is equal to ambient at-
mospheric pressure (patm), while downstream of the rotor it is equal to the sum
of atmospheric pressure and the disk loading (DL = Trotor/Arotor ? Trotor/Athroat).
Expressing the pressures in terms of the pressure coefficient, Cp, the following de-
velopment is obtained. Upstream of the rotor, or at the shroud inlet:
p = ptotal ?q = patm ? 12?v2 (A.31)
?Cp = patm ?pq
tip
=
1
2?v
2
1
2?v
2
tip
=
parenleftbigg v
vtip
parenrightbigg2
=
parenleftbigg v
i
vtip
parenrightbigg2
?
parenleftbiggv
vi
parenrightbigg2
= ?2i ?
parenleftbiggv
vi
parenrightbigg2
= (?dCT)?
parenleftbiggv
vi
parenrightbigg2
(A.32)
where v is the flow velocity at any given axial station in the flow. Downstream of
the rotor, or in the diffuser of a shroud, the development is similar:
p = ptotal ?q = patm + DL? 12?v2 (A.33)
279
?Cp = patm ?pq
tip
=
1
2?v
2 ?DL
1
2?v
2
tip
=
parenleftbigg v
vtip
parenrightbigg2
? DLq
tip
=
parenleftbigg v
i
vtip
parenrightbigg2
?
parenleftbiggv
vi
parenrightbigg2
? Trotor/Athroat1
2?v
2
tip
= ?2i ?
parenleftbiggv
vi
parenrightbigg2
?2CTrotor
= (?dCT)?
parenleftbiggv
vi
parenrightbigg2
? 2CT2?
d
= (?dCT)?
bracketleftBiggparenleftbigg
v
vi
parenrightbigg2
? 1?2
d
bracketrightBigg
(A.34)
By mass conservation, v/vi = Athroat/A, where A is the cross-sectional area of
the flow-field at the corresponding axial station. Therefore, expressions are needed
for the shape of the flow-field boundary as a function of distance from the plane of
the rotor disk. For an open rotor (?d = 1/2), the geometry of the wake, downstream
of the rotor, can be estimated using a model such as the Landgrebe wake model
[14, p. 450] ? this was done for Fig. 1.8c on page 19, repeated here as Fig. A.2
for convenience. In this figure, the pressure variation upstream of the open rotor
was obtained by assuming that the cross-sectional area of the upstream flow varied
?linearly? from infinity to the area of the rotor disk, i.e., that v/vi varied linearly
from 0 to 1.
For a shrouded rotor, the geometry of the wake is defined by the shape of the
diffuser. For a conical diffuser, Athroat/A = (R/r)2, where R is the radius of the
shroud throat and r is the radius at any axial station. For such a diffuser, r can be
280
Far upstream Far wakeNormalized distance
Static pressure
Shrouded Rotor, ?d = 1.5Shrouded Rotor, ?
d = 1.0Open Rotor (?d = 0.5)
Rotor disk plane 
Ambient pressure 
Diffuser exit plane(Shrouded rotor)   
Outer edge of inlet lip (Shrouded rotor)
Figure A.2: Variations in pressure in the hover flow-fields of open and shrouded
rotors, at the same thrust coefficient.
expressed as:
r = R+z?(Re ?R) (A.35)
where Re = ??dR is the radius of the diffuser exit and z is a normalized axial
coordinate that varies linearly from 0 at the rotor disk plane to 1 at the diffuser
exit plane (Fig. A.3b). Using this expression, the velocity ? or area ? ratio,
downstream of the rotor, is given by:
v
vi =
Athroat
A =
1
[1 +z(??d ?1)]2 (A.36)
For the inlet, the inflow geometry can be defined using the ?sphere-cap? model
proposed by Dyer [123], illustrated in Fig. A.3. A sphere cap is the surface obtained
when a spherical shell is intersected by a plane, and the portion of the shell above
that plane (the ?base plane?) is considered (Fig. A.3a). The base plane may lie either
above or below the center of the sphere. Applied to an inlet flow model, a sphere
cap is used to represent a potential surface, normal to which are the streamlines of
281
(a) Sphere cap geometry. (b) Sphere cap geometry applied to shroud inlet.
Figure A.3: Sphere-cap model for inlet flow geometry of a shrouded rotor.
the flow. The sphere caps, shown in cross-section by the dashed line in Fig. A.3b,
are centered on the shroud axis and intersect the inlet lip surface at right angles.
The line of intersection is a circle that represents the base plane of the sphere cap.
The surface area of a sphere cap is given by the equation:
Acap = 2pirCh = pi(r21 +h2) (A.37)
where rC is the radius of the sphere, h is the height of the sphere cap, and r1 is the
radius of the base circle. Applied to a shroud inlet, r1 and h are given by:
r1 = rC sin? = R+rlip(1?cos?) (A.38)
h = rC(1?cos?) = r1sin?(1?cos?) (A.39)
where ? is an inlet angular coordinate that varies from 0? at the throat to 180? at
the outer edge of the inlet lip. Using these expressions, the velocity and area ratios,
282
upstream of the rotor, are given by:
v
vi =
Athroat
Acap =
1
bracketleftbig1 + rlip
R (1?cos?)
bracketrightbig2 ?bracketleftBig1 +parenleftbig1?cos?
sin?
parenrightbig2bracketrightBig (A.40)
Equations (A.40) and (A.36) can now be substituted into Eqs. (A.32) and (A.34), re-
spectively, to obtain the pressure distributions over the inlet and diffuser surfaces of
a shroud, as a function of the inlet lip radius, diffuser expansion ratio and thrust co-
efficient. Figure A.4 shows such predictions of pressure distributions, while Fig. A.5
compares these theoretical predictions with the experimentally measured distribu-
tions for shrouded-rotor models LR06-D00-?0.6 and LR13-D20-?0.8. In Figure A.5,
the theoretical distributions were generated using the experimentally-measured val-
ues of CT for the shrouded rotors at the different rotor collectives. It can be seen
that the theoretical model qualitatively captures the general shape of the pressure
distributions quite well, but under-predicts the strength of the suction peak on the
inlet lip. The prediction is a little closer to the experimental values in the case of the
model with the larger tip clearance (Fig. A.5b); this correlates well with the obser-
vation, described in Chapter 3, that decreasing the tip clearance causes an increase
in the inlet suction. It is therefore quite probable that incorporating a theoretical
model for the effect of the tip clearance on the pressure distributions would improve
the predictions.
Finally, the values of the pressure just above and below the plane of the rotor
disk (stations 1 and 2 in Fig. A.1) are obtained by setting v = vi in Eqs. (A.32) and
283
?1 ?0.75?0.5?0.250 0.250.5 0.75 10
0.005
0.01
0.015
0.02
0.025
Non?dimensional distance, s
?C p
rlip = 6%Dtr
lip = 9%Dtrlip = 13%Dt
(a) Effect of inlet lip radius; CT = 0.02, ?d=
1.25.
?1 ?0.75?0.5?0.250 0.250.5 0.75 1
0
0.01
0.02
0.03
0.04
Non?dimensional distance, s
?C p
?d = 1.0?
d = 1.5?d = 2.0
(b) Effect of diffuser expansion ratio; CT = 0.02,
rlip= 13%Dt.
?1 ?0.75?0.5?0.250 0.250.5 0.75 10
0.01
0.02
0.03
0.04
Non?dimensional distance, s
?C p
CT = 0.01C
T = 0.02CT = 0.03
(c) Effect of thrust coefficient; rlip= 13%Dt, ?d=
1.25.
Figure A.4: Theoretical predictions of shrouded-rotor surface pressure
distributions.
284
?1 ?0.75?0.5?0.250 0.250.5 0.75 10
0.01
0.02
0.03
0.04
0.05
0.06
s
?C p
10o ? Theory10o ? Expt
20o ? Theory20o ? Expt
30o ? Theory30o ? Expt
40o ? Theory40o ? Expt
(a) Model LR06-D00-?0.6.
?1 ?0.75?0.5?0.250 0.250.5 0.75 10
0.01
0.02
0.03
0.04
0.05
0.06
s
?C p
10o ? Theory10o ? Expt
20o ? Theory20o ? Expt
30o ? Theory30o ? Expt
40o ? Theory40o ? Expt
(b) Model LR13-D20-?0.8.
Figure A.5: Shroud surface pressure distributions: comparison of theory with
experimental measurements.
(A.34), to obtain:
p1 = patm ?DL?(?2d) (A.41)
p2 = patm ?DL?(?2d ?1) (A.42)
Substituting the value of 1/2 for ?d in these equations results in the familiar expres-
sions (patm?DL/4) and (patm +3DL/4), respectively, for an open rotor. Figure A.6
shows the effect of increasing the expansion ratio on the (suction) pressures at the
rotor disk plane, above and below the rotor disk, at fixed total thrust. The decreas-
ing distance between the two lines illustrates the reduction in thrust borne by the
rotor itself (Trotor) with increasing ?d.
285
0.5 1 1.5 2
?0.02
0
0.02
0.04
?d
?C p
Above rotorBelow rotor
Figure A.6: Effect of expansion ratio on pressures at the rotor disk plane; CT =
0.02.
A.1.2 Climb (Axial Flight)
The development of the climb model parallels that for the hover condition,
with the difference that the velocities at the different stations in Fig. A.1 are now
incremented by the climb velocity, Vc:
v0 = Vc
v1 = v2 = Vc +vi
v3 = Vc +w (A.43)
Therefore, for an open rotor:
Conservation of mass: ?m = ?A(Vc +vi) (A.44)
Conservation of momentum: TOR = ?m(v3 ?v0) = ?mw (A.45)
Conservation of energy: PiOR = 12 ?m(Vc +w)2 ? 12 ?mV2c
= 12 ?mw(w+ 2Vc)
(A.46)
286
Actuator-disk model: PiOR = TOR(Vc +vi) (A.47)
Solving Eqs. (A.45), (A.46) and (A.47) yields:
w = 2vi (A.48)
which is the same result as for the hover condition. Using this result with Eqs. (A.44)
and (A.45) yields the formula for the ideal induced velocity at the rotor disk:
vi
vh = ?
Vc
2vh +
radicalBiggparenleftbigg
Vc
2vh
parenrightbigg2
+ 1 (A.49)
where vh = radicalbigTOR/2?A is the ideal induced velocity at the rotor in the hover
condition (cf. Eq. (A.8)). Note that the ?effective? expansion ratio for the open
rotor is no longer equal to 1/2, as it was in hover, and is now given by:
?d = A3A = Vc +viV
c +w
= Vc +viV
c + 2vi
(A.50)
which has a value between 1/2 and close to 1.0, depending on the climb velocity.
On the other hand, for a shrouded rotor, the expansion ratio is fixed by the
geometry of the diffuser, and is forced to remain the same (equal to Ae/A) at all
climb velocities ? assuming complete expansion at the diffuser exit plane. For a
shrouded rotor, therefore:
Conservation of mass: ?m = ?A(Vc +vi) = ?Ae(Vc +w) (A.51)
287
?w = Vc +vi?
d
?Vc (A.52)
Conservation of momentum: Ttotal = ?mw (A.53)
Solving these equations, and using the result vh =radicalbig?dTtotal/?A, yields the formula
for the ideal induced velocity:
vi
vh =
Vc
2vh(?d ?2) +
radicalBiggparenleftbigg
?dVc
2vh
parenrightbigg2
+ 1 (A.54)
The division of thrust between the rotor and the shroud is found, once again,
by using the actuator-disk model of the rotor:
Trotor = ?p?A = (p2 ?p2)?A
Applying the Bernoulli equation between stations 0 and 1, and between stations 2
and 3:
p0 + 12?V2c = p1 + 12?(Vc +vi)2
p2 + 12?(Vc +vi)2 = p3 + 12?(Vc +w)2
Solving these yields, for the pressure jump across the rotor:
p2 ?p1 = 12?(Vc +w)2 ? 12?V2c
= ?w(w/2 +Vc) (A.55)
288
Using Eqs. (A.52) and (A.53), the rotor thrust fraction is therefore given by:
Trotor
Ttotal =
(p2 ?p1)A
?mw
= Vc +w/2V
c +vi
= vi +Vc(?d + 1)2?
d(vi +Vc)
(A.56)
Thus, even though the total thrust is constant in a steady (unaccelerated) climb, the
rotor thrust is not, but depends on the climb velocity. The change in rotor thrust
from the hover condition is readily obtained as:
(Trotor)c
(Trotor)h =
(Trotor/Ttotal)c
(Trotor/Ttotal)h
= vi +Vc(?d + 1)v
i +Vc
(A.57)
The (ideal) power required to climb, Pc, is given by the product of the rotor
thrust and the velocity of the air through the rotor:
Pc = (Trotor)c ?(vi +Vc)
= (Trotor)h ?bracketleftbigvi +Vc(?d + 1)bracketrightbig (By Eq. (A.57))
Denoting Ph = (Trotor)h ?vh = T3/2total/?4?d?A (cf. Eq. (A.15)) as the ideal power
required in hover, the climb power can be expressed as:
Pc
Ph =
vi +Vc(?d + 1)
vh
=
parenleftbigg
3?d Vc2v
h
parenrightbigg
+
radicalBiggparenleftbigg
?dVc
2vh
parenrightbigg2
+ 1 (A.58)
289
For an open rotor, the rotor thrust is equal to the total thrust and is constant. The
climb power is therefore given simply by:
Pc
Ph =
vi +Vc
vh
= Vc2v
h
+
radicalBiggparenleftbigg
Vc
2vh
parenrightbigg2
+ 1 (A.59)
Equations (A.54), (A.54), (A.58) and (A.59) have been plotted in Fig. A.7 as
functions of the climb velocity ratio, Vc/vh, for an open rotor and for different values
of the expansion ratio for a shrouded rotor. Although it might seem, from these
figures, that the shrouded rotors have a severe power penalty compared to the open
rotor, these representations are misleading, because the open and shrouded rotors
have different values of vh and Ph. These figures only serve to illustrate the forms
of the equations, but cannot be used for comparing the different configurations.
Figure A.8 therefore shows a true comparison of the power requirements of the
configurations when they are producing the same total thrust and have the same
disk area, using Eq. (A.28). The ordinate shows the ratio of the shrouded-rotor
and open-rotor power requirements, while the abscissa shows the climb velocity
normalized by the ideal induced velocity of the open rotor in hover.1 It can be
clearly seen that shrouding the rotor provides power savings up to a climb velocity
nearly equal to vhOR, and increasing the expansion ratio increases the velocity up
to which these savings are obtained. At much higher climb velocities, however,
shrouding results in a power penalty, and lower expansion ratios are preferable, as
1I.e., the figure shows (Pc/Ph)SR/(Pc/Ph)OR ? (Ph)SR/(Ph)OR plotted versus (Vc/vh)SR ?
(vh)SR/(vh)OR.
290
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
2
2.5
3
Vc / vh
v i / v
h
Open rotor (?d = 0.5)Shrouded rotor, ?
d = 1.0Shrouded rotor, ?d = 1.5
Shrouded rotor, ?d = 2.0
(a) Ideal induced velocity.
0 0.5 1 1.5 2 2.5 30
2
4
6
8
10
12
Vc / vh
P c / P
h
Open rotor (?d = 0.5)   Shrouded rotor, ?
d = 1.0Shrouded rotor, ?d = 1.5
Shrouded rotor, ?d = 2.0
(b) Ideal power.
Figure A.7: Predicted variations in ideal induced velocity and power in climb, for
open and shrouded rotors.
0 0.5 1 1.5 2 2.5 30.4
0.6
0.8
1
1.2
1.4
1.6
Vc / vh,OR
P SR / P
OR
Shrouded rotor, ?d = 1.0Shrouded rotor, ?
d = 1.5Shrouded rotor, ?d = 2.0
Figure A.8: Comparison of open and shrouded rotor power requirements in climb.
would be expected. In addition, since these analyses do not consider the additional
thrust that would be required to overcome the drag of the shroud, the true penalties
of shrouding would be even greater.
Finally, as in the previous sections, the equations derived here for shrouded
rotors represent a general case; by substituting the expression in Eq. (A.50) for the
expansion ratio, the equations reduce to the special case of an open rotor, and by
substituting Vc = 0, they reduce to the special case of the hover condition.
291
A.2 Combined Blade-Element-Momentum Theory (Hover)
The Momentum Theory model of a rotor, with its various simplifying as-
sumptions and idealized conditions, is useful for predicting the upper performance
limits of a rotor and for gaining a broad understanding of the relationships between
the top-level rotor parameters: thrust, power, the size of the rotor, and the size
of the wake. For more realistic analysis and design, however, theoretical models
with greater degrees of refinement and detail are necessary. The Combined Blade-
Element-Momentum-Theory (BEMT) model is one such: it allows for modeling of
the effects of the discrete number of rotor blades, of the aerodynamic characteristics
of the airfoil sections used for the blades, and of the geometry of the blades them-
selves, i.e., of radial variations in chord length (blade ?taper?), pitch angle (blade
?twist?) and airfoil section. The model is thus able to predict the effects of the
profile drag of the blades and of a non-uniform inflow distribution. One remaining
assumption is that the two-dimensional (?2-D?) aerodynamic characteristics of the
individual airfoil sections are not affected by those of the neighboring sections. In
effect, BEMT is essentially Lifting-Line theory applied to a rotating wing.
In the BEMT model, the azimuthally discrete nature of the rotor is accounted
for by a parameter called the solidity ratio (?), equal to the total planform area of
all the blades divided by the area of the rotor disk:
? = NbcRpiR2 = NbcpiR (A.60)
292
and by a parameter called the Prandtl Tip-Loss Factor (F). The spanwise variation
in the geometry of the blades is modeled by dividing the rotor disk into concentric
annuli, and equating the change in axial momentum flux of the air passing through
each annulus with the sum of the differential thrust forces acting on all the blade
elements within that annulus. Since the total change in momentum flux of the air
passing through the shrouded rotor, equal to ?mw, is due to both the rotor thrust
as well as the shroud thrust, the change in momentum flux due only to the rotor
thrust is obtained from Eq. (A.17):
Trotor = 12?
d
Ttotal = 12?
d
( ?mw) = ?m
parenleftbigg v
i
2?2d
parenrightbigg
(A.61)
where w = vi/?d is the ?v ? the additional flow velocity in the far wake compared
to the freestream ? due to the total thrust of the rotor-shroud system, while vi/2?2d
is the ?v due to the rotor thrust only. The thrust on a single annular section of the
(shrouded) rotor is therefore given by:
dTannulus(r) = d ?m(r)??v(r)?F(r)
=parenleftbig??dA(r)?vi(r)parenrightbig?
parenleftbiggv
i(r)
2?2d
parenrightbigg
?F(r) (A.62)
or, in non-dimensionalized coefficient form:
dCTannulus(r) = r?dr? ?
2(r)
?2d ?F(r) (A.63)
where ? = vi/R is the (induced) inflow ratio. The Prandtl tip-loss correction factor
293
for an open rotor, F(r), should be replaced with the equivalent Goodman correction
factor [141] or the Prandtl-Shaidakov factor used in the analytical model presented
by Bourtsev and Selemenev [92], both of which were discussed in Section 1.4.4.1 of
this dissertation.
In Ref. [175, p. 323, Eq. 6.49], a different formulation for the differential rotor
thrust coefficient is given:
dCTannulus(r) = r?dr? 2?
2(r)
?d (A.64)
This latter expression would seem to have been obtained by multiplying the ro-
tor thrust fraction, Trotor/Ttotal, by the expression for the differential (total) thrust
coefficient of an open rotor:
dCTannulus,OR = 1?Av2
tip
?bracketleftbigparenleftbig??2piydy?vi(y)parenrightbig?parenleftbig2vi(y)parenrightbigbracketrightbig
= 4?2(r)?r?dr (A.65)
where 2vi is the ?v corresponding to the (total) thrust on an open rotor. The
formulation in Eq. (A.64) therefore cannot be correct, and that of Eq. (A.63) must
be used.
From the blade-element model, the differential rotor thrust coefficient is ob-
tained from the vector summation of the lift and drag acting on an airfoil section:
dTannulus(r) = Nb ?parenleftbigdL(r)?cos(?(r))?dD(r)?sin(?(r))parenrightbig (A.66)
294
? dCTannulus(r) = 12?(r)?r?
radicalbig
?2(r) +r2 ?
parenleftbigg
cl(r)?cd(r)?(r)r
parenrightbigg
?dr (A.67)
Equating the expressions in Eqs. (A.63) and (A.67) yields a 4th-order equation in
?, which is easily solved with a computer. If the inflow is assumed to be small,
compared to the rotational (tangential) velocity of the blade element (vi lessmuch ?y, i.e.,
?2 lessmuchr2), then a quadratic in ? is obtained, with the solution given by:
? = ??
2
d cd
4F
bracketleftBigg
?1 +
radicalBigg
1 + 8Frcl??2
d c
2
d
bracketrightBigg
(A.68)
where all terms except the expansion ratio (?d) are functions of the radial coordinate,
r. These equations must be iteratively solved, as they involve cl and cd, which are
? in general ? non-linear functions of angle of attack, which in turn is dependent
on the inflow ratio. If the further assumptions are made that cl is a linear function
of ? (i.e., in the pre-stall range) and that cd is a constant:
cl = a?(???0) = a?(?????0) (A.69)
cd = cd0 (A.70)
then a closed-form, analytic expression for ? is obtained:
? = ??
2
d (cd0 +a)
4F
bracketleftBigg
?1 +
radicalBigg
1 + 8Fra(???0)??2
d (cd0 +a)2
bracketrightBigg
(A.71)
295
which reduces to the familiar expression for the inflow ratio of an open rotor:
? = ?a16F
bracketleftBigg
?1 +
radicalbigg
1 + 32Fr(???0)?a
bracketrightBigg
(A.72)
when ?d = 1/2 and cd is ignored (cd lessmuchcl).
296
Appendix B
Measures of Rotor Efficiency
B.1 Introduction
The designer of a rotor is ultimately interested in minimizing the amount of
power required to generate a specified amount of thrust, or, equivalently, maximizing
the amount of thrust that can be generated with a given amount of power. Efficiency
parameters are therefore key to determining the optimal operating state of a rotor
and to comparing the performances of different rotor designs. A fair amount of
confusion exists, however, regarding the best way to describe the efficiency of a
hovering rotor. Although the direct ratio of thrust to power ? also known as
the Power Loading1 (PL) ? gives an immediately appreciable idea of the power
penalty of a system, it is dimensional in nature, having dimensions of 1/velocity,
and therefore non-dimensional quantities are sought which would be independent
of the scale of the system. For a fluid-dynamic propulsive device operating at a
freestream velocity, v0, and producing a thrust, T, efficiency (?) is traditionally
expressed as the ratio of the ?useful? propulsive power, T ?v0 = ?m(vf ?v0) ?v0,
to the total input power, this being equal to the change in kinetic energy of the
fluid, 12 ?m(v2f ?v20), where vf is the final velocity of the fluid in the wake of the
propulsor. This ratio simplifies to the expression 2v0/(v0 + vf), which implies an
1In some of the literature, this term has also been used to instead refer to the ratio of the input
power to propeller disk area.
297
efficiency of zero in hover (v0 = 0). Since this result is obviously not physical, a
different formulation is required for the special case of zero translational velocity.
In the field of helicopter aerodynamics, the traditional wisdom has been to consider
the maximum figure of merit as the benchmark of a rotor?s efficiency; however,
arguments have also been made for considering the point of maximum CT/CP ?
the ratio of the thrust and power coefficients ? as that for optimum performance
[14, p. 53],[24]. In this appendix, it will be shown that both of these measures
deserve consideration, but that, depending on the constraints faced in the design
of a rotor, one or the other will assume greater importance. In the special case
of rotary-wing micro air vehicles (MAVs), which are usually operated with varying
rotor tip speeds, it will be shown that the figure of merit is preferrable to the
CT/CP ratio as criterion for determining the optimum operating state of an open
rotor and for comparing the performances of different rotors. However, in the case
of the shrouded-rotor configuration, the use of the figure of merit for comparing the
performances of configurations with different diffuser expansion ratios can result in
misleading conclusions. For this configuration, a generalized form of the figure of
merit is proposed, which is equally applicable to the specific case of an open rotor
as well.
B.2 Figure of Merit
The Figure of Merit (FM), also known as the static thrust efficiency, is the
ratio of the ideal ? i.e., minimum possible ? amount of power required to produce
298
a given thrust, to the actual amount required by the propulsor:
FM = ideal poweractual power = PiP
= ideal powerinduced power + profile power = Pi?P
i +Po (B.1)
wherePi, the ideal power, is a function of the thust and the disk loading (DL = T/A)
of the rotor:
Pi = T ?
radicalBigg
DL
2? =
T3/2?
2?A (B.2)
By normalizing the power quantities by the rotor disk area and tip speed, the figure
of merit can be expressed in terms of non-dimensional coefficients:
FM = CPiC
P
= CPi?C
Pi +CPo
(B.3)
= C
3/2
T /
?2
CP (B.4)
where, for a given rotor, all terms are now solely functions of the collective angle
(see Fig. B.3a on page 308, for example). Similar to the propulsive efficiency, this
formulation informs that (100?FM)% of the total input power is converted into
useful thrust. The figure of merit therefore quantifies the additional power required
? ( 1FM ?1) times the ideal power ? to overcome losses due to the non-ideal char-
acteristics of the flow: blade profile drag losses due to the viscosity of the fluid,
represented by the profile power term (Po, CPo), and losses due to non-uniform in-
flow distributions and to rotational momentum imparted to the wake, represented
by the induced power correction factor, ?.
299
The relation between the figure of merit and the power loading is given by:
PL = TP = TP
i
?FM
=
radicalbigg2?
DL ?FM (B.5)
where T/Pi = radicalbig2?/DL is the ideal, or maximum possible, power loading for a
given disk loading. The variation in power loading with disk loading, as given by
this equation, is plotted in Fig. B.1 for different values of FM. Thus, for a high value
of power loading, Eq. (B.5) indicates that the figure of merit should be high and
the disk loading should be low. The equation also shows that if two different rotors
have the same thrust and area ? in other words, the same disk loading ? then
their (total) power requirements are inversely proportional to their figures of merit.
Just as the propulsive efficiency, ?, is used to compare the performances of different
propellers at the same advance ratio (J = v?/nD), the figures of merit of different
hovering rotors are only meaningfully compared when the rotors are at the same
disk loading [14, p. 47]. The reason for this can be seen from the following example
arguments:
1. If two rotors are producing the same thrust but have different disk areas, then
the rotor with the smaller area (higher disk loading) will have a greater ideal-
power requirement. If, additionally, the two rotors happen to have the same
non-ideal loss characteristics (?, Po), then, by Equation B.1, the higher-disk-
loading rotor will have a higher figure of merit, thereby giving the misleading
conclusion that it is the more efficient rotor, even though it actually consumes
300
0 5 10 15 20 25 300
10
20
30
40
50
60
Disk loading, DL [lb/ft2]
Power loading, PL [lb/hp]
FM = 0.50FM = 0.75
FM = 1.00
Figure B.1: Variation of power loading with disk loading for different values of
figure of merit (sea-level standard atmosphere).
a greater total amount of power for the same thrust.
2. If a given rotor (fixed disk area) is operated at a higher tip speed but at the
same collective angle, and hence at the same thrust and power coefficients,
the thrust (proportional to v2tip) and disk loading will be higher, but the power
loading will be lower, since the power is proportional to v3tip, even though the
figure of merit is the same.
At the most fundamental level, though, this constraint on the usefulness of the figure
of merit as a comparative measure can be understood by recognizing, from Eq. (B.5)
and Fig. B.1, that the power loading is a function of FM only when the disk loading
is constant.
According to this constraint, therefore, the efficiency of rotors should be ex-
pressed by plotting the variation in their figures of merit as a function of disk
loading. However, for a single rotor, this would result in a family of curves with
different curves for different tip speeds, and with rotor collective as a parameter that
301
changes along any given curve.2 The curves would be identical, except that they
would be expanded further ?sideways? along the x-axis for increasing values of vtip.
By normalizing the abscissa values ? the disk loading ? by the square of the tip
speed (and by the air density), the variation in FM can be plotted versus the thrust
coefficient, CT, and expressed as a single characteristic curve for a given rotor. The
restriction on comparing figures of merit at the same disk loading implies that it is
permissible to compare them at different thrust coefficients, or at different values
of ideal power, since it possible for rotors to have different values of CT or Pi while
having the same disk loading. A question that naturally arises, then, is: if two
rotors have the same (maximum) value of FM, but at different values of CT, then
is one better than the other? The answer is ?no ? not necessarily.? For the same
desired thrust, and with the same restriction on disk area (therefore, the same disk
loading), the two rotors will have the same power requirement (Eq. (B.5)); the only
difference will be that the rotor with the higher CT will allow for a lower tip speed,
which, depending on other design considerations, may or may not be desirable.
B.3 CT/CP
Although the figure of merit is undoubtedly a useful concept, it is arguably a
level of abstraction away from the more direct and intuitively appealing thrust:power
ratio. A parameter that retains this appeal, while also having the desired attribute
2Thedatacouldalsobe plottedasa familyofcurveswithdifferentcurvesfordifferentcollectives,
and with tip speed as the parameter that changes along a curve. However, since FM is constant for
a given collective, this method would result in a collection of straight lines parallel to the x-axis,
and would be of little use.
302
of being non-dimensional, is the ratio of the thrust and power coefficients: CT/CP.
The relation of this quantity to the power loading is given by:
PL = TP = 1v
tip
? CTC
P
(B.6)
and, by using Equation B.4, to the figure of merit by:
CT
CP =
FMradicalbig
CT/2 (B.7)
Therefore, by this formulation, a high power loading is obtained by having a low tip
speed and a high value of CT/CP. Following the method3 shown by Leishman [14,
p. 53], differentiating the ratio:
P
T = vtip
CP
CT = vtip
parenleftBigg
?
radicalbigg
CT
2 +
CPo
CT
parenrightBigg
(B.8)
with respect toCT, while holding the tip speed constant, reveals the maximum value
of CT/CP (minimum CP/CT) to lie at a thrust coefficient equal to 2(CPo/?)2/3.
Experimental data illustrating these results can be seen in Fig. B.2, which shows
plots of power loading versus disk loading taken from Refs. [16] and [20]. The
figures show families of curves, with each curve for a different tip speed. Within
each curve, the maximum power loading occurs at the rotor collective angle for
maximum CT/CP. Plotting the same data in non-dimensional form, as CT/CP
3This analysis is predicated on the assumption ? which is reasonable when the blades are op-
erating at small angles of attack ? that CPo remains approximately constant with rotor collective.
303
(a) Data from Bohorquez and Pines [16] (b) Data from Hein and Chopra [20]
Figure B.2: Experimental data showing variation of power loading with disk
loading for MAV-scale rotors. (1 N/W = 168 lb/hp, 1 lb/ft2 = 47.9 N/m2)
versus CT, would cause all the curves in each figure to collapse to a single line that
would completely characterize that rotor for any tip speed.
B.4 Comparison of FM and CT/CP
In the section describing the figure of merit, it was stated that the requirements
for high power loading were a high figure of merit and a low disk loading. In the next
section, describing the CT/CP ratio, it was found that a low tip speed is necessary
for obtaining a high power loading. These requirements can be combined by solving
Eqs. (B.6) and (B.7) to obtain:
PL = TP = 1v
tip
? FMradicalbigC
T/2
(B.9)
However, these requirements are contradictory: disk loading is reduced by lowering
the thrust [coefficient], while the highest figures of merit are obtained at the higher
values of CT. In addition, for any given thrust, the tip speed is minimized by
304
increasing CT. The optimal operating state of a rotor must therefore lie at some
intermediate value of thrust coefficient ? or collective angle, since the various non-
dimensional coefficients and ratios are functions of rotor collective only ? which
maximizes the product of these terms. At the collective for maximum CT/CP,
inserting the corresponding value of CT into Eqs. (B.4) and (B.3) shows that the
figure of merit has a value of 2/3?. In comparison, depending on the value of CPo,
the maximum FM achievable by a rotor is closer to 1/?, and occurs at a higher
collective angle. (This can also be intuitively seen from the fact that the figure of
merit is proportional to CT3/2/CP.) The question therefore arises as to which is
the optimal operating state (collective) of a rotor: that for maximum CT/CP, or
that for maximum FM; and indeed, which of these two parameters is the better, or
?correct? measure of efficiency for comparing different rotors. It can be immediately
recognized from Eq. (B.6) that, for a given rotor, if the collective is changed from
that for maximum CT/CP, then, although CT/CP has decreased, the same power
loading can be achieved by reducing the tip speed. As a further example, it can
be seen from the curves shown in Fig. B.2 that, at a fixed thrust (disk loading),
the required power can be reduced (PL increased) by reducing the tip speed and
simultaneously increasing the rotor collective to compensate.
It is now clear that, with tip speed being allowed to vary, the collective for
maximum T/P at a given thrust is higher than that for maximum CT/CP. With
vtip now being expressed as radicalbigT/?ACT, the right-hand-side of Eq. (B.8) becomes
radicalbigT/2?A?(?+C
Po/CPi), or
radicalbigT/2?A?(1/FM), which is the same as Eq. (B.5). For a
given thrust and constrained rotor disk area, therefore, maximizing the power load-
305
ing (minimizing P/T) is equivalent to maximizing the figure of merit. Under these
circumstances, a rotor should be operated at the thrust coefficient corresponding to
maximum FM, and the tip speed adjusted to produce the desired thrust. Conversely,
if constraints exist on the choice of tip speed ? for example, due to considerations of
noise, autorotational capability, rotor dynamic stability, advancing-blade compress-
ibility or retreating-blade stall ? then, in this case, the rotor should be operated at
the thrust coefficient corresponding to maximumCT/CP, and the disk area adjusted
to produce the desired thrust.
The same conditions apply whether for comparing different operating states
(collectives) of the same rotor, or for comparing the performances of different rotors.
If the tip speed is constrained, the maximum values of CT/CP should be compared
in order to determine the better rotor for the application, while if disk area is con-
strained, the maximum values of FM should be compared.4 In rotary-wing MAVs,
for reasons of mechanical simplicity, rotor thrust is typically varied by adjusting
rotational speed rather than by changing collective; also, by definition, a restriction
exists on the maximum size of these vehicles. Therefore, for this class of vehicles, it
would be the figure of merit that is more important as an efficiency parameter.
In most situations, however, constraints exist on both tip speed as well as on
4In the latter case, the requirement for equal disk loading is automatically fulfilled, since both
thrust and area are the same for both rotors. An analogous requirement for comparing CT/CP
would be equality of tip speed. If two rotors have the same disk loading and power loading, and
hence the same figure of merit, but are operated at different thrust coefficients to produce the
desired thrust (i.e., different tip speeds), then Eq. (B.7) shows that they would have different
values of CT/CP; using CT/CP as the measure of efficiency in this case would give misleading
results. Similarly, if two rotors have the same value of CT/CP, but at different thrust coefficients,
then the rotor with the higher CT will produce the same thrust at a lower tip speed, and have a
higher power loading and lower power requirement (Eq. (B.6)).
306
the rotor disk area. In these cases, the optimal operating point of a rotor may lie
at a collective that is neither that for maximum CT/CP nor that for maximum FM,
but, instead, somewhere in between. If the constraint on tip speed is more severe
than that on disk area, then the the optimal operating point will be closer to that for
maximum CT/CP; if vice versa, then that for maximum FM. In general, however, in
such cases, the best method for choosing the operating point would be to simply plot
the required power for a desired thrust on a 3-axis plot as a function of both tip speed
and disk area, and directly examine the region within the constraint boundaries for
the minimum power requirement. As an example, this procedure has been carried
out using the experimental data for the MAV-scale open rotor tested in this study,
and is illustrated in Fig. B.3. Figure B.3a shows the experimental values of CT,
CP, CT/CP and FM as functions of rotor collective angle, showing that maximum
CT/CP is achieved at ?0 = 10?, while maximum FM is achieved at ?0 = 20?. By
interpolating in between these measured values, the 3-axis carpet plots shown in
Figs. B.3b?d were obtained ? these show, respectively, the rotor collective, thrust
coefficient and input power that are required in order to produce a desired thrust
of 40 g, depending on the choice of rotor diameter and tip speed. Superimposed on
these plots are contour lines of constant collective which indicate the locations of
operation at maximum CT/CP and maximum FM. Figure B.3e reproduces the plot
of required power from Fig. B.3d, but in the form of 2-dimensional contour plot.
It clearly illustrates how, at a fixed tip speed, operating at the point of maximum
CT/CP results in the lowest power requirement, while at fixed rotor diameter, it is
the point of maximum FM that results in the lowest power requirement.
307
0 10 20 30 400
0.005
0.01
0.015
0.02
0.025
C T
?0 0 10 20 30 400
0.004
0.008
0.012
C P
?0 0 10 20 30 400
1
2
3
4
C T/C
P
?0 0 10 20 30 400
0.05
0.1
0.15
0.2
0.25
FM
?0
(a) Rotor source data
5
6
7
8 60 80 100 120
140 160
0
10
20
30
40
vtip [ft/s]D [in]
Required 
? 0 [deg]
(b) Required collective
5
6
7
8 60 80 100 120 140 160
0
0.01
0.02
0.03
0.04
0.05
vtip [ft/s]
D [in]
Required C
T Max FM Max C
T/CP 
(c) Required thrust coefficient
5
6
7
8 60 80 100
120 140 160
3
4
5
6
7
8
9
vtip [ft/s]D [in]
Required P [W]
Max FM 
Max CT/CP 
(d) Required power (e) Required power
Figure B.3: Values of collective angle, thrust coefficient, and power required to
produce a desired thrust of 40 g, as functions of rotor diameter and tip speed
308
B.5 Shrouded rotors
The preceding discussions have dealt with the special case of an ?open? or ?free?
rotor, for which, in hover, the wake is allowed to naturally contract to a final cross-
sectional area equal to one-half that of the rotor disk. In the case of a shrouded
rotor, for which the development of the wake is dictated by the geometry of the
shroud, the ideal power requirement is now additionally a function of the expansion
ratio of the shroud?s diffuser, ?d. In this, more general case5, the ideal power is given
by:
Pi = Trotor ?vi = T2?
d
?
radicalBigg
?dT
?A =
T3/2?
4?d?A (B.10)
CPi = CTrotor ??i = T2?
d
?
radicalbig
?dCT = C
3/2
T
2??d (B.11)
where T is the total thrust of the system, equal to Trotor + Tshroud. Equation (B.6),
relating the power loading and theCT/CP ratio, still holds for the case of a shrouded
rotor, so this measure of efficiency retains its importance for cases where contraints
exist on the choice of rotor tip speed, just as for an open rotor. However, when it is
the disk area that is constrained, use of the traditional formulation of the figure of
merit (Eq. (B.1)) can lead to incorrect conclusions. From Eq. (B.10) it can be seen
that, at fixed total thrust and rotor disk area, if the expansion ratio is increased,
then the ideal power requirement decreases. Since the rotor is being off-loaded ?
i.e., the rotor itself is producing less thrust ? the blade section airfoils would be
5These equations derived here also hold for the special case of an open rotor, with the effective
value of 1/2 for the expansion ratio.
309
operating at lower angles of attack, and therefore it can be expected that the profile
power losses would decrease as well. If the proportional decrease in the profile power
is the same as that for the ideal power, then the figure of merit is unchanged. On
the other hand, if the assumption is made, as it is often done in first-order analyses,
that the profile power is independent of the blade angle, and therefore remains the
same, then the decrease in ideal power would cause FM to decrease as well. In
either case, however, it is obvious that the total power requirement has decreased,
and the power loading increased. Therefore, while the figure of merit remains an
appropriate measure of efficiency for comparing different operating states of the
same rotor-shroud combination, or of different rotors within the same shroud, it
cannot be the correct means for comparing configurations with different expansion
ratios. Direct comparison of the rotor polars, that is, plots of CP versus CT, would
accurately enable such a comparison; however, this representation has the limitation
of not being able to indicate the location on the characteristic curve for optimal
operation of the rotor (i.e., the point of maximum FM). Thus, some other measure
of efficiency is required, which incorporates all the benefits of each of these methods,
for the case of constrained disk area.
Equation (B.10) can be re-arranged to show that the relationPi = T?radicalbigDL/2?
also holds for a shrouded rotor (cf. Eq. (B.2)), where DL, the disk loading, is equal
to Trotor/A = T/2?dA. With FM still defined as Pi/P, the relationship between FM
and power loading remains that given by Eq. (B.5), repeated below for convenience:
T
P =
radicalbigg2?
DL ?FM (B.12)
310
However, for a given total thrust and constrained disk area, the disk loadings of
different rotors are now not the same if they have different [effective] values of ?d.
Consequently, the power requirements would not be inversely proportional to the
figures of merit. On the other hand, expanding Eq. (B.12) results in the formulation:
T
P =
radicalbigg
4?d?A
T ?FM (B.13)
=
radicalbigg
4?A
T ?FM
??
d (B.14)
which shows that, at fixed total thrust and disk area, power is inversely proportional
to the product FM???d, which can therefore be referred to as a ?generalized? figure
of merit. At fixed power and disk area, manipulation of the same equation shows
that the thrust that would be produced is directly proportional to (FM??d)2/3.
A graph of FM ???d versus CT would therefore perform the same role as that of
the traditional FM-vs.-CT graph for an open rotor, while also allowing for valid
comparisons of the performances of configurations with different expansion ratios.6
Using Eq. (B.6), the relation between the generalized FM and CT/CP is:
CT
CP =
2?
CT ?FM
??
d (B.15)
while its relation to the power loading is as given above in Eq. (B.14).
6Note that the product FM??d is equal to C3/2
T /2CP, so plotting a graph of this ?generalized?
figure of merit is nothing more than plotting the variation in C3/2T /CP, which is applicable to any
rotor configuration. Strikingly, this parameter has the same form as the ratios c3/2l /cd and C3/2L /CD
which are used to determine the angle of attack for minimum power consumption (maximum
endurance) of airfoils and fixed-wing aircraft, respectively.
311
Bibliography
[1] Hundley, R. O., and Gritton, E. C., ?Future Technology-Driven Revolutions
in Military Operations: Results of a Workshop,? RAND National Defense
Research Institute DB-110-ARPA, Santa Monica, CA, 1994 Defense Technical
Information Center (DTIC) Accession Number ADA285478.
[2] McMichael, J. M., and Francis, USAF (Ret.), C. M. S., ?Micro Air Ve-
hicles Toward a New Dimension in Flight,? U.S. Department of De-
fense Weapons Systems Technology Information Analysis Center (WSTIAC)
Newsletter, Vol. 1, No. 1?3, Jan?Jul 2000 Originally published online at
http://www.darpa.mil/tto/MAV/mav auvsi.html, Aug. 7, 1997.
[3] Wilson, III, S. B., ?Micro Air Vehicles (MAV),? Presentation at the AIAA
Student Chapter meeting at the University of Maryland, College Park, Oct
19, 2000 Presentation prepared Dec 14, 1999.
[4] Grasmeyer, J. M., and Keennon, M. T., ?Development of the Black Widow
Micro Air Vehicle,? 39th AIAA Aerospace Sciences Meeting and Exhibit,
Reno, NV, Jan 8?11, 2001, no. AIAA-2001-127.
[5] Tarascio, M. J., and Chopra, I., ?Design and Development of a Thrust Aug-
mented Entomopter: An Advanced Flapping Wing Micro Hovering Air Vehi-
cle,? American Helicopter Society 59th Annual Forum Proceedings, Phoenix,
AZ, May 6?8, 2003.
[6] Singh, B., Dynamics And Aeroelasticity Of Hover Capable Flapping Wings:
Experiments And Analysis PhD thesis, University of Maryland, College Park,
Department of Aerospace Engineering, 2006.
[7] Vogel, S., Life in Moving Fluids: The Physical Biology of Flow, 2nd ed.,
Princeton University Press, Princeton, NJ, 1994.
[8] Lowson, M. V., ?Aerodynamics of Aerofoils at Low Reynolds Numbers,?
14th International Unmanned Air Vehicle Systems Conference, Bristol, United
Kingdom, Apr 12?14, 1999, no. 35, University of Bristol, Cranfield College of
Aeronautics and Royal Aeronautical Society.
[9] Tennekes, H., The Simple Science of Flight: From Insects to Jumbo Jets The
MIT Press, Cambridge, MA, 1997.
[10] Azuma, A., The Biokinetics of Flying and Swimming Springer-Verlag, Tokyo,
Japan, 1992.
[11] Taylor, J. W. R., Jane?s All the World?s Aircraft - 2000/2001 Jane?s Publish-
ing Company, London, U.K., 2000.
312
[12] Bohorquez, F., Rotor Hover Performance And System Design Of An Efficient
Coaxial Rotary Wing Micro Air Vehicle PhD thesis, University of Maryland,
College Park, Department of Aerospace Engineering, 2007.
[13] Jacobs, E. N., and Sherman, A., ?Airfoil Section Characteristics as Affected
by Variations of the Reynolds Number,? NACA Report No. 586, Langley
Memorial Aeronautical Laboratory, Langley Field, VA, Jun. 24, 1936.
[14] Leishman, J. G., Principles of Helicopter Aerodynamics, 1st ed., Cambridge
Aerospace Series. Cambridge University Press, New York, NY, 2000.
[15] Prouty, R. W., Helicopter Performance, Stability, and Control Krieger Pub-
lishing Company, Malabar, FL, 1995.
[16] Bohorquez, F., and Pines, D., ?Rotor and Airfoil Design for Efficient Rotary
Wing Micro Air Vehicles,? American Helicopter Society 61st Annual Forum
Proceedings, Grapevine, TX, Jun 1?3, 2005.
[17] Samuel, P., Sirohi, J., Rudd, L., Pines, D., and Perel, R., ?Design and Analysis
of a Micro Coaxial Rotorcraft,? American Helicopter Society Vertical Lift
Aircraft Design Conference, San Francisco, CA, Jan 19?21, 2000.
[18] Young, L., Aiken, E., Johnson, J., Andrews, J., R., D., and J., K., ?New Con-
cepts and Perspectives on Micro-Rotorcraft and Small Autonomous Rotary-
Wing Vehicles,? 20th AIAA Applied Aerodynamics Conference, St. Louis,
MO, Jun 24?27, 2002.
[19] Bohorquez, F., ?Hover Performance of Rotor Blades at Low Reynolds Num-
bers for Rotary Wing Micro Air Vehicles: An Experimental Study,? American
Helicopter Society 59th Annual Forum Proceedings, Phoenix, AZ, May 6?8,
2003.
[20] Hein, B. R., and Chopra, I., ?Hover Performance of a Micro Air Vehicle: Ro-
tors at Low Reynolds Number,? Journal of the American Helicopter Society,
Vol. 52, No. 3, Jul 2007, pp. 254?262.
[23] Sirohi, J., Tishchenko, M., and Chopra, I., ?Design and Testing of a Micro-
Aerial Vehicle with a Single Rotor and Turning Vanes,? American Helicopter
Society 61st Annual Forum Proceedings, Grapevine, TX, Jun 1?3, 2005.
[25] Hrishikeshavan, V., and Chopra, I., ?Design and Testing of a Shrouded Rotor
Micro Air Vehicle with Anti-torque Vanes,? American Helicopter Society 64th
Annual Forum Proceedings, Montreal, Canada, Apr 29 ? May 1, 2008.
[22] Bohorquez, F., Samuel, P., Sirohi, J., Pines, D., Rudd, L., and Perel, R.,
?Design, Analysis and Hover Performance of a Rotary Wing Micro Air Vehi-
cle,? Journal of the American Helicopter Society, Vol. 48, No. 2, Apr 2003,
pp. 80?90.
313
[24] Ramasamy, M., Johnson, B., and Leishman, J. G., ?Toward Understanding
the Aerodynamic Efficiency of a Hovering Micro-Rotor,? AHS International
Specialists? Meeting on Unmanned Rotorcraft: Design, Control and Testing,
Chandler, AZ, Jan 23?25, 2007.
[21] Hein, B., ?Rotor Blade Optimization Studies for MAVs Operating at Low
Reynolds Number,? AIAA Mid-Atlantic Region I Student Conference, College
Park, MD, Apr. 11?12, 2003.
[26] Fletcher, H. S., ?Experimental Investigation of Lift, Drag and Pitching Mo-
ment of Five Annular Airfoils,? NACA TN 4117, Langley Memorial Aeronau-
tical Laboratory, Langley Field, VA, Oct 1957.
[27] Lipera, L., ?Micro Craft Ducted Air Vehicle,? AHS/AIAA/RaeS/SAE Inter-
national Powered Lift Conference, Arlington, VA, Oct 30?Nov 1 2000.
[28] Mouille, R., ?The ?Fenestron,? Shrouded Tail Rotor of the SA.341 Gazelle,?
Journal of the American Helicopter Society, Vol. 15, No. 4, Oct 1970, pp. 31?
37.
[29] Mouille, R., and d?Ambra, F., ?The ?Fenestron? A Shrouded Tail Rotor
Concept for Helicopters,? American Helicopter Society 42nd Annual Forum
Proceedings, Washington, D.C., Jun 2?5, 1986, pp. 597?606.
[30] Vuillet, A., ?Operational Advantages and Power Efficiency of the Fenestron as
Compared to a Conventional Tail Rotor,? Vertiflite, Vol. 35, No. 5, Jul/Aug
1989, pp. 24?29.
[31] Clemmons, M. G., ?Antitorque Safety and the RAH-66 Fantail,? American
Helicopter Society 48th Annual Forum Proceedings, Washington, D.C., Jun
3?5, 1992, pp. 169?175.
[33] Basset, P.-M., and Brocard, M., ?A Fenestron Model for Improving the Heli-
copter Yaw Dynamics Flight Simulation,? 30th European Rotorcraft Forum,
Marseilles, France, Sep 14?16, 2004.
[32] Niwa, Y., ?The Development of the New Observation Helicopter (XOH-1),?
AHS International 54th Annual Forum Proceedings, Washington, D.C., May
20?22, 1998, pp. 1285?1297.
[34] Fleming, J., Jones, T., Lusardi, J., Gelhausen, P., and Enns, D., ?Improved
Control of Ducted Fan UAVs in Crosswind Turbulence,? AHS 4th Decen-
nial Specialists? Conference on Aeromechanics, San Francisco, CA, Jan 21?23,
2004.
[35] Weir, R. J., ?Aerodynamic Design Considerations for a Free-Flying Ducted
Propeller,? AIAA Atmospheric Flight Mechanics Conference, Minneapolis,
MN, Aug 1988, no. AIAA-88-4377, pp. 420?431.
314
[36] Kuhn, R. E., ?Review of Basic Principles of V/STOL Aerodynamics,? NASA
Conference on V/STOL Aircraft, Langley Field, VA, Nov 17?18, 1960, Langley
Research Center, pp. 1?17.
[37] Hamel, G., ?Aeroplane,? US Patent No. 1,463,694, Jul 31, 1923 Application
filed Apr 4, 1922; Serial No. 549,576.
[38] Kort, L., ?Combined Device of a Ship?s Propeller Enclosed by a Nozzle,? US
Patent No. 2,030,375, Feb 11, 1936 Application filed Sep 7, 1933; In Germany
Jul 25, 1933; Serial No. 688,505.
[39] Stipa, L., ?Experiments with Intubed Propellers,? NACA TM 655, Jan 30,
1932 Translated from L?Aerotecnica (Rome), pp. 923-953, Aug 1931.
[40] Stipa, L., ?Stipa Monoplane with Venturi Fuselage,? NACA TM 753, Wash-
ington, DC, Sep 1934 Translated from Rivista Aeronautica, Vol. IX, No. 7,
pp. 13?37, Jul 1933.
[41] van Manen, J. D., and van Oossanen, P., Principles of Naval Architecture,
Second Revision, 3rd ed.,, Vol. II: Resistance, Propulsion and Vibration The
Society of Naval Architects and Marine Engineers (SNAME), Jersey City, NJ,
1988, Ch. VI, Section 9: Ducted Propellers, pp. 213?225.
[42] Thurston, S., and Amsler, R. C., ?Review of Marine Propellers and Ducted
Propeller Propulsive Devices,? Journal of Aircraft, Vol. 3, No. 3, May?Jun
1966, pp. 255?261.
[43] Kruger, W., ?On Wind Tunnel Tests and Computations Concerning the Prob-
lem of Shrouded Propellers,? NACA TM 1202, Feb 1949 Translation of
ZWB Forschungsbericht Nr. 1949, Aerodynamischen Versuchsanstalt (AVA),
G?ottingen, Germany, Jan 21, 1944.
[44] Kuchemann, D., ?Concerning the Flow About Ring-Shaped Cowlings of Fi-
nite Thickness, Part I,? NACA TM 1325, Jan 1952 Translation of ZWB
Forschungsbericht Nr. 1235, Jun 13, 1940.
[45] Kuchemann, D., and Weber, J., ?Concerning the Flow About Ring-Shaped
Cowlings, Part II: Annular Bodies of Infinite Length with Circulation for
Smooth Entrance,? NACA TM 1326, Nov 1951 Translation of ZWB
Forschungsbericht Nr. 1236/2, Nov 11, 1940.
[46] Kuchemann, D., and Weber, J., ?Concerning the Flow About Ring-Shaped
Cowlings Part VI: Further Measurements on Inlet Devices,? NACA TM 1327,
Dec 1951 Translation of ZWB Forschungsbericht Nr. 1236/6, Mar 30, 1942.
[47] Kuchemann, D., and Weber, J., ?Concerning the Flow About Ring-Shaped
Cowlings, Part VIII: Further Measurements on Annular Profiles,? NACA TM
1328, Feb 1952 Translation of ZWB Forschungsbericht Nr. 1236/8, Mar 25,
1943.
315
[48] Kuchemann, D., and Weber, J., ?Concerning he Flow About Ring-Shaped
Cowlings, Part IX: The Influence of Oblique Oncoming Flow on the Incre-
mental Velocities and Air Forces at the Front Part of Circular Cowls,? NACA
TM 1329, Feb 1952 Translation of ZWB Forschungsbericht Nr. 1236/9, Jun
10, 1943.
[49] Kuchemann, D., and Weber, J., ?Concerning the Flow About Ring-Shaped
Cowlings, Part XII: Two New Classes of Circular Cowls,? NACA TM 1360,
Oct 1953 Translation of ZWB Untersuchen und Mitteilungen Nr. 3111.
[50] Kuchemann, D., ?Concerning the Flow About Ring-Shaped Cowlings, Part
XIII: The Influence of a Projecting Hub,? NACA TM 1361, Oct 1953 Trans-
lation of ZWB Untersuchen und Mitteilungen Nr. 3144.
[51] Kuchemann, D., and Weber, J., Aerodynamics of Propulsion, 1st ed.,
McGraw-Hill Publications in Aeronautical Science. McGraw-Hill Book Com-
pany, Inc., New York, NY, 1953, Ch. 6: The ducted propeller, pp. 125?139.
[52] Platt, Jr., R. J., ?Static Tests of a Shrouded and an Unshrouded Pro-
peller,? NACA RM L7H25, Langley Memorial Aeronautical Laboratory, Lan-
gley Field, VA, Feb 9 1948.
[53] Fairchild, J. E., Batra, N. N., and Stewart, R. L., ?Influence of Design Param-
eters on Fan-in-Fin Static Performance,? American Helicopter Society 29th
Annual National Forum, May 1973.
[54] Sacks, A. H., and Burnell, J. A., ?Ducted Propellers ? A Critical Review of
the State of the Art,? In Progress in Aeronautical Sciences, A. Ferri, D. Kuche-
mann, and L. H. G. Sterne, Eds., Vol. 3. Pergammon Press, The Macmillan
Company, New York, NY, 1962, Ch. Part 2, pp. 85?135.
[55] Parlett, L. P., ?Aerodynamic Characteristics of a Small-Scale Shrouded Pro-
peller at Angles of Attack from 0? to 90?,? NACA TN 3547, Langley Memorial
Aeronautical Laboratory, Langley Field, VA, Nov 1955.
[56] Ham, N. D., and Moser, H. H., ?Preliminary Investigation of a Ducted Fan
in Lifting Flight,? Journal of the American Helicopter Society, Vol. 3, No. 3,
Jul 1958, pp. 17?29.
[57] Taylor, R. T., ?Experimental Investigation of the Effects of Some Shroud De-
sign Variables on the Static Thrust Characteristics of a Small-Scale Shrouded
Propeller Submerged in a Wing,? NACA TN 4126, Langley Memorial Aero-
nautical Laboratory, Langley Field, VA, Jan 1958.
[58] Moser, H. H., ?Some Effects of the Shroud on the Forward Flight Performance
of the Lifting Fan,? Journal of the American Helicopter Society, Vol. 4, No. 2,
Apr 1959, pp. 39?44.
316
[59] Nelson, N. E., ?The Ducted Fan in VTOL Aircraft Design,? Journal of the
American Helicopter Society, Vol. 4, No. 1, Jan 1959, pp. 53?56.
[60] Parlett, L. P., ?Stability and Control Characteristics of a Small-scale Model of
an Aerial Vehicle Supported by Two Ducted Fans,? NASA TN D-920, Langley
Research Center, Langley Field, VA, Jul 1961.
[61] Yaggy, P. F., and Goodson, K. W., ?Aerodynamics of a Tilting Ducted Fan
Configuration,? NASA TN D-785, Ames Research Center, Moffett Field, CA
and Langley Research Center, Langley, VA, Mar 1961.
[62] Yaggy, P. F., and Mort, K. W., ?A Wind-Tunnel Investigation of a 4-Foot-
Diameter Ducted Fan Mounted on the Tip of a Semispan Wing,? NASA TN
D-776, Ames Research Center, March 1961.
[63] Goodson, K. W., and Grunwald, K. J., ?Aerodynamic Characteristics of a
Powered Semispan Tilting-Shrouded-Propeller VTOL Model in Hovering and
Transition Flight,? NASA TN D-981, Langley Research Center, Jan 1, 1962.
[64] Grunwald, K. J., and Goodson, K. W., ?Aerodynamic Loads on an Isolated
Shrouded-Propeller Configuration for Angles of Attack from ?10? to 110?,?
NASA TN D-995, Langley Research Center, Jan 1, 1962.
[65] Grunwald, K. J., and Goodson, K. W., ?Division of Aerodynamic Loads on a
Semispan Tilting Ducted-Propeller Model in Hovering and Transition Flight,?
NASA TN D-1257, Langley Research Center, May 1, 1962.
[66] Mort, K. W., and Yaggy, P. F., ?Aerodynamic Characteristics of a 4-Foot-
Diameter Ducted Fan Mounted on the Tip of a Semispan Wing,? NASA TN
D-1301, Ames Research Center, Apr 1, 1962.
[67] Kelley, H. L., and Champine, R. A., ?Flight Operating Problems and Aero-
dynamic Performance Characteristics of a Fixed-Wing, Tilt-Duct, VTOL Re-
search Aircraft,? NASA TN D-1802, Langley Research Center, Jul 1, 1963.
[68] Mort, K. W., ?Performance Characteristics of a 4-Foot-Diameter Ducted Fan
at Zero Angle of Attack for Several Fan Blade Angles,? NASA D-3122, Ames
Research Center, Moffett Field, CA, Dec 1965.
[69] Newsom, Jr., W. A., ?Aerodynamic Characteristics of Four-Duct Tandem
VTOL-Aircraft Configurations,? NASA TN D-1481, Langley Research Center,
Langley Station, Hampton, VA, Jan 1963.
[70] Paxhia, V. B., and Sing, E. Y., ?The X-22A Design Development,? 1st AIAA
Annual Meeting, Washington, D.C., Jun 29 ? Jul 2, 1964, no. AIAA-64-282,
American Institute of Aeronautics and Astronautics.
317
[71] Maki, R. L., and Giulianetti, D. J., ?Aerodynamic Stability and Control
of Ducted-Propeller Aircraft,? Conference on V/STOL and STOL Aircraft,
Ames Research Center, Moffett Field, CA, Apr 1966, no. 9 in NASA-SP-116,
NASA, pp. 115?128.
[72] Mort, K. W., ?Summary of Large-Scale Tests of Ducted Fans,? Conference
on V/STOL and STOL Aircraft, Ames Research Center, Moffett Field, CA,
Apr 1966, no. 8 in NASA-SP-116, NASA, pp. 97?113.
[73] Spreemann, K. P., ?Wind-Tunnel Investigation of Longitudinal Aerodynamic
Characteristics of a Powered Four-Duct-Propeller VTOL Model in Transi-
tion,? NASA TN D-3192, Langley Research Center, Langley Station, Hamp-
ton, VA, Apr 1966.
[74] Giulianetti, D. J., Biggers, J. C., and Maki, R. P., ?Longitudinal Aerodynamic
Characteristics in Ground Effect of a Large-Scale, V/STOL Model with Four
Tilting Ducted Fans Arranged in a Dual Tandem Configuration,? NASA
NASA TN-D-4218, Ames Research Center, Moffett Field, CA, Oct 1967.
[75] Mort, K. W., and Gamse, B., ?A Wind-Tunnel Investigation of a 7-Foot-
Diameter Ducted Propeller,? NASA TN D-4142, Ames Research Center,
Moffett Field, CA, Aug 1967.
[76] Rhodes, W. B., ?Initial Military Flight Tests of the X-22A VSTOL Research
Aircraft,? Journal of Aircraft, Vol. 7, No. 2, Mar?Apr 1970, pp. 169?174.
[77] Thompson, J. F., and Roberts, S. C., ?Shrouded propeller research at Mis-
sissippi State University leading to application of the United States Army
XV-11A,? Proceedings of the CAL/USAAVLABS Symposium on Aerody-
namic Problems Associated with V/STOL Aircraft, Vol II: Propulsion and
Interference Aerodynamics, Buffalo, NY, Jun 1966, Vol. 2.
[78] Lazareff, M., ?Aerodynamics of Shrouded Propellers,? The Aerodynamics of
V/STOL Aircraft, AGARD-VKI Lecture Series, AGARDograph 126, Rhode-
Saint-Genese, Belgium, May 13?17, 1968, P. E. Colin and J. Williams, Eds.,
no. D, AGARD and Von Karman Institute, pp. 237?289.
[79] Mouille, R., ?Ten years of Aerospatiale experience with the Fenestron and
conventional tail rotor,? American Helicopter Society 35th Annual Forum,
May 1979, no. Preprint No. 79-58.
[80] Andrews, III, J. R., Riley, Jr., R. G., and Rahnke, C., ?Design and Testing
of a Ducted Tail Rotor Concept Demonstrator for a Model 222U Helicopter,?
Twenty-Second European Rotorcraft Forum, Brighton, U.K., Sep 16?19, 1996,
no. Paper No. 4.
[81] Riley, Jr., R. G., ?Effects of Uneven Blade Spacing on Ducted Tail Rotor
Acoustics,? American Helicopter Society 52nd Annual Forum Proceedings,
Washington, D.C., Jun 4?6, 1996, pp. 1?12.
318
[82] Lemont, H. E., ?Ringfin Augmentation Effects,? American Helicopter Society
38th Annual Forum Proceedings, Anaheim, CA, May 4?7, 1982, pp. 451?460.
[83] Harr, H. K., Hollifield, P. J., and Smith, R. P., ?Flight investigation of Bell
Model 206 ring fin,? American Helicopter Society 41st Annual Forum Pro-
ceedings, Fort Worth, TX, May 15?17, 1985, pp. 327?337.
[84] Clark, D. R., ?Aerodynamic Design Rationale for the Fan-in-Fin on the S-
67 Helicopter,? American Helicopter Society 31st Annual National Forum,
Washington, D.C., May 13?15, 1975, no. Preprint No. 904.
[87] Rajagopalan, R. G., and Keys, C. N., ?Detailed Aerodynamic Design of the
RAH-66 FANTAILTM Using CFD,? American Helicopter Society 49th Annual
Forum Proceedings, St. Louis, MO, May 19?21, 1993, pp. 1193?1206.
[85] Keys, C., She?er, M., Weiner, S., and Heminway, R., ?LH Wind Tunnel
Testing: KeytoAdvancedAerodynamicDesign,? AmericanHelicopterSociety
47th Annual Forum Proceedings, Phoenix, AZ, May 6?8, 1991, pp. 77?87.
[86] Wright, G. P., Driscoll, J. T., and Nickerson, Jr., J. D., ?Handling Qualities
of the H-76 FANTAIL TM Demonstrator,? American Helicopter Society 47th
Annual Forum Proceedings, 1991, pp. 123?135.
[88] Kothmann, B. D., and Ingle, S. J., ?RAH-66 Comanche Linear Aeroservoe-
lastic Stability Analysis: Model Improvements and Flight Test Correlation,?
AHS International 54th Annual Forum Proceedings, Washington, D.C., May
20?22, 1998, pp. 620?643.
[89] Alpman, E., Long, L. N., and Kothmann, B. D., ?Toward a Better Un-
derstanding of Ducted Rotor Antitorque and Directional Control in For-
ward Flight,? American Helicopter Society 59th Annual Forum Proceedings,
Phoenix, AZ, May 6?8, 2003.
[90] Alpman, E., Long, L. N., and Kothmann, B. D., ?Understanding Ducted-
Rotor Antitorque and Directional Control Characteristics Part I: Steady-State
Simulation,? Journal of Aircraft, Vol. 41, No. 5, Sep?Oct 2004, pp. 1042?1053.
[91] Alpman, E., Long, L. N., and Kothmann, B. D., ?Understanding Ducted-
Rotor Antitorque and Directional Control Characteristics Part II: Unsteady
Simulations,? Journal of Aircraft, Vol. 41, No. 6, Nov?Dec 2004, pp. 1370?
1378.
[92] Bourtsev, B. N., and Selemenev, Serguei, V., ?Fan-in-fin Performance at
Hover ComputationalMethod,? 26thEuropeanRotorcraftForum, The Hague,
Netherlands, Sep 26?29, 2000, no. Paper No. 10.
319
[93] Vuillet, A., and Morelli, F., ?New Aerodynamic Design of the Fenestron
for Improved Performance,? Twelfth European Rotorcraft Forum, Garmisch-
Partenkirchen, Federal Republic of Germany, Sep 22?25, 1986, no. Paper No.
8.
[94] Fleming, J., Jones, T., Gelhausen, P., and Enns, D., ?Improving Control
System Effectivenss for Ducted Fan VTOL UAVs Operating in Crosswinds,?
2nd AIAA ?Unmanned Unlimited? Systems, Technologies and Operations ?
Aerospace, San Diego, CA, Sep 15?18, 2003, no. AIAA-2003-6514.
[95] ?Airborne Remotely Operated Device (1982-1988),? World-Wide Web docu-
ment, http://www.spawar.navy.mil/robots/air/arod/arod.html, Dec 10, 1999.
[96] Murphy, D. W., and Bott, J. P., ?On the Lookout: The Air Mobile Ground Se-
curity and Surveillance System (AMGSSS) Has Arrived,? Unmanned Systems,
Vol. 13, No. 4, Fall 1995.
[97] Cycon, J. P., ?Decoding the Cypher UAV,? Vertiflite, Vol. 36, No. 6, Nov/Dec
1990, pp. 56?58.
[98] Cycon, J. P., ?Sikorsky Aircraft UAV Program,? Vertiflite, Vol. 38, No. 3,
May/Jun 1992, pp. 26?30.
[99] Thornberg, C. A., and Cycon, J. P., ?Sikorsky Aircraft?s Unmanned Aerial
Vehicle, Cypher: System Description and Program Accomplishments,? Amer-
ican Helicopter Society 51st Annual Forum Proceedings, Fort Worth, TX, May
9?11, 1995, pp. 804?811.
[100] Walsh, D., and Cycon, J. P., ?The Sikorsky Cypher(R) UAV: A Multi-Purpose
Platform with Demonstrated Mission Flexibility,? AHS International 54th
Annual Forum Proceedings, Washington, D.C., May 20?22, 1998, pp. 1410?
1418.
[101] Murphy, D. W., Bott, J. P., Cycon, J. P., Bryan, W. D., Coleman, J. L.,
Gage, D. W., Marsh, W. C., Martin, B. F., Nguyen, H. G., and Schneider,
R., ?Air-Mobile Ground Surveillance and Security System (AMGSSS) Project
Summary Report,? Naval Command, Control and Ocean Surveillance Center
(NCCOSC), RDT&E Division NRaD TD 2914, San Diego, CA, Sep 1996.
[102] ?Multipurpose Security and Surveillance Mission Plat-
form (1992-1998),? World-Wide Web document,
http://www.spawar.navy.mil/robots/air/amgsss/mssmp.html, May 21,
2002.
[103] Cycon, J. P., Rosen, K. M., and Whyte, A. C., ?Unmanned Flight Vehicle
Including Counter Rotating Rotors Positioned Within a Toroidal Shroud and
Operable to Provide All Required Vehicle Flight Controls,? US Patent No.
5,152,478, Oct 6, 1992 Filed May 18, 1990.
320
[104] Cycon, J. P., Millea, V. F., Kohlhepp, F. W., and Hansen, B. D., ?Drive Train
Assembly for a Rotor Assembly Having Ducted, Coaxial Counter-Rotating
Rotors,? US Patent No. 5,226,350, Jul 13, 1993 Filed Jun 22, 1992.
[105] Cycon, J. P., Kohlhepp, F. W., and Millea, V. F., ?Coaxial Transmis-
sion/Center Hub Subassembly for a Rotor Assembly Having Ducted, Coaxial
Counter-Rotating Rotors,? US Patent No. 5,351,913, Oct 4, 1994 Filed Jun
22, 1992.
[106] Cycon, J. P., Furnes, K. M., Kohlhepp, F. W., Darrell, M. D., and Sandy,
D. F., ?Toroidal Fuselage Structure for Unmanned Aerial Vehicles Having
Ducted, Coaxial, Counter-Rotating Rotors,? US Patent No. 5,277,380, Jan
11, 1994 Filed Jun 22, 1992.
[107] Cycon, J. P., DeWitt, C. W., and Scott, M. W., ?Unmanned Multi-Mode
Vertical Take Off and Landing Aircraft,? US Patent No. Des. 418,805, Jan
11, 2000 Filed Nov 24, 1998.
[108] Cycon, J. P., Scott, M. W., and DeWitt, C. W., ?Method of Reducing a Nose-
Up Pitching Moment on a Ducted Unmanned Aerial Vehicle,? US Patent No.
US 6,170,778 B1, Jan 9, 2001 Filed Apr 22, 1999.
[109] Cycon, J. P., Scott, M. W., and DeWitt, C. W., ?Unmanned Aerial Vehi-
cle with Counter-Rotating Ducted Rotors and Shrouded Pusher-Prop,? US
Patent No. US 6,270,038 B1, Aug 7, 2001 Filed Apr 22, 1999.
[110] McGonigle, K. P., Ferraro, J., and Cycon, J. P., ?Unmanned VTOL Ground
Surveillance Vehicle,? US Patent No. 5,575,438, Nov 19, 1996 Filed May 9,
1994.
[111] Ebbert, M. D., Gustin, R. G., Horbett, E. G., Edwards, J. J., and Adcock,
C. L., ?Unmanned Vertical Take-off and Landing, Horizontal Cruise, Air
Vehicle,? US Patent No. 5,295,643, Mar 22, 1994 Filed Dec 28, 1992.
[112] Lipera, L., Colbourne, J. D., Tischler, M. B., Mansur, M. H., Rotkowitz,
M. C., and Patangui, P., ?The Micro Craft iSTAR Micro Air Vehicle: Control
System Design and Testing,? American Helicopter Society 57th Annual Forum
Proceedings, Washington, D.C., May 9?11, 2001.
[113] Wilson, III, S. B., and Schnepf, S., ?Ground Loiter for OAVs and MAVs,?
American Helicopter Society 58th Annual Forum Proceedings, Montreal,
Canada, Jun 11?13, 2002.
[114] Black, D. M., Wainauski, H. S., and Rohrbach, C., ?Shrouded Propellers
? A Comprehensive Performance Study,? AIAA 5th Annual Meeting and
Technology Display, Philadelphia, PA, Oct 21?24, 1968, no. AIAA-68-994,
AIAA.
321
[115] Ahn, J., ?Parametric Design of a Ducted Fan System in Forward Flight and
Static Conditions,? AHS International Specialists? Meeting on Unmanned
Rotorcraft: Design, Control and Testing, Chandler, AZ, Jan 18?20, 2005.
[116] Kriebel, A. R., and Mendehall, M. R., ?Predicted and Measured Performance
of Two Full-Scale Ducted Propellers,? Proceedings of the CAL/USAAVLABS
Symposium on Aerodynamic Problems Associated with V/STOL Aircraft, Vol
II: Propulsion and Interference Aerodynamics, Buffalo, NY, Jun 1966, Vol. 2.
[117] Martin, P., and Tung, C., ?Performance and Flowfield Measurements on a 10-
inch Ducted Rotor VTOL UAV,? American Helicopter Society 60th Annual
Forum Proceedings, Baltimore, MD, Jun 7?10, 2004.
[118] Bell, E. B., ?Test of a Single-Stage Axial-Flow Fan,? NACA Report No. 729,
Langley Memorial Aeronautical Laboratory, Langley Field, VA, Sep 22 1941.
[119] Hubbard, H. H., ?Sound Measurements for Five Shrouded Propellers at Static
Conditions,? NACA TN 2024, Langley Memorial Aeronautical Laboratory,
Langley Field, VA, Apr 1950.
[120] Nygaard, T. A., Dimanlig, A. C., and Meadowcroft, E. T., ?Application of
a Momentum Source Model to the RAH-66 Comanche FANTAIL,? AHS 4th
Decennial Specialists? Conference on Aeromechanics, San Francisco, CA, Jan
21?23, 2004.
[121] Lee, H. D., Kwon, O. J., and Joo, J., ?Aerodynamic Performance Analysis of
a Helicopter Shrouded Tail Rotor Using an Unstructured Mesh Flow Solver,?
5th Asian Computational Fluid Dynamics, Busan, Korea, Oct 27?30, 2003,
no. Paper No. 8.
[122] Gupta, N., and Plump, J., ?Design of a VTOL Micro Air Vehicle that Incor-
porates Counter-Rotating, Reversible Thrust Propellers,? AUVSI Unmanned
Systems Conference, Baltimore, MD, Jul 31 ? Aug 2, 2001.
[123] Dyer, K. G., Aerodynamic Study of a Small, Ducted VTOL Aerial Vehicle PhD
thesis, Massachusetts Institute of Technology, Department of Aeronautics and
Astronautics, Cambridge, MA, Jun 2002.
[124] Martin, P. B., and Boxwell, D. A., ?Design, Analysis and Experiments on a
10-inch Ducted Rotor VTOL UAV,? AHS International Specialists? Meeting
on Unmanned Rotorcraft: Design, Control and Testing, Chandler, AZ, Jan
18?20, 2005.
[125] Graf, W., Fleming, J., and Gelhausen, P., ?Ducted Fan Aerodynamics in
Forward Flight,? AHS International Specialists? Meeting on Unmanned Ro-
torcraft: Design, Control and Testing, Chandler, AZ, Jan 18?20, 2005.
[126] Kondor, S., Amitay, M., Parekh, D., Fung, P., and Glezer, A., ?Active Flow
Control Application on a Mini Ducted Fan Uav,? No. AIAA-2001-2440, AIAA.
322
[127] Kondor, S., and Heiges, M., ?Active Flow Control for Control of Ducted
Rotor Systems,? Proceedings of the 39th AIAA Aerospace Sciences Meeting
& Exhibit, Reno, NV, Jan 8?11, 2001, no. AIAA-2001-0117.
[128] Fung, P. H., and Amitay, M., ?Control of a Miniducted-Fan Unmanned Aerial
Vehicle Using Active Flow Control,? Journal of Aircraft, Vol. 39, No. 4, Jul?
Aug 2002, pp. 561?571.
[129] Plump, J. M., and Gupta, N. R., ?Vertical Takeoff and Landing Aerial Ve-
hicle,? US Patent No. 6,691,949 B2, Feb 17, 2004 Application filed Jul 8,
2002.
[130] Moller, P. S., ?Robotic or Remotely Controlled Flying Platform,? US Patent
No. 4,795,111, Jan 3, 1989 Application filed Feb 17, 1987.
[131] Moller, P. S., ?Stabilizing Control Apparatus for Robotic or Remotely Con-
trolled Flying Platform,? US Patent No. US 6,450,445 B1, Sep 17, 2002 Filed
Dec 10, 1999.
[132] Kelly, Mark, W., ?Large-Scale Wind-Tunnel Studies of Several Vtol Types,?
NASA Conference on V/STOL Aircraft, Langley Field, VA, Nov 17?18, 1960,
Langley Research Center, pp. 35?48.
[133] Milla, J. H., and Blick, E. F., ?Lift and Drag Characteristics of a Biannular
Wing Airplane,? Journal of Aircraft, Vol. 3, No. 6, Nov?Dec 1966, pp. 586?
587.
[134] Abrego, A. I., and Bulaga, R. W., ?Performance Study of a Ducted Fan
System,? American Helicopter Society Aerodynamics, Acoustics, and Test
and Evaluation Technical Specialists? Meeting, San Francisco, CA, Jan 23?25,
2002.
[135] Kriebel, A. R., ?Theoretical Stabiity Derivatives for a Ducted Propeller,?
Journal of Aircraft, Vol. 1, No. 4, Jul?Aug 1964, pp. 203?210.
[136] Mendenhall, M. R., and Spangler, S. B., ?Theoretical Study of Ducted Fan
Performance,? NASA CR 1494, Jan 1, 1970.
[137] Mendenhall, M. R., and Spangler, S. B., ?A Computer Program for the
Prediction of Ducted Fan Performance,? NASA CR 1495, Feb 1, 1970.
[138] Sheehy, T. W., ?Computer Aided Shrouded Propeller Design,? Proceedings of
the AIAA 9th Annual Meeting and Technical Display, 1973, no. AIAA-73-54,
AIAA.
[139] Gray, R. B., and Wright, T., ?A Vortex Wake Model for Optimum Heavily
Loaded Ducted Fans,? Journal of Aircraft, Vol. 7, No. 2, Mar?Apr 1970,
pp. 152?158.
323
[140] McCormick, Jr., B. W., Aerodynamics of V/STOL Flight Aerodynamic Press,
New York, NY, 1976.
[141] Goodman, T. R., ?The Tip Correction for Wind-Tunnel Tests of Propellers,?
Journal of the Aeronautical Sciences, Vol. 23, No. 12, Dec 1956, pp. 1094?1098.
[142] Gibson, I. S., ?Theoretical Studies of Tip Clearance and Radial Variation of
Blade Loading on the Operation of Ducted Fans and Propellers,? Journal of
Mechanical Engineering Science, Vol. 16, No. 6, 1974, pp. 367?376.
[143] de Campos, J. F., ?A Three-Dimensional Theory for the Design Problem of
Propeller Ducts in a Shear Flow,? Eighteenth Symposium on Naval Hydro-
dynamics, 1991, National Academy of Sciences, pp. 645?665.
[144] Page, G. S., ?Design and Analysis of Single and Dual Rotation Ducted Fans,?
Proceedings of the AIAA 34th Aerospace Sciences Meeting and Exhibit, Reno,
NV, Jan 15?18, 1996, no. AIAA-96-0677.
[145] Guerrero, I., Londenberg, K., Gelhausen, P., and Myklebust, A., ?A Powered
Lift Aerodynamic Analysis for the Design of Ducted Fan UAVs,? 2nd AIAA
?Unmanned Unlimited? Systems, Technologies and Operations ? Aerospace,
San Diego, CA, Sep 15?18, 2003, no. AIAA-2003-6567.
[146] Quackenbush, T. R., Wachspress, D. A., Boschitsch, A. H., and Solomon,
C. L., ?Aeromechanical Analysis Tools for Design and Simulation of VTOL
UAV Systems,? American Helicopter Society 60th Annual Forum Proceedings,
Baltimore, MD, Jun 7?10, 2004.
[147] Rajagopalan, R. G., and Zhaoxing, Z., ?Performance and Flow Field of a
Ducted Propeller,? AIAA/ASME/SAE/ASEE 25th Joint Propulsion Confer-
ence, Monterey, CA, Jul 10?12, 1989, no. AIAA-89-2673.
[148] Alpman, E., and Long, L. N., ?Unsteady RAH-66 Comanche Flowfield Sim-
ulations Including Fan-in-fin,? 16th AIAA Computational Fluid Dynamics
Conference, Orlando, FL, Jun 23?26, 2003, no. AIAA-2003-4231.
[149] Ruzicka, G. C., Strawn, R. C., and Meadowcroft, E. T., ?Discrete Blade CFD
Analysis of Ducted Tail Fan Flow,? 42nd AIAA Aerospace Sciences Meeting
and Exhibit, Reno, NV, Jan 5?8, 2004, no. AIAA-2004-48.
[150] Ruzicka, G. C., Strawn, R. C., and Meadowcroft, E. T., ?Discrete-Blade,
Navier-Stokes Computational Fluid Dynamics Analysis of Ducted-Fan Flow,?
Journal of Aircraft, Vol. 42, No. 5, Sep?Oct 2005, pp. 1109?1117.
[151] Ahn, J., and Lee, K., ?Performance Prediction and Design of a Ducted Fan
System,? 40th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and
Exhibit, Fort Lauderdale, FL, Jul 11?14, 2004, no. AIAA-2004-4196.
324
[152] Drela, M., andGiles, M.B., ?ISES-ATwo-DimensionalViscousAerodynamic
Design and Analysis Code,? 25th AIAA Aerospace Sciences Meeting, Jan 12?
15, 1987, no. AIAA 1987-424.
[153] Srivastava, R., Reddy, T. S. R., and Stefko, G. L., ?A Numerical Aeroelastic
Stability Analysis of a Ducted-Fan Configuration,? Proceedings of the 32nd
AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, Buena
Vista, FL, Jul 1?3, 1996, no. AIAA-96-2671.
[154] Keith, Jr., T. G., and Srivastava, R., ?Final Technical Report for
Euler/Navier-Stokes Solvers Applied to Ducted Fan Configurations,? Uni-
versity of Toledo NASA CR 97-112953, Toledo, OH, Dec 1997.
[155] Srivastava, R., and Reddy, T. S. R., ?Application of Three Flutter Analysis
Methods to a Ducted-Fan Configuration,? AIAA.
[156] Elena, L., and Schiestel, R., ?Turbulence Modeling of Rotating Confined
Flows,? International Journal of Heat and Fluid Flow, Vol. 17, No. 3, Jun
1996, pp. 283?289.
[157] Randriamampianina, A., Elena, L., Fontaine, J. P., and Schiestel, R., ?Nu-
merical Prediction of Laminar, Transitional and Turbulent Flows in Shrouded
Rotor-Stator Systems,? Physics of Fluids, Vol. 9, No. 6, Jun 1997, pp. 1696?
1713.
[158] Drischler, J. A., ?Analytic Studies of Sound Pressures Inside the Duct of
Ducted Propellers,? NASA TN D-6345, Langley Research Center, Sep 1,
1971.
[159] Fricke, W., Bissell, J. R., Bamberg, W. T., and Martina, C. K., ?Analyt-
ical and Experimental Studies of Sound Pressures on Ducted Propellers,?
Bell Aerosystems Company NASA-CR-66270; Report No. 9500-950001, Dec
1, 1966 NASA Contract NASW-1217.
[160] Gearhart, W. S., ?Tip Clearance Cavitation in Shrouded Underwater Propul-
sors,? Journal of Aircraft, Vol. 3, No. 2, Mar?Apr 1966, pp. 185?192.
[161] Oweis, G. F., and Ceccio, S. L., ?Instantaneous and Time-Averaged Flow-
Fields of Multiple Vortices in the Tip Region of a Ducted Propulsor,? Exper-
iments in Fluids, Vol. 38, Mar 24, 2005, pp. 615?636.
[162] Oweis, G. F., Fry, D., Chesnakas, C. J., Jessup, S. D., and Ceccio, S. L.,
?Development of a Tip-Leakage Flow ? Part 1: The Flow Over a Range of
Reynolds Numbers,? Journal of Fluids Engineering, Vol. 128, No. 4, Jul 2006,
pp. 751?764.
[163] Oweis, G. F., Fry, D., Chesnakas, C. J., Jessup, S. D., and Ceccio, S. L.,
?Development of a Tip-Leakage Flow ? Part 2: Comparison Between the
325
Ducted and Un-ducted Rotor,? Journal of Fluids Engineering, Vol. 128, No. 4,
Jul 2006, pp. 765?773.
[164] Avanzini, G., D?Angelo, S., and de Matteis, G., ?Performance and Stability
of a Ducted-Fan Uninhabited Aerial Vehicle,? Proceedings of the 39th AIAA
Aerospace Sciences Meeting & Exhibit, Reno, NV, Jan 8?11, 2001, no. AIAA-
2001-0844.
[165] Avanzini, G., de Matteis, G., and Fresta, F., ?Robust Multivariable Control
of a Shrouded-Fan Uninhabited Aerial Vehicle,? AIAA Atmospheric Flight
Mechanics Conference and Exhibit, Monterey, CA, Aug 5?8, 2002.
[166] Avanzini, G., D?Angelo, S., and de Matteis, G., ?Performance and Stability of
Ducted-Fan Uninhabited Aerial Vehicle Model,? Journal of Aircraft, Vol. 40,
No. 1, Jan-Feb 2003, pp. 86?93.
[167] Salluce, D. N., ?Comprehensive System Identification of Ducted Fan UAVs,?
Master?s thesis, California Polytechnic State University, San Luis Obispo, CA,
Jan 2004.
[168] Bohorquez, F., and Pines, D., ?Hover Performance and Swashplate Design of
a Coaxial Rotary Wing Micro Air Vehicle,? American Helicopter Society 60th
Annual Forum Proceedings, Baltimore, MD, Jun 7?10, 2004.
[169] Bohorquez, F., and Pines, D., ?Rotor Design for Efficient Rotary Wing Micro
Air Vehicles,? AHS International Specialists Meeting on Unmanned Rotor-
craft: Design, Control and Testing,, Chandler, AZ, Jan 18?20, 2005.
[170] Hein, B., and Chopra, I., ?Hover Performance of a Micro Air Vehicle: Rotors
at Low Reynolds Number,? AHS International Specialists? Meeting on Un-
manned Rotorcraft: Design, Control and Testing, Chandler, AZ, Jan 18?20,
2005.
[171] Ramasamy, M., Leishman, J. G., and Lee, T. E., ?Flow Field of a Rotating-
Wing Micro Air Vehicle,? American Helicopter Society 62nd Annual Forum
Proceedings, Phoenix, AZ, May 9?11, 2006.
[172] Schroeder, E. J., ?Low Reynolds Number Flow Validation Using Computa-
tional Fluid Dynamics With Application To Micro Air Vehicles,? Master?s
thesis, University of Maryland, College Park, College Park, MD, 2005.
[173] Schroeder, E. J., and Baeder, J. D., ?Using Computational Fluid Dynamics
for Micro-Air Vehicle Airfoil Validation and Prediction,? 23rd AIAA Applied
Aerodynamics Conference, Toronto, Canada, Jun 6?9, 2005, no. AIAA-2005-
4841.
[174] Boatwright ?Measurements of Velocity Components in the Wake of a Full-
Scale Helicopter Rotor in Hover,? USAAMRDL TR 72-33, 1972 In Prouty,
326
R. W., Helicopter Performance, Stability, and Control , Krieger Publishing
Company, Malabar, FL, 1995, p.40.
[175] Leishman, J. G., Principles of Helicopter Aerodynamics, 2nd ed., Cambridge
Aerospace Series. Cambridge University Press, New York, NY, 2006.
[176] Armutcuoglu, O., Kavsaoglu, M. S., and Tekinalp, O., ?Tilt Duct Verti-
cal Takeoff and Landing Uninhabited Aerial Vehicle Concept Design Study,?
Journal of Aircraft, Vol. 41, No. 2, Mar?Apr 2004, pp. 215?223.
[177] Behun, M., Rom, F. E., and Hensley, R. V., ?Evaluation of a Ducted-Fan
Power Plant Designed for High Output and Good Cruise Fuel Economy,?
NACA RM E50E01, Lewis Flight Propulsion Laboratory, Cleveland, OH, Oct
17, 1950.
[178] Brocard, J. M. R., ?Aircraft,? US Patent No. 3,128,062, Apr 7 1964 Filed
Feb 19, 1962; In France, Jul 19, 1961; Serial No. 174,314.
[179] Campbell, J. P., ?Development of V/STOL Aircraft 1950 to 1970,? Journal of
the American Helicopter Society, Vol. 24, No. Special V/STOL Aircraft Issue,
Jun 1979, pp. 7?16.
[180] Corson, Jr., B. W., ?The Aerodynamics of a Wind-Tunnel Fan,? NACA
Report No. 820, Langley Memorial Aeronautical Laboratory, Langley Field,
VA, Aug. 1941.
[181] Drela, M., and Youngren, H., ?Axisymmetric Analysis and Design of Ducted
Rotors,? Jan 30, 2005.
[182] Gilmore, A. W., and Grahame, W. E., ?Research Studies on a Ducted Fan
Equipped with Turning Vanes,? Journal of the American Helicopter Society,
Vol. 4, No. 4, Oct 1959, pp. 11?21.
[183] Gruzling, J., ?Propeller Shroud,? US Patent No. 4,789,302, Dec 6, 1988
Application filed Feb 6, 1987.
[184] Kort, L., ?Combined Propelling and Steering Device for Screw Propelled
Ships,? US Patent No. 2,139,594, Dec 6, 1938 Application filed Jun 25, 1937;
In Germany Feb 8, 1936; Serial No. 150,398.
[185] Lindenbaum, B., and Blake, W., ?The VZ-9 ?Avrocar?,? Vertiflite, Vol. 44,
No. 2, Spring 1998.
[186] Milde, Jr., K. F., Pruszenski, Jr., A. S., and Wagner, J., ?A Safe and Fuel-
Efficient VTOL ?Personal Aircraft?,? American Helicopter Society 59th An-
nual Forum Proceedings, Phoenix, AZ, May 6?8, 2003.
327
[187] Okan, A., Tekinalp, O., and Kavsaoglu, M. S., ?Flight Control of a Tilt-
Duct Vtol Uav,? Proceedings of the AIAA 1st Technical Conference and
Workshop on Unmanned Aerospace Vehicles, Portsmouth, VA, May 20?23,
2002, no. AIAA-2002-3466.
[188] Perl, W., and Tucker, M., ?A General Representation for Axial-Flow Fans
and Turbines,? NACA Report No. 814, Langley Memorial Aeronautical Lab-
oratory, Langley Field, VA, May 4 1945.
[189] de Piolenc, F. M., and Wright, Jr., G. E., ?Ducted Fan Design, Volume 1,?
522 E. Vine, 213, West Covina, CA 91790, USA., 2001.
[190] Sato, T., and Engeda, A., ?Design and Performance Analysis of Ducted Fans
for Micro UAV Applications,? AHS International Specialists? Meeting on Un-
manned Rotorcraft: Design, Control and Testing, Chandler, AZ, Jan 18?20,
2005.
[191] Van Weelden, S. D., Smith, D. E., and Mullins, Jr., B. R., ?Preliminary
Design of a Ducted Fan Propulsion System for General Aviation Aircraft,?
Proceedings of the AIAA 34th Aerospace Sciences Meeting and Exhibit, Reno,
NV, Jan 15?18, 1996, no. AIAA-96-0376.
[192] Wallis, A. M., Denton, J. D., and Demargne, A. A. J., ?The Control of
Shroud Leakage Flows to Reduce Aerodynamic Losses in a Low Aspect Ratio,
ShroudedAxialFlowTurbine,? Transactions of the ASME,Vol.123, Apr2001,
pp. 334?341 Presented at the 45th International Gas Turbine and Aeroengine
Congress and Exhibition, Munich, Germany, May 811, 2000. International Gas
Turbine Institute Paper No. 2000-GT-475.
[193] Xu, G.-H., and Wang, S.-C., ?Effects of the Shroud on Aerodynamic Perfor-
mance in Helicopter Shrouded Tail Rotor System,? Aircraft Engineering and
Aerospace Technology, Vol. 73, No. 6, 2001, pp. 568?572.
328