ABSTRACT 
 
Title: DISCHARGE CHARACTERISTICS OF 
CANONICAL SPRINKLER SPRAYS 
 
Andrew F. Blum, Master of Science, 2007 
 
Directed By: Andre Marshall, Assistant Professor, Department of 
Fire Protection Engineering 
 
Detailed characterization of spray behavior and its relationship to nozzle geometry, 
fluid properties, and injection characteristics is needed to advance water-based 
suppression technology and fire related computational fluid dynamics (CFD) tools.  In 
this study, a series of experiments have been conducted to measure discharge 
characteristics of sprays produced by basic injector configurations modeled after 
conventional pendant sprinklers.  Liquid jets of various sizes were injected downwards 
onto flat deflectors, tined deflectors, and boss-modified tined deflectors to establish the 
three canonical configurations explored in this study.  Spray measurements including the 
initial angle of the sheet at the deflector exit, the sheet breakup radius, the drop size 
distribution 1 m below the deflector surface, and the volume density distribution were 
performed for these configurations.  These systematic experiments provide discharge 
characteristics of practical interest while providing valuable data for CFD based 
atomization model development.   
DISCHARGE CHARACTERISTICS OF CANONICAL SPRINKLER SPRAYS 
 
By 
 
Andrew F. Blum 
 
Thesis 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 
Master of Science 
2007 
 
Advisory Committee: 
Professor Dr. Andre Marshall, Chair 
Professor Dr. Arnaud Trouve 
Professor Dr. James Milke 
 
? Copyright by 
Andrew Blum 
2007 
 
ii 
 
Acknowledgements 
This work is supported by the National Fire Sprinkler Association (NFSA).  I want to 
thank the NFSA program manager Mr. Russell Fleming for his support of the project.  I 
want to express my appreciation to Dr. Andre Marshall, my advisor, for his helpful 
guidance, support and the tireless hours he spent with me in the lab working on my 
research.  
 
I would also like to thank the rest of my thesis committee, Dr. Arnaud Trouve and Dr. 
James Milke, for their guidance through this process. In addition, I would like to thank 
the Maryland Fire Rescue Institute for their support and facilities for my research, most 
notably Marty LePore and Scott Wood.  
 
I also want to thank the rest of the Fire Protection Engineering staff at the University 
of Maryland for passing along their many years of knowledge to me and, as well as my 
fellow students that worked with me throughout my years, most notably Ning Ren, for his 
help with my research.  
 
Finally, I would like to thank my girlfriend, Leigh Stewart, for her continuous support 
throughout the completion of my research. I could not have done it without your love and 
encouragement. 
 
iii 
 
Table of Contents 
 
Acknowledgements............................................................................................................. ii 
Table of Contents............................................................................................................... iii 
List of Figures......................................................................................................................v 
Nomenclature..................................................................................................................... vi 
Chapter 1: Introduction........................................................................................................1 
1.1 Overview....................................................................................................................1 
1.2 Literature Review.......................................................................................................2 
1.2.1 Sheet Disintegration Studies...............................................................................2 
1.2.2 Drop Size Measurements ....................................................................................4 
1.3 Research Objectives...................................................................................................6 
Chapter 2: Approach............................................................................................................8 
2.1 Atomization Physics ..................................................................................................8 
2.2 Atomization Measurements .......................................................................................9 
2.3 Diagnostics...............................................................................................................11 
2.3.1 Sheet Trajectory................................................................................................11 
2.3.2 Sheet Breakup ...................................................................................................13 
2.3.3 Volume Density ................................................................................................15 
2.3.4 Drop Size ..........................................................................................................17 
Chapter 3: Results and Discussion.....................................................................................21 
3.1 Sheet Trajectory.......................................................................................................21 
3.2 Sheet Breakup ..........................................................................................................23 
3.3 Volume Density .......................................................................................................25 
iv 
 
3.4 Drop Size Distributions............................................................................................26 
Chapter 4: Conclusions......................................................................................................33 
 
v
List of Figures 
Figure 1. Previous drop size measurements.........................................................................6 
Figure 2. Theoretical sheet breakup processes ....................................................................9 
Figure 3. The anatomy of a sprinkler.................................................................................10 
Figure 4. Sheet trajectory and sheet breakup diagnostic set-up.........................................13 
Figure 5. An inverted PLIF photo of the Standard nozzle at 2.07 bar...............................14 
Figure 6. Sheet breakup photo from above the spray of the Basis nozzle.........................15 
Figure 7. MFRI facility plan view and diagnostics............................................................16 
Figure 8. Malvern diffraction technique ............................................................................18 
Figure 9. Malvern drop size limitations.............................................................................20 
Figure 10. Initial angle results ...........................................................................................21 
Figure 11. Inverted PLIF images at radial locations within the spray...............................23 
Figure 12. Dimensionless sheet breakup distances............................................................24 
Figure 13. Dimensionless linear density results ................................................................27 
Figure 14. Local d
v50
 measurements ..................................................................................28 
Figure 15. Rosin-Rammler/log-normal distributions.........................................................29 
Figure 16. Dimensionless drop size results........................................................................32 
Figure 17. Dimensionless drop size results for spaces and tines .......................................32 
 
vi 
 
Nomenclature 
 
CVF   Cumulative volume fraction  
d Drop diameter, ?m 
10
d Average mean drop diameter, ?m 
30
d Average mean volume drop diameter, ?m 
50v
d Characteristic drop diameter, ?m 
D
boss
 Boss diameter, mm 
D
def
 Deflector diameter, mm 
D
inlet
 Nozzle inlet diameter, mm 
o
D Orifice diameter, mm 
g Gravitational acceleration constant, m/s
2
h Measurement elevation, m 
L
inlet
 Length of nozzle inlet, mm 
L
jet
 Length of jet before deflector impact, mm 
q
 Linear volume density 
q Measured volume flux, kg/sm
2
Q Volumetric flow rate, lpm 
QF  Volume flux fraction 
r Radius, m 
r
 Non-dimensional radial coordinate (r/R)
r
bu
, Sheet breakup distance, mm 
R Characteristic dispersion length scale, m 
vii 
 
Re  Reynolds number 
U
jet
 Jet velocity, m/s 
(v
o
)
r
Initial radial sheet velocity, m/s 
(v
o
)
z
Initial vertical sheet velocity, m/s 
VF  Spatial volume fraction 
w Rosin-Rammler/log-normal correlation coefficient 
We  Weber number 
x Coordinate 
y Coordinate 
z Coordinate 
Greek letters 
 Surface tension, N/m 

boss  
Angle of deflector boss, ? 

space
 Angle of deflector space, ? 

tine
 Angle of deflector tine, ? 
 Density, kg/m
3
Subscripts 
i Drop size bin 
j Measurement station 
l Water liquid 
T Total volume 
1
Chapter 1: Introduction 
1.1 Overview
Sprinkler systems have been used to suppress fires for over a hundred years.  The 
extinguishing ability of water and the physical mechanisms of water-based suppression 
are well understood.  Those mechanisms include, heat extraction from the fire via drop 
vaporization and expansion, attenuation of heat feedback via absorption and scattering of 
thermal radiation, and surface cooling by drop-wise wetting.  These mechanisms, along 
with other water-based fire suppression topics, were discussed in a comprehensive review 
by Grant, et al. [1].  
Despite the extensive use of sprinklers, little attention has been given to atomization 
and spray dispersion processes.  Although several experimental studies have been 
conducted to characterize drop size, mass flux, and velocity distributions from sprinklers, 
this research has not been fully utilized to formulate physical models characterizing the 
initial spray characteristics and their relationship to sprinkler geometry [2-10].  The 
absence of atomization models for sprinklers continue to result in large uncertainties in 
the specification of initial spray characteristics essential for computer simulations of 
water-based suppression design and analysis.  
 Atomization physics indicate that intermediate processes, including sheet 
formation and disintegration, affect the spray characteristics before drops are formed [11-
20].  Therefore it is important to investigate the sheet breakup dynamics of the initial 
spray in order to gain insight into sprinkler discharge characteristics.  The current study 
focuses on establishing relationships between discharge characteristics and critical 
2
geometric features common to most sprinklers through carefully conceived experiments 
centered on sprinkler geometry, sheet breakup dynamics, volume flux, and initial drop 
size.  
 
1.2 Literature Review
Sprinklers typically use an impinging jet configuration where water is supplied to a 
fire by injecting a continuous liquid stream onto a deflector, forming a radially expanding 
sheet that disintegrates and disperses water in the form of drops.  When studying the 
atomization and dispersion processes of a sprinkler, several physical mechanisms 
contribute to the atomization process governing the initial spray and its subsequent 
dispersion.  These atomization processes have been studied extensively for a variety of 
liquid injection devices.  A few atomization studies are summarized in the following 
section, including those specifically focused on sprinkler sprays.   
1.2.1 Sheet Disintegration Studies 
Dombrowski, et al. conducted a series of seminal studies exploring the physics of 
liquid sheet atomization [12-14].  The experiments in these studies typically involved 
small fan nozzles employed to create flat, wavy liquid sheets at pressures ranging from 
0.21-3.45 bar.  The sheets were visualized using a high-speed flash photography 
technique providing high contrast images of a bright, liquid sheet with a dark, black 
background.  These studies characterized the aerodynamic wave motion on high-velocity 
sheets and provided analysis for determination sheet breakup and ultimately drop size 
based on wave instability concepts.  
3
Huang investigated the breakup radius of axisymmetric sheets formed by the 
impingement of two co-axial jets by exploiting a photographic technique similar to 
Dombrowski [15].  Huang determined a semi-empirical correlation for liquid sheets in 
ambient air, relating the dimensionless sheet breakup radius, 2r
bu
/D
o
, to the Weber 
number at conditions having 800 < We < 40,000. The breakup radius, r
bu
, and the orifice 
diameter, D
o
, describes the dimensionless breakup location while the Weber number is 
defined as the ratio of the inertial forces divided by the surface tension forces, 
 /
2
ol
DUWe = [21].  The liquid density, *
l
, jet velocity U
jet
, and orifice diameter 
describe the inertial forces and surface tension effects are represented by J. Scaling the 
sheet breakup distance with the Weber number was first performed by Ostrach & 
Koestel, who determined the dimensionless breakup distance scaled with 1/We for Weber 
numbers greater than 10,000, an obvious divergence between the experimental results of 
Huang [20].  
Prahl and Wendt conducted a series of experiments to determine breakup locations of 
a sheet generated by a jet impinging on a flat disk [17].  In these moderate Weber number 
experiments having 1600 < We < 4000, controlled disturbances were created using a 
vibrating deflector.  The critical, or most unstable, wavelength was determined by finding 
the forcing frequency (and corresponding wavelength) that provided the earliest breakup 
determined by high-speed photography. These sheet breakup location measurements also 
demonstrated the We
-1/3
 scaling law proposed by Huang. 
More recently, Clanet and Villermaux measured sheet breakup for smooth and 
flapping sheets in an impinging jet configuration using laser induced fluorescence in 
conjunction with short exposure, high-speed photography [18-19].  They established that 
4
the smooth, or steady, sheet transitions to a flapping, unsteady sheet at a critical Weber 
number of 1200.  Once this critical Weber number is exceeded, the dimensionless sheet 
breakup, 2r
bu
/D
o
, follows a We
-1/3
 scaling law.  Both of these conclusions mirror the 
results established by Huang when describing the different sheet breakup regimes and the 
Weber number scaling laws.  
1.2.2 Drop Size Measurements 
Classical sprinkler experiments were conducted by Dundas where drop size 
measurements were performed for six sprinkler configurations with orifice diameters 
ranging from 3.1 - 25.4 mm over a wide range of pressures, 0.345 ? 5.52 bar [2].  A high 
speed flash photography technique was applied to measure droplets in the initial breakup 
region, 0.61 meters from the sprinkler centerline, which involved a tedious photographic 
drop counting method.  Dundas concluded that the dimensionless characteristic drop size,
d
v50
/D
o
, follows a We
-1/3
 scaling law.  The dimensionless characteristic drop size, d
v50
/D
o
,
from Dundas and others is provided in Figure 1.  
Yu performed drop size measurements for 16.3 mm, 13.5 mm, and 12.7 mm diameter 
upright sprinklers at radial locations spanning the entire spray, 3 and 6 meters below the 
head [3].  A laser-based imaging technique for measuring drop size was adopted for these 
experiments, which had previously been used for measurements inside of rain clouds.  
Yu?s results for one sprinkler, operated at two pressures, followed the We
-1/3
 scaling law 
first observed by Dundas, and have been included in Figure 1. 
More recently, Widmann measured discharge characteristics of four residential 
sprinklers with orifice diameters of 8.0 ? 11.0 mm, operated at pressures 0.69 ? 2.0 bar 
[4-6].  A Phase Doppler Interferometer (PDI) was used to measure the drop size and 
5
velocity at radial locations approximately 1 m below the sprinkler head.  The main 
objective of the study was to validate the PDI technique, not the drop size dependence on 
Weber number, therefore only one sprinkler was investigated for a Weber number 
dependency.  Widmann?s measurement of the average mean volume diameter, d
30
, shown 
in Figure 1, demonstrates the We
-1/3
 dependence, except at low pressures, 0.69 bar.   
Clanet and Villermaux measured drop sizes for a jet impinging onto a disk while 
conducting their laser induced fluorescence experiments [19].  The drop size, an average 
value denoted as d
10
, appears to hit a minimum value at large Weber numbers having 
3000 < We < 30,000, as depicted in Figure 1. 
Sheppard performed a series of extensive initial spray experiments testing over 16 
common residential sprinklers, both pendent and upright, for drop size, velocity, and 
trajectory [7-8].  The sprinklers varied in orifice size from 9.5 - 25.4 mm and were 
operated at pressures of 0.345 ? 5.52 bar.  The drop sizes were determined by the PDI 
technique perfected by Widmann, at various circumferential and azimuthal angles within 
the initial spray, either 0.38 or 0.61 m from the sprinkler centerline.  A local 
characteristic drop size, d
v50
, was determined at each location and is provided n Figure 1.  
However, the drop size trend with respect to Weber number was difficult to evaluate 
because an overall drop size for the entire spray was not available; only the local 
measurements were provided. 
Putorti applied a Particle Tracking Velocimetry and Imaging (PTVI) technique, 
which is a laser-based fluorescence technique, to measure drop size, velocity, trajectory, 
and mass flux [9-10]. The nozzles consisted of a jet impinging onto conical sprinkler 
plates at angles of 60?, 90?, and 120?, and orifice diameters ranging from 4.0 ? 8.5 mm  
6
Figure 1. Previous drop size measurements:      Dundas (d
v50
);       
 Yu (d
v50
);      Widmann (d
30
); Sheppard (local d
v50
);       
 Putorti (d
v50
);     Clanet & Villermaux (d
10
).  
 
operated at pressures between 0.21 and 4.34 bar.  Putorti observed a characteristic drop 
size, d
v50
, dependence of We
-2/3
 and his measurements are presented in Figure 1. 
1.3 Research Objectives
The primary purpose of this study is to provide measurements of discharge 
characteristics in canonical sprinkler configurations.  The critical discharge 
characteristics of interest are the drop size, velocity, and location distributions describing 
the spray.  These measurements will provide a foundation for physics-based atomization 
models to be used for sprinkler design and CFD analysis.   
As demonstrated in Figure 1, there is a wide range of drop size measurements that 
have already been conducted for various sprinkler geometries by a variety of 
experimental techniques.  This wide range of drop size measurements is not ideal for the 
7
purposes of model development and validation.  Therefore, a second objective of this 
study is to conduct several careful experiments that yield reproducible and dependable 
results for model development.   
 
8
Chapter 2: Approach 
 
Historically, the most sought after quantity in sprinkler sprays has been the drop size, 
as this quantity closely affects the penetration cooling and oxygen displacement 
performance of the spray.  Characterizing the drop size is the principal focus of this 
study; however, an understanding of atomization physics suggests there are important 
intermediate processes that govern the drop formation.  In this study, these intermediate 
processes are characterized through a series of careful experiments for canonical 
sprinkler configurations.  This approach will provide insight into the atomization physics 
and generate valuable data to support development and validation of a parallel modeling 
study. 
2.1 Atomization Physics
The atomization process for low and medium pressure sprinklers can be described by 
three distinct stages, as illustrated in Figure 2.  The injector forms a vertical water jet that 
impinges upon a striker plate, or boss.  The redirected jet forms a thin, horizontal film 
traveling along the top of the deflector.  Once the film travels past the deflector, it 
transforms into an unconfined, expanding sheet.  The sheet expands radially outwards 
from the deflector becoming increasingly unstable, creating aerodynamic waves.  These 
sinuous waves grow until the sheet begins to breakup at a critical wave amplitude.  The 
sheet disintegrates into ring-like ligaments that are also inherently unstable.  Dilatational 
waves grow on the ligament until they reach a critical wave amplitude, initiating ligament 
breakup into smaller water fragments.  These fragments will eventually contract to form  
9
Figure 2. Theoretical sheet break up processes [22]. 
 
spherical water drops.   A more detailed discussion of atomization processes along with 
the stages of sheet formation and disintegration can be found in Wu [23] and Lefebvre 
[24]. 
2.2 Atomization Measurements
When dissected and studied in great detail, the anatomy of a sprinkler is a complex 
configuration that needs to be characterized carefully to ensure each geometric 
characteristic is represented accurately.  Figure 3 depicts the sprinklers investigated in 
this study as well as the important geometric characteristics common to most sprinklers.   
In this study, three different canonical sprinkler configurations were investigated, 
identified as the Basis, Tined, and Standard nozzles.  The Basis nozzle consists of a 
separate injector and deflector disk.  It does not have tines, frame arms, or a boss.  This 
Y
GothofWaves
Lgamen

op
Seet
Fomatin
Jet
Deflcor
Seet

igamt
Growth  
of waves 
Jet 
Deflector 
Sheet 
 Formation 
Ligaments        Drops 
Drop 
 Formation 
Sheet       Ligaments 
10
Figure 3. The anatomy of a sprinkler: (1) Inlet (2) Frame Arms (3) Boss (4) Deflector. 
 
canonical configuration provides a useful baseline for evaluating the impact of additional 
geometric features on discharge characteristics.  The Tined nozzle configuration helps to 
isolate and explore the effect of tines and spaces on a deflector.  The nozzle was 
fabricated by removing the conical boss from a standard Tyco D3 nozzle, leaving a flat, 
notched deflector.  A conventional Tyco D3 nozzle was utilized for the Standard nozzle 
adding boss effects and extending the nozzle geometry to a commercially available 
configuration.  Table 1 summarizes the important geometric and flow characteristics for 
all three nozzles used in this study. 
The influence of sprinkler spray geometry on discharge characteristics is aided by 
measurement and analysis of intermediate processes occurring during atomization.  The 
two intermediate processes measured in this study were sheet trajectory and sheet 
breakup location.  These measurements provide insight into the dispersion behavior of 
the spray along with valuable information for atomization model validation and 
development.  In addition to examining the intermediate processes leading to drop  
boss 
D
D


def 
tine 
space
Section A-A 
t
arm
D
L
L

boss 
o
inlet 
jet 
inlet 
A A
D
L
arm 
1
3
4
2
Center Tine (0?) 
11
Table 1. Experimental Sprinkler Dimensions 
 
Basis Nozzle 
Tined  
Nozzle 
Standard 
Nozzle 
D
inlet
 (mm) 19 19 19 19.5 19.5 
L
inlet
 (mm) 25.4 25.4 25.4 19.7 19.7 
L
jet 
(mm) 25.4 25.4 25.4 23 23 
D
o
(mm) 3.5 6.7 9.7 6.35 6.35 
Inlet  
Characteristics 
K-Factor 
(lpm/bar
1/2
)
7.2 25.9 49.0 25.9 25.9 
D
def 
(mm) 38 38 38 25.4 25.4 

tine  
(?) None None None 21 21 
Deflector  
Characteristics 

space  
(?) None None None 9 9
D
boss 
(mm) None None None 12 12 
Boss  
Characteristics 

boss 
(?) None None None 180 65 
formation, other spray characteristics were measured including volume flux and drop size 
distributions along the radial extent of the spray.  These detailed measurements were used 
to determine global spray characteristics such as the overall d
v50
.
2.3 Diagnostics
Sheet trajectory, sheet breakup, volume density, and local drop size experiments were 
conducted to quantify discharge characteristics for each of the experimental nozzle 
configurations.  The experiments were performed at 0.69, 1.38, 2.07, and 2.76 bar to 
investigate the effect of injection pressure on spray characteristics.  Table 2 summarizes 
the pertinent experimental injection parameters including injection pressure, P, jet 
velocity, U
jet
, nozzle flow rate, Q, Weber number, We, and Reynolds number, Re, where 
the Weber and Reynolds numbers are based upon the jet velocity.  
2.3.1 Sheet Trajectory 
Sheet trajectory experiments were conducted to track the path of the radially 
12
Table 2. Experimental Injection Parameters 
 D
inlet
 (mm) PP (bar) U
jet 
(m/s) Q (lpm) We  Re  
0.69 11.8 1.6 6509 41059 
1.38 16.6 2.2 13019 58066 
2.07 20.4 2.7 19528 71116 
D
o
= 3.5 
2.76 23.5 3.2 26038 82118 
0.69 11.8 5.7 12461 78598 
1.38 16.6 8.0 24922 111155 
2.07 20.4 9.9 37383 136136 
D
o
= 6.7 
2.76 23.5 11.4 49844 157197 
0.69 11.8 10.8 18041 113792 
1.38 16.6 15.2 36081 160926 
2.07 20.4 18.6 54122 197093 
Basis Nozzle 
D
o
= 9.7 
2.76 23.5 21.5 72162 227583 
0.69 11.8 5.7 11810 74493 
1.38 16.6 8.0 23620 105348 
2.07 20.4 9.9 35430 129025 
Tined Nozzle  D
o
= 6.35 
2.76 23.5 11.4 47240 148985 
0.69 11.8 5.7 11810 74493 
1.38 16.6 8.0 23620 105348 
2.07 20.4 9.9 35430 129025 
Standard Nozzle D
o
= 6.35 
2.76 23.5 11.4 47240 148985 
expanding sheet beyond the edge of the deflector.  The experiments were conducted 
inside a vented 1.7 m x 1.7 m x 1.9 m chamber illustrated in Figure 4.  Planar Laser 
Induced Fluorescence (PLIF) was utilized to visualize a cross-section of the expanding 
sheet.  Illumination was provided by a 500 MW, air cooled, argon ion laser and a 20 face 
rotating mirror spinning at 20 Hz.  The water supply was seeded with a rhodamine 6G 
dye having a mass concentration of 0.5 mg/l.  The sheet was imaged with a low noise, 16-
bit, 2.0 mega-pixel, Cooke SE ? high-speed digital video camera fitted with a high pass 
optical filter operated with an electronic shutter speed of 900 ?s at 5 frames per second.  
Comparison of several images at each flow condition revealed that the trajectory, before 
13
Figure 4. Sheet trajectory and sheet breakup diagnostic set-up. 
 
breakup, is essentially time-independent.  A single image was thus used along with a 
spatial calibration for determination of the sheet trajectory as shown in Figure 5. 
2.3.2 Sheet Breakup 
The PLIF technique described previously also provides images for determining the 
radial location where the flapping, wavy sheet breaks up completely into ligaments.  The 
PLIF image shown in Figure 5 is also representative of those used to determine the 
breakup distance, where the breakup distance was defined as the location where the 
continuous sheet no longer exists and only ligaments remain.  At least 100 images were 
analyzed at each experimental condition to ensure statistically steady results.  However, 
the PLIF technique did not provide satisfactory results for very thin sheets with long 
Rotating 
Mirror 
Laser 
Source
Nozzle 
Flow 
Meter 
Water and 
Rhodamine Dye 
Foam 
Water 
Pump 
Camera View 
Computer 
1.7 m 
1.9 m 
Pressure 
Transducer 
Laser Beam 
Reflector 
Flash 
Camera
Alternative 
sheet breakup 
technique 
14
Figure 5. An inverted PLIF photo of the Standard nozzle at 2.07 bar. 
 
breakup distances, such as those produced by the smallest Basis nozzle.  An alternative 
photographic technique was used to determine sheet breakup locations in these cases, as 
previously illustrated in Figure 4.  In this photographic method, the sheet was illuminated 
with a 15.6 ?s diffuse-reflected Canon 550EX flash.  A Canon D30 Digital SLR camera 
was placed above the nozzle to photograph the sheet breakup producing images similar to 
the one depicted in Figure 6.  Where, once again, the breakup distance was defined as the 
location where the sheet is completely broken up into ligaments.  At least 10 images were 
recorded at each experimental condition and multiple breakup locations were obtained at 
circumferential stations distributed around the sheet.  This alternative technique was used 
for all Basis nozzle breakup measurements because they produced thin, long sheets 
(especially the smallest nozzle).  On the other hand, the Tined and Standard nozzles were 
determined by the PLIF technique because the complexities of their sprays resulted in 
indecipherable images using the alternative technique.  Comparisons of the breakup 
locations for the 6.7 mm Basis nozzle, using both photographic techniques, showed 
agreement within 12.8%, verifying that both techniques can effectively determine the 
breakup distance.  
10 mm 
Jet 
Deflector 
Boss
15
Figure 6. Characteristic sheet breakup photo from above the spray [22].  
 
2.3.3 Volume Density 
Knowledge of the volume distribution dispersed onto the floor is helpful in model 
development as well as critical to the determination of the overall characteristic drop size, 
d
v50
, described later.  Volume density distributions were obtained using a 2.6 m 
patternator positioned 1 m below the nozzle deflector surface and 1 m above the floor.  
To permit analysis of the entire sprinkler spray, volume density measurements were 
conducted inside a large 8.6 m x 7.2 m x 3 m room located at the Maryland Fire Rescue 
Institute, as illustrated in Figure 7.  The nozzles were discharged for 10 minutes to 
average over short time scale aerodynamic or water supply fluctuations, after which the 
water in each cup was weighed to determine the volume at each radial station.  After 
verifying the axisymmetry and repeatability of the Basis nozzle spray at 0? and 30? 
stations (? 2%), radial volume density distributions were obtained only at the 0? station.  
Meanwhile, the Tined and Standard nozzles were tested at the 0? and 15? stations.  These 
stations were aligned with the middle of the center tine and adjacent space, respectively, 
as illustrated in Figure 2. 
Deflector 
50 mm 
16
Figure 7. MFRI facility plan view and diagnostics. 
 
A characteristic dispersion length scale, R, first introduced by Prahl and Wendt [17], 
was employed to facilitate analysis of the measurements.  This reference quantity 
provides an inviscid radial location at the measurement elevation for each experimental 
condition, and is given by 



	























+








=
2/1
2
2/1
2
2/1
2
)(
2
)(
1
2
)(
gh
v
gh
v
g
h
vR
zozo
ro
, (2.1) 
where h is the measurement elevation (below the nozzle), g is the gravitational constant, 
(v
o
)
r
is the initial radial sheet velocity, and (v
o
)
z
is the initial vertical sheet velocity.  The 
velocity magnitude is estimated using a model describing viscous interaction with the 
deflector [25], and the angle is determined from the average initial angle at that 
experimental condition determined by the trajectory measurements, yielding sheet 
velocities (v
o
)
r
and (v
o
)
z
. The resulting volume density distributions in the r/R coordinate 
30? 
0? 
15? 
Nozzle 
2.0 m 3.0 m 4.0 m 5.0 m 1.0 m 
Spraytec 
8.6 m 
7.2 m 
 Patternator  
 
2.0 m 
17
describe the relative effect of drag on dispersion.  The volume density measurements 
were described non-dimensionally so that  
1=

 rq
i
, (2.2) 
where Rrr /=
 is the dimensionless station width and the dimensionless linear density 
of dispersed volume flow,
i
q
 , is given by  
()r
RQ
q
q
i
i

=
 2
/
2

, (2.3) 
where 
i
q is the volume flux measured at drop size i and Q is the nozzle flow rate.  The 
length scale R was modified at every experimental condition except at measurements 
aligned with spaces (15? station) for the Tined and Standard nozzles.  At these 
measurement stations the initial angle and corresponding velocity magnitude could not be 
determined from the trajectory experiments described previously.  For these cases, R
values at measurement locations aligned with the tine (0? station) were used for 
measurement locations aligned with the space (15? station) to facilitate comparison. 
2.3.4 Drop Size 
Local drop size measurements were also conducted inside the large room illustrated in 
Figure 7 to investigate the drop size variations along the radial span of the spray.  An 
overall drop size distribution and a characteristic drop size, d
v50
, for each experimental 
condition can be derived from these measurements.  The local drop sizes were measured 
using a Spraytec particle analyzer developed by Malvern Instruments [26].  This laser-
based instrument employs a light diffraction technique for counting and sizing drops or 
particles.  As illustrated in Figure 8, the instrument?s sampling volume is created with a 
collimated laser diode.  Drops entering this sampling volume diffract light at various 
18
angles according to their size.  This light is collected by detector rings to measure the 
intensity distribution.  The intensity distribution is used along with proprietary 
correlations to calculate the drop size distribution.  The entire signal is then focused onto 
a power detector which measures attenuation of the incident light providing an estimate 
of the concentration.   
Local measurements were taken at 12 radial stations starting at 0.5 m and separated by 
0.5 m, positioned 1 m below the nozzle and 1 m above the floor.  The Spraytec 
measurement volume was configured to be 12 mm in diameter and 130 mm long.  At 
least 100 drops were estimated to fill the measurement volume for valid measurement 
records.  Measurements were taken at each station for 1 minute at 50 Hz providing local 
drop size distribution realizations.  The drop sizes were measured at the same 
circumferential stations as the volume flux experiments and in the case of the Basis 
nozzle demonstrated similar axisymmetric behavior.   
The drop size distribution determined by the Malvern RTSizer software  
v 5.3.1.0 is a local drop size distribution within the Spraytec?s measurement volume [27]. 
 
Figure 8. Malvern diffraction technique. 
 
Laser 
Diode
Laser  
Beam 
Scattering 
Detector 
Lens
Beam Power 
Detector 
Drops 
Scattered 
Incident Light 
19
However, in this study an overall characteristic drop size, d
v50
, for the entire spray is 
of interest.  To determine the overall d
v50
 the local Spraytec measurement is weighted 
with the local volume density measurements to transform the spatial Malvern 
measurements into the flux-based drop size distributions described in equations 2.4 ? 2.6.  
 
(2.4) 
 
(2.5) 
 
(2.6) 
 
where 
ji
VF
,
is the local spatial volume fraction within the Spraytec measurement volume 
for the ith drop size at the jth measurement location.  The spray quantities 
i
Q and 
T
Q are 
estimates of the drop-wise volume flux and total volume flux from all drops, respectively.  
The quantity 
i
QF is the flux-based drop-wise volume flux fraction for the entire spray.  
Drop size distributions based on 60 drop size bins ranging from 0.29 ? 2000 ?m are 
easily calculated from 
i
QF  for determining flux based drop characteristics.    
The upper measurement limit of the Spraytec (d
v50
 = 850 ?m) presents a challenge for 
measuring the larger drop sizes produced by the nozzles used in this study.  This limit is 
clearly observed in the radial drop size distribution measurements for the Basis nozzle 
configurations presented in Figure 9.  The figure also demonstrates that under quiescent 
conditions, the drag effects result in reduced penetration of smaller drops in the 
dimensionless coordinate when compared to larger drops, regardless of the experimental 
Tii
M
i
N
j
jijT
N
j
jiji
QQFQ
VFrqQ
VFrqQ
/
)(
)(
1 1
,
1
,
=


=


=


= =
=
M
i
N
j
jijT
VFrqQ )(
1 1
,


=

= =
Tii
QQFQ /=
20
condition.  However, at extreme radial locations, surprisingly drop size measurements did 
not change significantly with location and remained slightly below the Spraytec d
v50
 
limit.  The spatial separation of drops is easily predicted through drop dispersion 
calculations, including drag effects, for the various drops [25].  The favorable comparison 
between model predictions and valid Spraytec measurements, demonstrated in Figure 9, 
provides an opportunity to correct the erroneous drop size measurements at extreme 
radial locations.  At these locations, drop size measurements were replaced with those 
from the drop dispersion model estimates.  It should be noted that the contaminated 
region typically consisted of only 5 -15% of the total mass of the spray for the Basis and 
Tined nozzles and even less for the Standard nozzle, resulting in only small adjustments 
to the overall drop size distribution and the associated characteristic drop sizes. 
 
Figure 9. Malvern drop size limitations:        Drop Dispersion Analysis; 
Basis nozzle:      D
o
= 3.5 mm;      D
o
= 6.7 mm;      D
o
= 9.7 mm. 
Spraytec Limit 
Drop Dispersion  
Analysis
21
Chapter 3: Results and Discussion 
 3.1 Sheet Trajectory
The initial angle, , was determined for each nozzle configuration and experimental 
condition through analysis of the PLIF images.  Trajectory measurements performed on 
the Basis nozzle and on the other nozzles aligned with the center of the tine (0? station) 
revealed that the sheet does not deviate from its initial angle.  Significant curvature was 
only observed in the Standard nozzle operating at low pressures.  The initial angle does 
not exhibit any consistent Weber number functionality for a given nozzle.  Instead, 
dependence between the initial angle and orifice diameter was observed, demonstrated in 
Figure 10 and summarized in Table 3.  
The addition of spaces on the deflector in the Tined nozzle does not have a significant 
 
Figure 10. Sheet trajectory measurements;           Average Angle; 
 Basis nozzle:      D
o
= 3.5 mm;      D
o
= 6.7 mm;      D
o
= 9.7 mm;  
 Tined nozzle:       D
o
= 6.35 mm; Standard nozzle:      D
o
= 6.35 mm. 
22
effect upon the initial angle when compared to the Basis nozzle value with a similar 
orifice diameter.  These nozzles have initial angles of 4.2? and 4.7?, respectively.  
However, the Standard nozzle had an average initial angle of 3.5?, which is smaller than 
both the Tined and Basis nozzles at approximately the same orifice diameter. 
For the Tined and Standard nozzles the trajectory of the sheet exiting the tine was 
easy to distinguish.  However, the sheet trajectory in the spaces between the tines was 
difficult to discern due to the limitations of the measurement equipment and complexities 
of the spray in the spaces.  To obtain insight into the flow of water in the spaces, 
qualitative PLIF images were acquired in planes orthogonal to the center tine at radial 
locations 12.7 mm, 22.7 mm, and 62.7 mm, from the center of the deflector, as illustrated 
in Figure 11(a).  
 
Table 3. Summary of Discharge Characteristics 
 
Basis  
Nozzle 
Tined  
Nozzle 
Standard 
Nozzle 
PP
(bar) 
U
jet
 
(m/s) 
D
o
= 3.5 D
o
= 6.7 D
o
= 9.7 
Tine 
(0?) 
Space 
(15?) 
Tine 
(0?) 
Space 
(15?) 
0.69 11.8 9.47 4.00 3.99 4.56 N/A 2.67 N/A 
1.38 16.6 7.87 4.19 3.50 4.03 N/A 4.05 N/A 
2.07 20.4 10.10 5.23 3.49 4.10 N/A 3.41 N/A 
Initial Angle, 
 (?) 
2.76 23.5 9.65 5.53 3.50 4.25 N/A 3.81 N/A 
Average,  (?) N/A N/A 9.27 4.74 3.62 4.23 N/A 3.49 N/A 
0.69 11.8 49.43 37.58 37.11 30.56 N/A 21.48 N/A 
1.38 16.6 39.18 31.91 33.81 29.27 N/A 17.62 N/A 
2.07 20.4 37.39 30.72 N/A 27.31 N/A 15.30 N/A 
Sheet Breakup 
Distance, 
2r
bu
/D
o
2.76 23.5 34.64 28.72 N/A 25.87 N/A 13.46 N/A 
0.69 11.8 0.155 0.091 0.085 0.101 0.104 0.087 0.056 
1.38 16.6 0.149 0.081 0.076 0.097 0.100 0.073 0.053 
2.07 20.4 0.143 0.082 0.080 0.094 0.103 0.062 0.048 
Overall 
Characteristic 
Drop Size, 
d
v50
/D
o
2.76 23.5 0.146 0.081 0.074 0.089 0.101 0.059 0.045 
23
Figure 11. Inverted PLIF images depicting flow through sprinkler spaces: (a) Top view 
of measurement locations (b) Tined nozzle (c) Standard nozzle.  
 
Figure 11(b) demonstrates the Tined nozzle creates a relatively flat sheet that is not 
significantly affected by the spaces between the tines.  In contrast, the Standard nozzle, 
shown in Figure 11(c), appears to direct a significant amount of water through the spaces.  
This creates a three dimensional sheet formed by the flow over the tines and the flow 
forced through the spaces by the boss.  
3.2 Sheet Breakup
High-speed flash photography was used to determine sheet breakup for the Basis 
nozzle, while a PLIF technique was employed for the more complex Tined and Standard 
nozzles.  As described in Figure 12 and summarized in Table 3, the dimensionless sheet 
breakup location for each experimental condition is presented with respect to the Weber 
number along with data from Huang, a previous investigator who measured the  
(12.7 mm) 
(12.7 mm) 
(22.7 mm) 
(62.7 mm) 
Tine 
Edges 
(22.7 mm) 
(62.7 mm) 
Image
Planes 
(12.7 mm) 
(22.7 mm) 
(62.7 mm) 
(a) (b) (c)
24
Figure 12. Dimensionless sheet breakup distances:         Huang 
Correlation;          Basis Correlation;          Standard Correlation; Basis 
nozzle:      D
o
= 3.5 mm;      D
o
= 6.7 mm;      D
o
= 9.7 mm; Tined nozzle:         
 D
o
= 6.35 m; Standard nozzle:      D
o
= 6.35 mm. 
 
sheet breakup distance for axisymmetric sheets [15].  Huang proposed a semi-empirical 
correlation for his axisymmetric sheets described as  
2r
bu
/D
o
= 1250We
-1/3
 (3.1) 
The breakup distance of the Basis nozzle for all three orifice diameters correlates well 
with one another and follows the We
-1/3
 scaling law proposed by Huang.  This study?s 
empirical correlation for the Basis nozzle was determined to be 
2r
bu
/D
o
= 964We
-1/3
, (3.2) 
which falls below the correlation determined by Huang (2r
bu
/D
o
= 1250We
-1/3
).  This 
discrepancy could be a result of differing methodologies for creating the horizontal, 
axisymmetric sheets.  Huang used two opposed impinging jets to create his radially 
25
expanding sheets, while a single jet impinging upon a flat deflector surface was used in 
this study. 
Sheet breakup measurements of the Tined nozzle also follow a We
-1/3
 scaling at 
higher pressure conditions, with the exception of the lowest pressure (0.69 bar) data 
point.  However, the sheets created with the Tined nozzle break up sooner than that of the 
Basis nozzle at a similar orifice diameter.  The scaling of the Tined nozzle appears to 
follow the We
-1/3
 at higher pressure conditions, with a break down at the lowest pressure 
(0.69 bar), resulting in an outlaying data point.  The Standard nozzle sheet breakup 
distances also demonstrated the We
-1/3
 scaling law and were the shortest when compared 
to the two other configurations.  The empirical correlation for Standard nozzle 
configuration was determined to be 
 2r
bu
/D
o
= 495We
-1/3
, (3.3) 
breaking up at approximately one-half the distance of the Basis nozzle.  From these 
experiments it is clear the addition of first tines and spaces, and then the boss promotes 
sheet instability, resulting in earlier sheet disintegration.   
3.3 Volume Density
In Figure 13, results from volume distribution experiments are presented in terms of a 
dimensionless linear volume density, q
 , and a dimensionless radial location, r/R. The 
Basis nozzle presented in Figure 13(a) ? (c), demonstrates the effect of increasing the 
orifice diameter.  As the orifice diameter increases the linear density peak shifts radially 
outwards, indicating more volume is delivered at extreme radial locations, reflecting a 
greater contribution from large drops.  In contrast, pressure has little effect on the 
26
dimensionless volume density distribution as all four experimental conditions have 
similar shape. 
Comparisons between the Basis, Tined, and Standard nozzles aligned with the tine (0? 
station) are provided in Figure 13(d) ? (f) and comparisons between the three 
configurations aligned with the space (15? station) are depicted in Figure 13(g) ? (i).  
Comparison of Figure 13(b), Figure 13(e) and Figure 13(h) reveal that the addition of 
tines in the Tined nozzle has little effect on the volume distribution when compared to the 
Basis nozzle with a similar orifice diameter.  On the other hand, the addition of the boss 
in the Standard nozzle, Figure 13(f) and Figure 13(i), had a profound effect on the 
volume distribution throughout the spray when compared to the similar orifice diameters 
of the Basis nozzle, Figure 13(b), and Tined nozzle, Figure 13(e) and Figure 13(h).  Also 
in contrast to the Tined and Basis nozzle, the volume density of the Standard nozzle 
depends significantly on the pressure, especially for measurements aligned with the tine 
(0? station).  
3.4 Drop Size Distributions
Local drop size measurements were performed at 0.5 m stations spanning the entire 
sprinkler spray.  The results are described in Figure 14 and summarized in Table 3 for all 
nozzle configurations and experimental conditions.  Once again, the similarities between 
measurements from the Basis and Tined nozzles, Figure 14(a), are immediately apparent; 
whereas the spray from the Standard nozzle, Figure 14(b), behaves differently than the 
other configurations.  For the Basis and Tined nozzles, the dimensionless location of 
drops passing through the measurement elevation is determined by the size of the drop.  
As drops travel from the deflector they are separated by drag effects in the air, resulting  
27
Figure 13. Volume Density measurements for all nozzle configurations and conditions: 
0.69 bar;        1.38 bar;         2.07 bar;         2.76 bar; Basis nozzle: (a) D
o
= 3.5 mm, 
(b) D
o
= 6.7 mm, (c) D
o
= 9.7 mm, (d) D
o
= 6.7 mm; Tined nozzle (Tine, 0?): (e) D
o
= 6.35 
mm; Standard nozzle (Tine, 0?): (f) D
o
= 6.35 mm; Basis nozzle: (g) D
o
= 6.7 mm; Tined 
nozzle (Space, 15?): (h) D
o
= 6.35 mm; Standard nozzle (Space, 15?): (i) D
o
= 6.35 mm. 
(d) (g) 
(h) (e) 
(b) 
(i) 
(f) 
(c) 
(a) Basis nozzle 
D
o
= 3.5 mm 
Tined nozzle  
(Tine, 0?) 
D
o
= 6.35 mm 
Basis nozzle 
D
o
= 9.7 mm 
Basis nozzle 
D
o
= 6.7 mm 
Basis nozzle 
D
o
= 6.7 mm 
Standard nozzle  
(Tine, 0?) 
D
o
= 6.35 mm 
Basis nozzle 
D
o
= 6.7 mm 
Standard nozzle  
(Space, 15?) 
D
o
= 6.35 mm 
Tined nozzle  
(Space, 15?) 
D
o
= 6.35 mm 
28
Figure 14. Local d
v50
 measurements for all nozzle configurations and conditions:  
(a) Basis nozzle:     D
o
= 3.5 mm;      D
o
= 6.7 mm;      D
o
= 9.7 mm; Tined nozzle (Tine, 
0?):      D
o
= 6.35 m; (b) Standard nozzle (Tine, 0?):     D
o
= 6.35 mm. 
 
in smaller drops traveling shorter distances and larger drops traveling the furthest, as 
clearly seen in Figure 14(a).  In contrast, with the addition of the boss in the Standard 
nozzle, larger drops appear much closer to the centerline, resulting from drops being 
formed by the flow of water directed through the spaces.  This process creates a 
significantly different drop dispersion curve when compared to the Basis and Tined 
nozzles, as shown in Figure 14(a) and Figure 14(b).  
When presenting drop size data, it is often useful to find overall spray characteristics 
for a given nozzle and experimental condition.  The cumulative volume fraction is an 
extremely useful overall spray quantity describing the percentage of the total spray 
volume contained in drop sizes smaller than a specific drop diameter, for each nozzle and 
experimental condition.  It has been shown that a Rosin-Rammler/log-normal drop size 
(a) 
(b) 
Contaminated 
Measurement 
Area 
Contaminated 
Measurement 
Area 
29
distribution predicts the cumulative volume fraction for sprinkler sprays, which in turn is 
useful for modeling purposes [28].  The Rosin-Rammler/log-normal distribution is 
defined as 
 
() ()
()[]














 



=







w
vCVF
CVF
v
dd
d
w
dd
e
ddedw
CVF
)/(693.0
0
2
/ln
1
2
1
50
2
2
50
1
2
where CVF is the cumulative volume fraction of drops with diameters less than d
CVF
 , w
is a correlation coefficient, ()( )( ) www /15.12ln22
1
2/1
==


 , found empirically.  Figure 
15(a) and Figure 15(b) provide sample drop size distributions for the Basis nozzle, D
o
=
9.7 mm and the Standard nozzle at 2.07 bar, respectively.  A Rosin-Rammler/log-normal 
equation curve fit to the cumulative volume fraction is also included to determine w,
Figure 15. Drop size distribution at 2.07 bar: (a) Basis nozzle, D
o
= 9.7 mm, (b) 
Standard nozzle, D
o
= 6.35 mm;           CVF;          Rosin-Rammler/log-normal. 
(b) (a) 
50vCVF
dd !
(3.4)                                         
CVFv
dd <
50
30 
 
yielding w = 2.6 with a d
v50
 = 781 ?m for the Basis nozzle and w = 2.0 with a d
v50
 = 393 
?m for the Standard nozzle.  The agreement between the Standard nozzle and the Rosin-
Rammler/log-normal distribution is very close, while the curve fit doesn?t appear to 
predict the large number of smaller drops found in the Basis nozzle.  
Although local drop size measurements are useful, the primary objective of these 
measurements was to determine an overall characteristic drop size, d
v50
, for the entire 
spray.  Dimensionless overall characteristic drop sizes, d
v50
/D
o
, for each experimental 
condition and nozzle configuration are presented in Figure 16 and have been summarized 
previously in Table 3.  Similar to drop size results provided by Clanet and Villermaux 
[19] for a flat disk deflector configuration, the characteristic drop size for all three Basis  
 
Figure 16. Experimental drop size results: (a) Clanet & Villermaux; Basis nozzle:       
 D
o
= 3.5 mm;      D
o
= 6.7 mm;      D
o
= 9.7 mm; Tined nozzle:       D
o
= 6.35 m; (b) 
Standard nozzle:     D
o
= 6.35 mm;          Dundas Correlation;        Standard Correlation.     
 
(b)(a)
We
-1/3
31 
 
nozzles shown in Figure 16(a) did not demonstrate a strong dependence on the Weber 
number.  Following the same trend, the characteristic drop size measurements for the 
Tined nozzle, aligned with the tine (0? station), did not significantly differ from the Basis 
nozzle results.  A very different trend is depicted in Figure 16(b), where the drop size 
measurements for the Standard nozzle aligned with the tine (0? station), followed the  
We
-1/3 
scaling law proposed by Dundas [2] for sprinkler configurations. Dundas proposed 
the empirical correlation  
3/1
50
/

= CWeDd
ov
(3.5) 
shown in Figure 16(b), where C = 1.41, for the sprinkler configuration he tested.  The 
constant, C, has been shown by Yu [3] and Sheppard [7], both testing a variety of 
commercially available sprinklers, to depend upon the sprinkler type.  In this study, the 
experimental constant that best matched the drop size measurements was determined to 
be 2.04 for the Standard nozzle.   
In addition to drop size measurements at locations aligned with the tine (0? station), 
measurements were also performed at locations aligned with the space (15? station).  
Overall drop size measurements for the Tined and Standard nozzles at each experimental 
condition have been included in Figure 17.  The overall characteristic drop sizes from 
measurements of the Tined nozzle aligned with the space essentially have no Weber 
number dependence and are similar in drop diameter to the measurements aligned with 
the tine.  However, drop size measurements from the Standard nozzle aligned with the 
space follow the expected We
-1/3
 power law decay with considerably smaller drop sizes 
when compared to the measurements aligned with the tine.  For the Standard nozzle the  
32 
 
Figure 17. Dimensionless drop size results for spaces and tines:
Tined nozzle:      Tine (0?);       Space (15?); Standard nozzle:        
 Tine (0?);      Space (15?).         
 
constant C from equation (3.5) when aligned with the space (15? station) was found to be 
C = 1.48.   
 
33 
 
Chapter 4: Conclusions 
 
A series of experiments have been conducted to characterize the initial spray of 
canonical sprinkler configurations.  Measurements were taken to determine the initial 
angle of the sheet as it exits the deflector, the distance the sheet travels before break-up, 
the volume and drop size distribution 1 meter below the sprinkler head.  
The initial sheet angle exiting the deflector appears to be a function of the orifice 
diameter, and not pressure, where the initial angles fell in a range of 3.6? - 9.3? for the 
Basis nozzles with orifice diameters ranging from 3.5 mm ? 9.7 mm.  The addition of 
tines appeared to have a minimal impact on the initial angle of the sheet as it exits the 
deflector.  However, the addition of a boss with tines in the Standard nozzle does appear 
to create a smaller average initial angle, 3.5?, when compared to the Tined nozzle, at 4.2?.  
Unfortunately, the trajectory of the water falling through the spaces of the tines could 
only be estimated qualitatively in this study. 
The sheet breakup distances followed the previously proposed scaling law of We
-1/3
 
for the Basis and Standard nozzle with empirical correlations of 2r
bu
 /D
o
= 964We
-1/3
 and 
2r
bu
 /D
o
= 495We
-1/3
, respectively.  Similar to the initial angle measurements, the addition 
of tines on the deflector does not have a significant influence on the sheet breakup 
distance, unless a boss is also present.   
The volume density distributions, when normalized by the characteristic dispersion 
length scale, R, did not vary significantly with an increasing Weber number for the Basis 
and Tined nozzles.  Therefore, the addition of tines in a deflector did not have an effect 
on the volume distribution to the floor.  However, with the addition of tines and a boss on 
the Standard nozzle, the volume density distributions were dramatically different than 
34 
 
those seen in the Basis and Tined nozzles.  The volume density for the Standard nozzle 
saw significant amounts of water lost through the spaces at early radial locations, a 
characteristic not seen in the previous sprinkler configurations.    
For the Basis and Tined nozzles, the characteristic drop sizes did not significantly 
change with respect to Weber number for a given configuration.  This result, though 
unexpected, was in good agreement with the work of Clanet and Villermaux, who 
measured the drop size of an impinging jet configuration similar to this study?s Basis 
nozzle.  The Standard nozzle, the only commercially available nozzle tested, did follow 
the We
-1/3
 scaling laws proposed by Dundas for sprinklers.  This result suggests that the 
Basis nozzle, with its flat disk deflector, has unique and different Weber number 
functionality than what has been determined for conventional sprinklers.  At this time it is 
not known which atomization mechanisms are responsible for this difference.  With 
better diagnostics a more detailed look at the complex sheets produced by common 
sprinkler nozzles can be studied to determine those physical mechanisms.   
 
35 
 
References 
[1]    Grant, G., Brenton, J., and Drysdale, D., ?Fire Suppression by Water Sprays,? 
Progress in Energy and Combustion Science, Vol. 26, pp. 79-130, 2000. 
 
[2]  Dundas, P. H., ?Technical Report Optimization of Sprinkler Fire Protection The     
Scaling of Sprinkler Discharge: Prediction of Drop Size,? FMRC Serial No. 18792 
RC73-T-40, Factory Mutual Research Corporation, Norwood, MA, June 1974.  
 
[3]  Yu, H. Z., ?Investigation of Spray Patterns of Selected Sprinklers with the FMRC 
Drop Size Measuring System,? First International Symposium on Fire Safety 
Science, New York, pp. 1165-1176, 1986. 
 
[4]  Widmann, J. F., Sheppard, D. T., and Lueptow, R. M., ?Non-Intrusive 
Measurements in Fire Sprinkler Sprays?, Fire Technology, Vol. 37, pp. 297-315, 
2001. 
 
[5]  Widmann, J. F., ?Phase Droppler Interferometry Measurements in Water Sprays 
Produced by Residential Fire Sprinklers?, Fire Safety Journal, Vol. 36, pp. 545-
567, 2001. 
 
[6]  Widmann, J. F., ?Characterization of a Residential Fire Sprinkler using Phase 
Doppler Interferometry?, NISTIR 6561, National Institute of Standards and 
Technology, Gaithersburg, MD, 2000. 
 
[7]  Sheppard, D. T., ?Spray Characteristic of Fire Sprinklers,? NIST GCR 02-838, 
National Institute of Standards and Technology, Gaithersburg, MD, 2002. 
 
[8]  Sheppard, D. T., Gandhi, P.D., ?Understanding Sprinkler Sprays: Trajectory 
Analysis? NISTIR 6561, National Institute of Standards and Technology,
Gaithersburg, MD, 2000. 
 
[9]  Putorti A. D., Belsinger T. D., and Twilley W. H., ? Determination of Water Spray 
Drop Size and Speed from a Standard orifice, Pendent Spray Sprinkler?, Report of 
Test. NIST, Gaithersburg, MD, May 1999. 
 
[10]  Putorti A.D., ?Simultaneous Measurements of Drop Size and Velocity in Large-
Scale Sprinkler Flow Using Particle Tracking and Laser-Induced Fluorescence?, 
NIST GCR-861, National Institute of Standards and Technology, Gaithersburg, 
MD, 2004 
 
[11] Fraser, R. P., Eisenklam, P., ?Research into the Performance of Atomizers for 
Liquids?, Imp. Coll. Chem. Eng. Soc. J., Vol. 7, pp. 52-68, 1953. 
 
36 
 
[12] Dombrowski, N., Fraser, R. P., ?Photographic Investigation into the Disintegration 
of Liquid Sheets?, Philos. Trans. R. Soc. London Ser. A, Math. Phys. Sci., Vol. 247, 
No. 924, pp. 101-130, 1954.  
 
[13] Dombrowski, N., Hasson, D., Ward, D. E., ?Some Aspects of Liquid Flow Through 
Fan Spray Nozzles?, Imp. Coll. Chem. Eng. Soc. J., Vol. 12, pp. 35-50, 1960.  
 
[14] Dombrowski, N., Hooper, P. C., ?The effect of Ambient Density on Drop 
Formation in Sprays?, Imp. Coll. Chem. Eng. Soc. J., Vol. 17, pp. 291-305, 1962.  
 
[15] Huang, J. C., ?The Break-up of Axisymmetric Liquid Sheets?,  Journal of Fluid 
Mech. Vol. 43, part 2, pp. 305-319, 1970. 
 
[16] Arai, T., Hashimoto, H., ?Disintegration of a Thin Liquid Sheet in a Cocurrent Gas 
Stream?, Proceesings of the 3
rd
 International Conference on Liquid Atomization 
and Spray Systems, London, 1985, pp. V1B/1/1-7.  
 
[17] Prahl, J. M., Wendt, B., ?Dishcharge Distribution Performance for an Axisymmetric 
Model of a Fire Sprinkler Head?, Fire Safety Journal, Vol. 14, pp. 101-111, 1988.  
 
[18] Clanet, C., Villermaux, E., ?Life of a Smooth Liquid Sheet?, Journal of Fluid 
Mechanics, Vol. 462, pp. 307-340, 2002.  
 
[19] Villermaux, E., Clanet, C., ?Life of a Flapping Liquid Sheet?, Journal of Fluid 
Mechanics, Vol. 462, pp. 341-363, 2002.  
 
[20] Weber, Z., Angew Math Mech. Vol.11 pp 136-154, 1931. 
 
[21] Ostrach, A., Koestel, S., ?Film instabilities in two-phase flows?, AIChE Journal,
Vol 11, pp. 294-303, 1965.  
 
[22] Wu, D., Guillemin, D.,  Marshall, A.W., ?A Modeling Basis for Predicting the 
Initial Sprinkler Spray?, Fire Safety Journal, In Press, 2007. 
 
[23] Wu, D., ?Atomization Model for Fire Suppression Devices?, University of 
Maryland, College Park, MD, 2005. 
 
[24] Lefevre, A. H., ?Atomization and Sprays? Indiana: Taylor and Francis, 1989.  
 
[25] Crowe, C.T., Sommerfeld, M., and Tsuji, Y., ?Multiphase Flows with 
Droplets and Particles?,  CRC Press, 1998, ISBN 0-8493-9469-4. 
[26] Malvern Instruments/Insitec Inc., ?Technical Specifications: EPCS?, Issue 1.0, 
1997. 
 
37 
 
[27] Krarup, A.,?Spraytec Cookbook: RTSizer software Vers. 5.1 to 5.3?, Version 1.3, 
2002.  
 
[28]   Chan, T.S., ?Measurements of Water Density and Droplet Size Distributions of 
Selected ESFR Sprinklers?, Journal of Fire Protection Engineering, 6(2):79?87, 
1994. 
 
[29] Lawson J. R., Walton W. D., and Evans, D. D., ? Measurement of Droplet Size in 
Sprinkler Sprays?, National Bureau of Standards, U.S. Department of Commerce, 
NBSIR 88-3715, Gaithersburg, MD, Feb. 1988. 
 
[30] Jackman. L. A., Nolan P. F., Morgan H. P. ?Characterization of Water Drops from 
Sprinkler Sprays?, Fire Suppression Research ? First International Conference, 
Stockholm, Sweden, 1992. 
 
[31]  Chow W. K. Shek L. C., ?Physical Properties of a Sprinkler Water Spray?, Fire 
Mater, Vol. 17, pp. 279-292, 1993. 
 
[32] Chan T. S., ?Measurements of Water Density and Drop Size Distributions of 
Selected ESFR Sprinklers?, Journal of Fire Prot Engr, Vol. 6, pp. 79-87, 1994. 
 
[33] Reitz, R. D., and Bracco, F. V. ?Mechanism of Atomization of a Liquid Jet?, 
Physics of Fluids, Vol. 25, pp. 1730-1742, 1982. 
 
[34] Marshall, A. W. and Guillemin, D., ?An Analytical Model for Prediction Initial 
Spray Properties from Liquid Suppression Devices,? Workshop on Fire Suppression 
Technologies Proceedings, Mobile, AL, 2003. 
 
[35]  Heskestad, G., ?Proposal for Studying Interaction of Water Sprays with Plume in 
Sprinkler Optimization Program,? Memorandum to C. Yao, June 16, 1972. 
[36] Dombrowski, N. and Johns, W. R., ?The Aerodynamics Instability and 
Disintegration of Viscous Liquid Sheets?, Chemical Engineering Science, Vol.18, 
pp.203-214, 1963. 
 
[37] McGrattan, K., ?Fire Dynamics Simulator (Version 4) Technique Reference 
Guide?, NIST Special Publication 1018, 2004. 
 
[38] Watson, E. J., ?The Radial Spread of a Liquid Jet over a Horizontal Plane?, Journal 
of Fluid Mechanics, Vol. 20, pp.481-499, 1964. 
 
[39] Ibrahim, E. A., and Przekwas, A. J., ?Impinging Jets Atomization?, Physics of 
Fluids A, Vol. 3, pp. 2981, 1991.