ABSTRACT 
 
 
Title of Thesis: AN ALTERNATIVE TO HYDRAULIC FRACTURING:  
AN EXPLOSIVELY DRIVEN MECHANICAL DEVICE 
  
 David T. Horst, Master of Science, 2018 
  
Thesis Directed By: Professor William L. Fourney 
Department of Mechanical Engineering 
 
 
Hydraulic fracturing is currently used to extract natural gas from shale formations in the 
U.S, but there are major concerns regarding the environmental impact.  As an alternative 
method of well stimulation, a mechanical device driven by explosive loading was 
developed.  The goal of the device is to control the orientation of crack growth, in an 
unfavorable in situ stress field.  Small scale model testing was conducted in PMMA, 
simulating an unconventional reservoir.  A hydraulic press was used to simulate the in 
situ stress and high speed cameras were used to capture images of detonation, device 
function, and crack development.  Variations were made to device geometry and the 
resulting crack growth is visually compared.  Optimal device dimensions for fracture 
initiation are determined, and the effect of borehole characteristics on crack formation is 
discussed.  Scaling recommendations for testing the device at full scale are provided.  
 
 
 
 
AN ALTERNATIVE TO HYDRAULIC FRACTURING:  
AN EXPLOSIVELY DRIVEN MECHANICAL DEVICE 
 
by 
 
David T. Horst 
 
 
 
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 
2018 
 
 
 
 
 
 
 
 
 
    Advisory Committee: 
   Professor William L. Fourney, Chair 
   Professor Amr Baz  
   Professor Hugh Bruck 
 
 
Acknowledgements 
I would like to start to by thanking my advisor, Dr. William Fourney for his 
endless support and patience.  Throughout the many challenges I faced, I knew I could 
count on Dr. Fourney’s expertise and advice to help me along the way.  I will always be 
thankful for the opportunities he has given me and the lessons he has taught me. 
Next, I would like to thank the Center for Engineering Concepts Development 
(CECD) at UMD, for the financial support they provided to conduct this research.  
Without their backing, I would not have been able to pursue my research interests and 
benefit from the experiences I have gained during this project. 
Also, I would like to thank my committee members, Dr. Amr Baz and Dr. Hugh 
Bruck, for their valuable input and flexibility to help.  Their eagerness to help and 
contributions towards this thesis are highly appreciated. 
I would like to give a special thanks to Mike Perna from the UMD Aerospace 
Machine Shop.  His machining expertise and advice were crucial to my research.  He was 
always willing to take a break from his busy day to help me out and teach me something 
new.  I could not have done it without him. 
Last but not least, I would like to thank Dr. Uli Leiste and Leslie Taylor.  The two 
friendly faces that make the Dynamic Effects Laboratory the place that it is.  Over the 
years I have received a healthy balance of friendly banter and expert advice from Uli and 
Les.  Their presence is lab is irreplaceable and greatly improved my experience at UMD. 
  
 ii 
 
Table of Contents 
Acknowledgements .............................................................................................................. ii 
Table of Contents ................................................................................................................ iii 
List of Figures ...................................................................................................................... v 
Chapter 1: Introduction and Background.............................................................................1 
1.1 Introduction ............................................................................................................... 1 
1.2 Overview of the Current Hydraulic Fracturing Method............................................ 2 
1.3 Drawbacks of Current Method .................................................................................. 5 
1.4 Background ............................................................................................................... 6 
1.5 Developing an Alternative Method of Crack Initiation ............................................ 9 
1.6 Small-Scale Testing................................................................................................... 9 
Chapter 2: Overview of Device Design .............................................................................12 
Chapter 3: Materials and Methods .....................................................................................16 
3.1 Testing Equipment .................................................................................................. 16 
3.1.1 Explosive Equipment........................................................................................ 16 
3.1.2 Optical Equipment ............................................................................................ 20 
3.1.3 Loading Equipment .......................................................................................... 22 
3.2 Method of Model Assembly and Testing ................................................................ 24 
3.2.1 First Stage of Testing........................................................................................ 24 
3.2.2 Second Stage of Testing ................................................................................... 34 
3.2.3 Third Stage of Testing ...................................................................................... 36 
Chapter 4: Experimental Results .......................................................................................38 
4.1 Preliminary Testing and Design Iteration ............................................................... 38 
4.2 Establishing Planes of Reference ............................................................................ 41 
4.3 Resulting Crack Growth from Test 1 ...................................................................... 42 
4.3.1 Test 1, Stage 1 .................................................................................................. 42 
4.3.2 Test 1, Stage 2 .................................................................................................. 46 
4.3.3 Test 1, Stage 3 .................................................................................................. 51 
4.4 Model Repeatability ................................................................................................ 53 
4.4.1 Test 2, Stage 1 .................................................................................................. 53 
4.4.2 Test 2, Stage 2 .................................................................................................. 55 
 iii 
 
4.4.3 Test 2, Stage 3 .................................................................................................. 58 
4.4.4 Discussion of Repeatability .............................................................................. 60 
4.5 No-Load Condition ................................................................................................. 61 
4.5.1 Top View of Functioning Device ..................................................................... 62 
4.5.2 Test 3, Stage 1 .................................................................................................. 64 
4.5.3 Test 3, Stage 2 .................................................................................................. 65 
4.5.4 Test 3, Stage 3 .................................................................................................. 67 
4.5.5 No-Load Condition Discussion ........................................................................ 69 
4.6 Variations to Wedge Geometry ............................................................................... 70 
4.6.1 Test 4, Stage 1 .................................................................................................. 70 
4.6.2 Test 4, Stage 2 .................................................................................................. 74 
4.6.3 Test 4, Stage 3 .................................................................................................. 76 
4.6.4 Test 5, Stage 1 .................................................................................................. 78 
4.6.5 Test 5, Stage 2 .................................................................................................. 80 
4.6.7 Test 6, Stage 1 .................................................................................................. 81 
4.6.8 Test 6, Stage 2 .................................................................................................. 84 
4.6.9 Test 6, Stage 3 .................................................................................................. 86 
4.6.10 Discussion of Variations in Wedge Geometry ............................................... 88 
4.7 Sensitivity to borehole characteristics..................................................................... 89 
4.7.1 Borehole Wall Roughness ................................................................................ 90 
4.7.2 Borehole Discontinuities .................................................................................. 91 
4.8 Implications and Recommendations for Full-Size Testing ..................................... 92 
Chapter 5: Conclusions and Future Work..........................................................................95 
5.1 Conclusion............................................................................................................... 95 
5.2 Suggested Future Work ........................................................................................... 96 
Bibliography.......................................................................................................................98 
 
  
 iv 
 
List of Figures 
Figure 1: U.S natural gas production by source [2] ............................................................ 1 
Figure 2: Volumetric composition of a generic fracture fluid [3]  ...................................... 4 
Figure 3: Vertical section view of a hydraulically fractured horizontal well [7] ................ 4 
Figure 4: Exploded top view of CAD model for preliminary design ............................... 14 
Figure 5: An overview of a custom explosive charge....................................................... 17 
Figure 6: Completed custom explosive charge ................................................................. 18 
Figure 7: FS-10 Firing System.......................................................................................... 20 
Figure 8: Phantom v12.1 high speed camera .................................................................... 21 
Figure 9: LED array for analyzing crack development after loading  ............................... 22 
Figure 10: Hydraulic hand press used to simulate in situ stress ....................................... 23 
Figure 11: Hydraulic hand pump used for quasi-static loading ........................................ 24 
Figure 12: Engineering drawing of model block .............................................................. 25 
Figure 13: Engineering drawing of tapered wedge blades................................................ 27 
Figure 14: Engineering drawing of slotted aluminum rod ................................................ 28 
Figure 15: Aluminum rod and steel wedges subassembly ................................................ 29 
Figure 16: Engineering drawing of tapered steel pin ........................................................ 30 
Figure 17: Assembled device (pin, rod, and wedges) ....................................................... 31 
Figure 18: Engineering drawing of first stage plug .......................................................... 32 
Figure 19: Assembled model prepared for first stage of testing ....................................... 33 
Figure 20: Engineering drawing of second stage plug...................................................... 35 
Figure 21: Engineering drawing of third stage plug ......................................................... 37 
Figure 22: Unsuccessful Preliminary Test ........................................................................ 40 
Figure 23: CAD Drawing of Section Design Iteration ..................................................... 40 
Figure 24: Front view of a block prior to testing .............................................................. 41 
Figure 25: Top view of Test 1, Stage 1............................................................................. 43 
Figure 26: Top view, close up of horizontal cracks at bottom of borehole  ...................... 44 
Figure 27: Right view, detailing side profile and unwanted vertical cracks ..................... 45 
 v 
 
Figure 28: Bottom view of Test 1, Stage 2 ....................................................................... 49 
Figure 29: Back view of Test 1, Stage 2 ........................................................................... 50 
Figure 30: Top view of Test 1, Stage 3 results ................................................................. 51 
Figure 31: Back view of Test 1, Stage 3 results ............................................................... 52 
Figure 32: Bottom view of Test 2, Stage 1 horizontal cracks........................................... 54 
Figure 33: Right view of Test 2, Stage 1 detonation location .......................................... 55 
Figure 34: Top view of Test 2, Stage 2............................................................................. 57 
Figure 35: Back view of Test 2, Stage 2 ........................................................................... 58 
Figure 36: Bottom view of Test 2, Stage 3 ....................................................................... 59 
Figure 37: Front view of Test 2, Stage 3 detailing curvature  ........................................... 60 
Figure 38: High speed images of functioning device ....................................................... 63 
Figure 39: Top view of Test 3, Stage 1............................................................................. 65 
Figure 40: Top view of Test 3, Stage 2............................................................................. 67 
Figure 41: Front view of Test 3, Stage 3 .......................................................................... 68 
Figure 42: Top view of Test 4, Stage 1............................................................................. 72 
Figure 43: Back view of Test 4, Stage 1 ........................................................................... 73 
Figure 44: Right view close-up of detonation point, of Test 4, Stage 1 ........................... 73 
Figure 45: Top view of Test 4, Stage 2............................................................................. 75 
Figure 46: Back view of Test 4, Stage 2 ........................................................................... 76 
Figure 47: Top view of Test 4, Stage 3............................................................................. 77 
Figure 48: Front view of Test 4, Stage 3 .......................................................................... 78 
Figure 49: Top view of Test 5, Stage 1............................................................................. 79 
Figure 50: Left side view of Test 5, Stage 1 ..................................................................... 79 
Figure 51: Left side view of Test 5, Stage 2 ..................................................................... 80 
Figure 52: Right side view of Test 5, Stage 2 ................................................................... 81 
Figure 53: Bottom view of Test 6, Stage 1 ....................................................................... 83 
Figure 54: Left side view of Test 6, Stage 1 ..................................................................... 83 
Figure 55: Bottom view of Test 6, Stage 2 ....................................................................... 85 
Figure 56: Left side view of Test 6, Stage 2 ..................................................................... 86 
 vi 
 
Figure 57: Back view of Test 6, Stage 3 ........................................................................... 87 
Figure 58: Right side view of Test 6, Stage 3 ................................................................... 88 
Figure 59: Roughness Comparison ................................................................................... 91 
Figure 60: Engineering Drawing of Scaled Wedge .......................................................... 93 
Figure 61: Engineering Drawing of Scaled Pin ................................................................ 93 
Figure 62: Engineering Drawing of Scaled Rod ............................................................... 94 
 
 
 vii 
 
Chapter 1: Introduction and Background 
1.1 Introduction 
 In recent years, the role of natural gas has become increasing important for the 
United States energy sector.  In 2016, natural gas was the largest source of primary 
energy production accounting for 33% of U.S. domestic energy production [1].  The 
increase in production of natural gas is mostly due to the access of unconventional 
reservoirs.  The distribution of natural gas production by source can be seen in Figure 1 
[2].  
 
Figure 1: U.S natural gas production by source [2] 
The historical data is presented along with a projection of future performance, as of 2015.  
As can be seen in Figure 1, gas production from shale and other low permeability 
 1 
 
formations (tight gas) has seen a drastic increase in recent years.  The increasing role of 
gas production from shale reservoirs can be attributed to developments in drilling and 
fracturing technologies, allowing for economic potential in low permeability reservoirs 
[3].  Three important factors leading to the economic viability of shale gas production are 
advances in horizontal drilling, hydraulic fracturing, and increasing prices in natural gas 
[3].  Horizontal drilling creates wells with much greater reservoir exposure compared to 
that of vertical wells due to the lamina structure of shale [3][4] .  Hydraulic fracturing is a 
formation stimulation technique used to increase permeability in a producing formation, 
allowing gas to flow more readily to the wellbore [5][6]. 
                                 
1.2 Overview of the Current Hydraulic Fracturing Method 
 The process of hydraulically fracturing a shale formation begins after a suitable 
reservoir has been selected and a horizontal wellbore has been drilled.  Before any 
fracking can occur, a steel pipe is inserted into the vertical section of the wellbore to 
serve as a protective casing.  Additionally, a casing composed of either steel or concrete 
may be inserted into the horizontal section of the wellbore, depending on the structural 
integrity of the rock formation.  Cement is then used to secure the casing and bond it to 
the adjacent rock surface.  The casing provides structural support for the wellbore, 
preventing unwanted collapsing and acting as a protective barrier for the rock, separating 
the fracking fluid from entering protected zones such as aquifers [3][7] .  Due to the 
length of the exposed wellbore, it is usually impossible to effectively pressurize the entire 
borehole at once [8].  Instead, the majority of the well is plugged, isolating a small 
 2 
 
portion called a stage.  Stages are fractured sequentially until the entire well has been 
stimulated [9].  For each stage, small shaped charges called perforators are placed in a 
helical/screw pattern around the borehole wall [7].  These explosive charges are 
detonated next, in order to rough up the borehole and create small cracks that will later be 
repressurized to cause further crack growth.  Then, a liquid commonly known as “frack 
fluid” is inserted into the wellbore.  As seen below in Figure 2, frack fluid is primarily 
composed of water and a proppant (sand), with a small volumetric percentage being 
chemicals that are used to alter specific performance characteristics of the fluid.  The 
exact chemicals used will vary depending on the well operator, but some common uses 
are acid for dissolving minerals and expanding fractures, biocide for eliminating 
corrosion causing bacteria, and gelling agents to aid the suspension of the proppant [3].  
The proppant is an essential component of the frack fluid, holding open fractures and 
allowing gas to flow back into the wellbore.  After the frack fluid fills the borehole and 
existing cracks, it is pressurized with high power pumps from the surface.  The high 
pressure fluid is forced outwards, expanding fractures vertically into the shale and 
allowing access to natural gas.  After the cracks have arrested, the fluid is treated with 
chemicals to lower viscosity.  The fluid is then pumped back up to the surface leaving the 
proppant behind to hold open fractures.  When the fluid is removed, the wellbore 
becomes lower in pressure than the gas held within the shale formation and gas begins to 
flow back up to the surface for production.  An overview of a horizontal well after 
hydraulic fracture is presented in Figure 3 [7]. 
 3 
 
 
Figure 2: Volumetric composition of a generic fracture fluid [3] 
 
Figure 3: Vertical section view of a hydraulically fractured horizontal well [7] 
 4 
 
1.3 Drawbacks of Current Method 
 Although hydraulic fracturing has been used successfully from an economic 
perspective, there are concerns over the environmental costs of the method [13].  A study 
by D. Zhang et al. on the environmental impacts of hydraulic fracturing of shale 
formations found that there is potential to consume fresh water resources, contaminate 
surface and groundwater, induce earthquakes, and pollute air.  However, the water usage, 
as well as the intensity, are lower than conventional oil.  Additionally, the earthquakes 
induced are non-destructive micro-earthquakes below 2 in magnitude [12].  It is likely 
that the reported earthquakes induced by fracturing and were larger than 3 in magnitude 
were a result of existing faults prior to fracturing [14].  The air pollution caused by 
hydraulic fracturing comes from pollutant emissions and gas leakage during extraction 
and transportation [12].  Some estimations place the greenhouse gas footprint of shale gas 
being greater than conventional oil [13], while other estimations expect the emissions to 
be significantly lower [15].  The environmental impacts stated above are substantial, but 
comparable to conventional reservoir extraction methods.  The most distinctive and 
controversial environmental concern associated with hydraulic fracturing is water 
contamination.  Studies have shown possible contamination risks from both stray natural 
gas [16][17] and wastewater leakage [18]. 
The environmental impacts presented have led to regulatory and social 
repercussions for the practice of hydraulic fracturing.  Government agencies have 
restricted the practice and completely banned it in some states [10].  Other regions have 
seen large-scale rejection of the practice from angered citizens due to groundwater 
 5 
 
pollution.  D. Evensen et al. conducted a national survey on the support or opposition to 
both the terms “fracking” and “shale gas development.”  The results showed that 
nationally only 46% of the survey participants supported “fracking,” while 62% of the 
participants supported “shale gas development” [11].  The author suggests that these 
results are most likely influenced by the language and connotation associated with each 
phrase.  The media often uses fracking in a negative sense, while the media often uses 
developing the shale in a positive sense.  This survey data could also be interpreted to 
mean that the method of fracking has large social rejection, while the idea of accessing 
the shale in general is more acceptable.  Furthermore, this shows that citizens of the U.S. 
are not satisfied with the current hydraulic fracturing procedure and developing a new, 
more environmentally method of accessing unconventional reservoirs could be highly 
desirable. 
 
1.4 Background 
 The researchers at the Dynamic Effects Laboratory at the University of Maryland 
have previously had success modeling boreholes in small-scale fracture testing.  PMMA 
models have been tested using various well stimulation techniques.  It was found that the 
results can be used to qualitatively predict results in full-scale testing of nonhomogeneous 
material [19][20] .  The field conditions and stress states acting on the borehole dictate 
crack propagation and orientation.  Model testing has been conducted showing the effects 
of the in situ stress field on a horizontal wellbore.  Also, the use of machined flaws or 
notches as well as the influence on crack initiation and orientation due to variations in 
 6 
 
flaw geometry, numbers of flaws, and locations of flaws has been studied.  Notching 
techniques have been used to successfully control the orientation of crack growth [21].  
The significance of the pressure versus time loading profile has been explored and 
variations in pressure magnitude and duration have been determined to be an important 
parameter for quantity and quality of crack development [22][23] .  
 Additional work has been performed to investigate the method of fracture 
initiation and the influence of borehole conditions.  An immensely important and well 
documented concept for hydraulic fracturing is the in situ stress imposed by the 
overburden rock above a horizontal wellbore.  This stress field acts as a vertical 
compressive stress on the borehole, making it very difficult to grow cracks in the 
horizontal plane.  The in situ stress makes the direction of preferred fracture propagation 
parallel to the compressive stress and perpendicular to the least principle stress 
[24][25][26].   
Mechanisms of brittle fracture in rock has been studied and a fracture theory has 
been developed.  A relationship between crack velocity, driving energy, crack 
propagation, and crack roughness has also been established [24].  The principle of 
applying explosive energy in a borehole to initiate rock fracture has been studied and a 
method for finding borehole pressure generated from a detonation has been determined.  
The mechanism of how the resulting borehole pressure causes fracture growth is known 
[26].  The volume of reservoir accessible by a fracture network has been established as a 
metric for effectiveness.  Increased surface roughness is desirable as it increases the 
volume of rock accessed and aides in keeping the crack open without the use of proppant.  
 7 
 
Explosively driven cracks are driven at a higher speed and result in a higher crack 
roughness making them a superior source of crack propagation when compared to quasi-
static loading [21].  One concern of explosively loading a wellbore is the formation of a 
residue stress cage after detonation is complete.  Too much explosive can crush the 
immediate area around a borehole, where plastic effects cause the resulting fractures to 
close and resist opening to allow gas flowback.  Filling the borehole with a fluid such as 
water has been shown to be an effective method of minimizing unwanted crushing 
[23][26].  The relationship between fluid medium, coupling ratio, and borehole pressure 
generation has been studied [26]. 
Another source of unwanted crack development can be discontinuities in the 
borehole.  It has been found that circumferential cracks form at the ends of the borehole 
or at a change in diameter.  Stress wave theory has been investigated and applied to 
fracture initiation by explosives [26].  The crack development at discontinuities can be 
attributed to a doubling in pressure wave magnitude and duration of loading due to 
reflection at the discontinuities [23]. 
This paper strives to use the established principles of controlling fracture growth, 
and apply them to developing a device that can be inserted into a wellbore.  With the 
previous methods of fracture control, there is no practical way to reproduce results at full-
scale due to the challenges of manufacturing flaws and notches in deep horizontal wells.  
The goal of this device is to replicate the successful fracture control demonstrated with 
notching techniques, by implementing a new mechanism of fracture initiation.  The 
 8 
 
recognized fracture theory will be applied to design a device capable of producing 
desirable crack characteristics for natural gas extraction. 
 
1.5 Developing an Alternative Method of Crack Initiation 
 At the Dynamic Effects Laboratory at the University of Maryland, small-scale 
model testing is conducted to simulate hydraulic fracturing.  A mechanical device driven 
by explosives was designed to initiate cracks in wellbores.  The purpose of this device is 
to replace the need for chemical additives in frack fluid, reducing ground water pollution, 
and initiating cracks in any desired orientation within the shale formation.  The device 
was manufactured and inserted into PMMA blocks, a transparent and brittle material, 
allowing for monitoring of crack formation.  Variations in device design and geometries 
were investigated to find the optimal configuration.   An analysis was performed on crack 
growth with high speed camera footage and post-mortem photography.  A discussion of 
scaling the device to full-scale is presented and a final recommendation of device 
geometry for future testing is provided. 
 
1.6 Small-Scale Testing 
 The small-scale explosive testing performed at the Dynamic Effects Laboratory at 
the University of Maryland has advantages over large-scale testing.  The most obvious 
advantage of scaled testing is the amount of money that is required to conduct tests.  The 
reduction in scale of explosives and other materials used saves a considerable amount of 
funds.  Another benefit offered is that the time required for testing is much lower.  When 
 9 
 
developing a device or method, custom manufacturing can be very time consuming.  
Conducting these tests at small-scale means less time and materials are needed for 
manufacturing and drilling.   
 An important requirement of small-scale testing is that the results have some real 
meaning and implication for the real scenario.  The Dynamic Effects Lab and other 
organizations investigating fracture mechanics of rock have provided field verification 
showing that small-scale testing in PMMA blocks can provide data with significance for 
full-scale testing.  Specifically, the results found in small-scale testing provide qualitative 
information about how cracks will form from different techniques and loading scenarios 
[19][20].  With scaled testing it is important to look for general patterns of crack 
formation.  Exact repeatability and scaling is not to be expected.  This is due to small and 
randomly distributed cracks and flaws within materials that impact the development of 
the fracture system [24]. 
 A common method for scaling blast waves is cube root scaling.  The scaling law 
from Hopkinson’s scaling states that for any distance R, the peak pressure blast will be 
proportional to the cube root of the weight of explosive [27][28].  A common form used 
from these scaling laws is given by the following equation:  
1
𝑅 𝑤 3
= ( )  
𝑅1 𝑤1
In this equation, R is distance from the charge and w is the weight of the charge [29]. The 
subscript 1 denotes the new explosive.  An example of this scaling can be applied to a 
change in borehole diameter.  If an explosive charge of 250 mg is detonated in a 1 inch 
 10 
 
borehole, an equivalent charge mass for a 4 inch borehole would be 16 grams of 
explosive. 
 
  
 11 
 
Chapter 2: Overview of Device Design 
 As stated above, the Dynamic Effects Laboratory has had success controlling 
crack growth orientation under a large compressive load simulating in situ stress by using 
machined notches.  However, it is not practical to use this method in a horizontal well 
due to the challenges of manufacturing.  Instead, a device was designed to reproduce the 
results of notches while remaining compact and capable of being inserted.  The design 
requirements are that the device must be manufactured and assembled at the surface, and 
also be capable of being inserted into a horizontal borehole and then detonated to initiate 
fractures.   
Shale reservoirs are an anisotropic media with a laminar structure, generally 
oriented in the horizontal plane.  Natural gas is contained within these layers of the shale.  
For this reason, the horizontal plane will be the desired orientation of crack growth, as it 
will allow for a greater volume of natural gas access when compared to growth in the 
vertical direction.  Access to the reservoir in the vertical direction is limited by the 
thickness of the shale reservoir. Fracture growth in the horizontal direction has the 
potential to be extended a much greater distance, but this is not possible with current 
methods due to the in situ stress field acting on the formation.  The primary concern for 
horizontal crack growth is the vertical compressive stress imposed by the overburden 
rock.  A compressive stress in the vertical direction impedes horizontal crack growth and 
makes the vertical direction the preferred direction of fracture initiation.  The device 
 12 
 
developed must be capable of overcoming the unfavorable conditions imposed by this 
compressive stress, and initiate crack growth in the horizontal direction. 
The premise of this device is based around using hardened steel wedges with a 
tapered edge (3 degrees), which will be explosively driven into the side of the wellbore.  
A thin-walled aluminum rod with slots is used to hold and align the wedges.  A tapered 
steel pin with a matching degree of taper (3 degrees) to the wedges is inserted and 
epoxied into the rod and wedge subassembly.  The assembly of the three components are 
then inserted into a 15/16 inch diameter by 5.5 inch long borehole within a PMMA block.  
PMMA is chosen due to its transparency, allowing internal deformation to be monitored 
more easily.  Additionally, PMMA has been replicated in field testing, proving it serves 
as an effective qualitative model of anisotropic rock.  An explosive charge composed of 
PETN and lead azide is placed in close proximity to widest end of the steel pin and wires 
are run out of the block to an external firing box, and then a plug is inserted into the 
borehole and bonded with an adhesive.  Finally, the entire block is inserted into a 
hydraulic press where a significant load is applied perpendicular to the plane of desired 
cracking to model the in situ stress.  An exploded top view of a computer-aided design 
 13 
 
model for an assembled block can be seen below in Figure 4.
 
Figure 4: Exploded top view of CAD model for preliminary design 
To initiate the test, a high voltage is applied to the firing cables causing a 
detonation of the charge.  The detonation causes pressure to build up within the borehole, 
driving the pin down through the wedges.  Due to the taper on the wedges, they are 
driven outward into the wellbore wall and embed themselves within it.  This deformation, 
combined with the pressure build up in the borehole, initiates a crack in the orientation of 
the blades.  This method successfully overcomes the preferred direction of cracking 
caused by the applied compressive stress and allows controlled crack orientation.   
To further expand the initial fractures, a second stage of testing is performed on 
the same block.  The plug and device are removed from the borehole and the borehole is 
drilled fresh again to ensure no unwanted material is left behind.  A second smaller 
 14 
 
charge is inserted and the borehole is filled with water.  The borehole is plugged again 
and reinserted into the hydraulic press under the same compressive load.  The charge is 
detonated and the initial fractures are expanded farther horizontally.   
A third and final stage of testing is conducted.  Now that the dynamic fracturing 
potential has been performed via explosive loading, the block is connected to a hydraulic 
pump to further expand the cracks by quasi-static loading.  A steel plug with NPT threads 
is bonded into the borehole, the borehole is filled with hydraulic oil, and the block is 
reinserted into the hydraulic press.  The hydraulic pump applies a slow incremental 
pressure increase until the fracture grows.  One of the cracks initiated by explosive 
loading (the most well defined crack) will continue to grow until it reaches an outer 
surface of the block and loses pressure.  
 It is worth noting the reasoning and implications of using hydraulic oil as a 
pressurization medium.  The first and most obvious reason is that it was an inexpensive 
fluid compatible for use with the hydraulic hand pump.  However, the high viscosity of 
the hydraulic oil is also advantageous.  The increase in viscosity helps pressurize major 
flaws, while avoiding the fluid from pressurizing and extending minor flaws.  In a porous 
material such as rock, a high viscosity fluid will avoid leakage and flow into unwanted 
areas.  Furthermore, most of the current fluids used in hydraulic fracturing at full-scale 
are very viscous, so using a viscous hydraulic oil will be an effective model for small-
scale testing.  
 15 
 
Chapter 3: Materials and Methods 
3.1 Testing Equipment 
 In order to conduct small-scale explosive testing, there are equipment 
requirements and test preparation is needed.  Specifically, there is explosive equipment 
needed for conducting detonations, optical equipment needed for recording and observing 
models, and preparation equipment needed to assemble models and modify their 
characteristics.  A detailed discussion of the typed of equipment needed and the methods 
of test preparation are provided. 
 
3.1.1 Explosive Equipment 
 The most obvious and fundamental equipment for explosive testing are the 
charges themselves.  For initial testing, premanufactured charges from a large-scale 
distributor were purchased.  However, an iterative design process was used for early 
testing in order to quickly narrow down device geometry and charge sizing.  During this 
period, it was determined that using fixed size charges purchased in bulk from retailers 
were not economically feasible.  Instead, creating custom charges was preferred.  This 
option offers flexibility in charge sizes between tests and the ability to manipulate charge 
mass with ease.  An overview of an unassembled custom charge can be seen below in 
Figure 5. 
 16 
 
 
Figure 5: An overview of a custom explosive charge 
 Manufacturing an explosive charge begins with two 18 AWG aluminum wires.  
One end of each wire is stripped and fastened together to form a single length of wire.  
This is a safety precaution taken to avoid an unintentional voltage difference from static 
build up when in contact with sensitive initiating explosive.  The other ends of the wires 
are stripped to expose roughly 3 mm of the inner aluminum core.  The exposed lengths of 
wire are separated by a small gap, about a millimeter in width.  A heat shrink tube with a 
diameter of 5 mm (0.196 in) is placed around the exposed sections and a heat gun is 
applied.  The heat shrink tube secures the wires together while holding the gap at a fixed 
distance.  The heat shrink tube now has the structure of a small firm cylinder which will 
 17 
 
house the initiating explosive, lead azide.  10 milligrams of lead azide is carefully packed 
into the heat shrink, filling the gap between the two aluminum wires.  This assembly 
functions as a detonator for the charge.  A Delrin tube with an OD of 9.52 mm (0.375 in) 
and an ID of 7.39 mm (0.291 in) serves as a casing for the charge.  The Delrin casing is 
epoxied to a sheet of paper, closing one end of the tube.  The other end of the casing is 
filled with the desired amount of powdered PETN.  The end of the aluminum wires with 
heat shrink tubing and lead azide are inserted into the casing and embedded in the PETN 
powder.  The rest of the casing is filled with epoxy and the adjacent wires are coated in 
epoxy as well.  This provides a waterproof seal and structure integrity for the assembled 
charge.  Figure 6 shows a completed charge. 
 
Figure 6: Completed custom explosive charge 
 18 
 
 To initiate the custom charge, the free ends of the wires are unfastened and a high 
voltage difference is applied to them.  This causes an arc to occur across the small gap 
located at the opposite end, inside the charge.  The arc causes a spark which initiates a 
detonation in the sensitive lead azide.  A chain reaction occurs, setting off the PETN 
powder and detonating the remainder of the charge. 
 In order to provide a high voltage in a controlled and safe manner, a firing system 
is necessary.  For the purpose of these tests, the FS-10 control unit was obtained from 
Reynolds Industries Systems Incorporated.  The firing system is capable of supplying 
4,000 volt energy source to the explosives, creating a significant arc and spark.  An 
additional feature of the firing system, is its capability to simultaneously send a trigger 
signal to other electrical equipment upon firing.  This is a very useful tool for timing the 
high speed camera that is used to record images after detonation.  The FS-10 firing 
system can be seen in Figure 7.  
 19 
 
 
Figure 7: FS-10 Firing System 
3.1.2 Optical Equipment 
 To observe the function of the device and the resulting growth of cracks, a 
Phantom v12.1 high speed camera was used.  The camera was equipped with a Nikon 
Nikkor 50 mm lens for optimal image quality.  The Phantom camera was placed upon an 
adjustable tripod that allowed for fine tuning of position.  Figure 8 shows the camera with 
the lens attached, and mounted on the adjustable tripod.  The camera sensor allows for a 
maximum resolution of 1280 x 800 pixels.  However, in order to obtain higher frame 
rates, the resolution is often reduced.  Due to the high speed nature of explosive and 
fracture development, the frame rate most commonly used for testing was 91,056 frames 
per second.  This framerate allows for an image to be recorded every 11 microseconds.  
The resolution of the images recorded were 256 x 184 pixels.  With the chosen frame rate 
 20 
 
and image size, the camera is then positioned to have the area of interest within the field 
of view.  Then, the attached lens can be focused by hand to find the highest clarity image.  
An Ethernet cable is connected between the camera and a computer allowing for remove 
control and managing various settings.  Also, a cable is attached between the firing 
system and the camera to provide a trigger signal upon detonation. 
 
Figure 8: Phantom v12.1 high speed camera 
 21 
 
 One challenge with high speed photography is generating enough light to 
illuminate the area of interest and reach the sensor.  At high frame rates the sensor is not 
exposed to light for very long, so the images appear darker.  To account for this, 
additional lighting is provided by two incandescent 250 W lamps.  Two lights were 
positioned to shine light at the model within the hydraulic press.  This proved to be an 
effective way of illuminating cracks as they formed and grew during testing.  Additional 
lighting is also useful for inspecting and analyzing models after conducting tests.  A 
combination of the incandescent lights and a custom LED array were used to illuminate 
the cracks.  The custom LED array is shown in Figure 9. 
 
Figure 9: LED array for analyzing crack development after loading 
 
3.1.3 Loading Equipment 
 An essential component for this model testing was a hand-operated hydraulic 
press.  This piece of equipment was responsible for simulating the in situ stress field 
generated by overburden rock.  The press is pictured in Figure 10.  The assembled model 
 22 
 
PMMA block is inserted with the borehole oriented horizontally and parallel to the 
loading platens.  A steel plate is placed between the loading platens and both sides of the 
block, to evenly distribute the applied load.  It features a hand lever, using mechanical 
advantage to generate incremental increases in load.  A gauge measuring force is 
mounted at the top of the press to monitor the current load.  Once a desired load it 
reached, the press will maintain it until a release value is opened.  The maximum load 
capacity of the press is 88.9 kN (20,000 lbf), but in the interest of the longevity of the 
press, the load was restricted to 80 kN (18,000 lbf) during testing. 
 
Figure 10: Hydraulic hand press used to simulate in situ stress 
 As mentioned in the discussion of the preliminary design and testing process, the 
third pressurization of the model borehole is performed using quasi-static loading.  This 
 23 
 
loading is applied by the hydraulic hand pump shown in Figure 11.  The pump features a 
hand lever, giving the user mechanical advantage.  There is also an attached pressure 
gauge that provides readings of the current pressure being applied.  The output of the 
pump is through a 3/8 inch NPT thread.  A tapered thread is an effective method for 
maintaining a seal for high pressure applications.  The threads are wrapped in Teflon pipe 
sealing tape and inserted into a threaded steel plug bonded within the model block. 
  
 
Figure 11: Hydraulic hand pump used for quasi-static loading 
 
3.2 Method of Model Assembly and Testing 
3.2.1 First Stage of Testing 
 The process of preparing for a test begins with a PMMA block machined to the 
desired dimensions.  The machined surfaces of the block are then sanded with 600 grit 
water proof sand paper while wetted.  The surfaces must then be polished to remove any 
remaining surface flaws.   The result is a smooth and transparent model that allows for 
 24 
 
easy monitoring of internal deformation and crack growth.  The final dimensions of the 
model block can be seen in Figure 12.  Note that that the displayed units are in inches.  
The model has a cross section of approximately 102 mm (4 in) by 152 mm (6 in).  The 
length of the model is approximately 203 mm (8 in).  A borehole with a diameter of 23.8 
mm (15/16 in) is machined at the center of the cross section, extending 140 mm (5.5 in) 
in total length.  This leaves 63.5 mm (2.5 in) between the bottom tip of the borehole and 
the bottom surface of the block.  From preliminary testing, it was determined that this 
spacing was necessary to ensure that the bottom portion of the block does not blow out 
when the borehole is loaded.  There is some variation in borehole roughness throughout 
the manufactured blocks.  This is due to deterioration of the condition of the drill bit over 
time.  The effect of borehole roughness on crack growth is analyzed and discussed. 
 
Figure 12: Engineering drawing of model block  
 25 
 
The next step of assembling a block for testing is putting together the aluminum 
rod and steel tapered wedge subassembly.  The wedges are a critical component of the 
device since they are directly responsible for causing deformation and causing crack 
growth in the borehole wall.  The wedge geometry was changed between tests to see how 
the variations affect the success of crack formation.  A general schematic for the tapered 
wedge component is presented in Figure 13, the listed dimensions are inches.  An 
important dimension for assembly and alignment is the 3.2 mm (1/8 in) thickness of the 
tapered end.  This dimension remained constant throughout all tests.  Similarly, the width 
of the two ends remained constant at 8.1 mm (0.32 in) and 9.4 mm (0.37 in) for all tests.  
This dimension determined the gap between the wedges and borehole wall when the 
model was assembled.  The goal of this was to have a small consistent gap size that 
allowed for the wedges to be inserted without damaging the borehole prior to testing, 
while also remaining relatively snug within the borehole.  If the gap is too large, the 
blades have a tendency to misalign during test preparation and can deviate from the 
desired orientation.  The angle of the pointed side of the wedge was varied to observe its 
influence on crack growth.  The angle was tested at 10°, 20°, and 30°, and the resulting 
crack formation was studied.  The mass of each wedge is 1.58 g, 2.8 g, and 3.39 g 
respectively.  Additionally, the length of the wedge was varied to study its impact.  One 
test was conducted at half the normal length of 25.4 mm (1 in) with an angle of 30° and a 
mass of 1.61 g. 
 26 
 
 
Figure 13: Engineering drawing of tapered wedge blades  
 In order to provide a controlled method of alignment, the wedges are assembled 
into a slotted rod.  The aluminum rod can be seen in Figure 14, dimensioned in inches.  
As noted above, the tapered wedges have a constant thickness of 3.2 mm (1/8 in).  The 
rod has a corresponding slot with a gap size of 3.2 mm (1/8 in).  If machined correctly, 
the slot and wedge will fit with an interference fit.  Occasionally, adjustments are 
necessary and a thin steel file is used to remove burrs or open the gap by a small amount.  
Once the wedge can be inserted into the rod’s slot, it is aligned so the distance from the 
outer tip of each wedge’s blade is 23.4 mm (0.92 in).  This is an appropriate distance for 
providing a small gap between the wedges and borehole wall when inserted.  After the 
wedges are aligned, quick set epoxy is applied.  The epoxy does not serve any structural 
purpose during testing, but is helpful in maintaining the desired alignment while 
 27 
 
assembling the remainder of the model.  The completed subassembly composed of the 
aluminum rod and steel tapered wedges can be seen in Figure 15. 
 
Figure 14: Engineering drawing of slotted aluminum rod  
 28 
 
 
Figure 15: Aluminum rod and steel wedges subassembly  
 The tapered steel pin is the next component to be assembled.  A dimensioned 
drawing of the pin with units of inches, can be seen in Figure 16.  The pin features a 3° 
taper on its side to match the tapered edge of the wedges.  This is helpful for alignment 
purposes, but is also responsible for driving the wedges outward into the borehole wall 
upon detonation.    For assembly, the pin is inserted into the rod and wedges 
subassembly.  The pin’s tapered sides rest upon the tapered ends of the wedges which 
centers the pin within the rod.  The small end of the pin reaches the midpoint of the 
wedges when assembled, which is 12.7 mm (0.5 in) from the top surface of the aluminum 
rod.  Similarly to the assembly of the wedges, once the pin is in the desired position it is 
secured with quick set epoxy.  The assembled device containing the pin, rod, and wedges 
is shown in Figure 17.  Another important dimension of the tapered pin is the 23.4 mm 
(59/64 in) diameter top surface.  This dimension was chosen to be sized just under the 
 29 
 
diameter of the borehole, allowing for a tight fit without impeding assembly.  When 
detonation occurs, this tight fit will restrict the rapidly expanding gases from the charge, 
generating pressure on the pin’s top surface.  The pressure drives the pin which initiates 
the movement of the wedges and the resulting crack growth.  The last noteworthy feature 
of the pin is the 1/4 -20 threads machined in the top surface.  The threads serve as an 
alignment tool once the device is inserted into the borehole.  When the device is inserted, 
a screw can be lightly threaded into the top of the pin and used to fine tune the alignment 
of the wedges.   
 
Figure 16: Engineering drawing of tapered steel pin  
 
 30 
 
 
Figure 17: Assembled device (pin, rod, and wedges)  
 The final step of assembly is inserting a charge and plugging the borehole to hold 
pressure upon detonation.  The charge size used for this stage of testing is 250 mg of 
PETN and 10 mg of lead azide as an initiating explosive.  The plug material used is 
PMMA for earlier stages of testing or Polycarbonate (PC) for later tests.  The material 
type is not critical; the switch to PC was made solely for easier manufacturing.  The 
geometry of the first stage plug remained constant throughout all tests and can be seen in 
Figure 18.  The 7.9 mm (5/16 in) through hole is for running wires from the explosive to 
the external firing system.  The larger diameter hole at the bottom of the plug is used to 
center the explosive charge.  The outer surface of the plug has a diameter of 23.75 mm 
(0.935 in), providing a very close fit to the borehole wall.  The outer surface is roughed 
with coarse sand paper and cleaned with a degreaser safe for plastic use, such as 
isopropanol alcohol.  The same surface treatment is applied to the borehole wall where 
the plug will be in contact.  This technique increases the bonding strength by increasing 
 31 
 
the surface area and removing trace amounts of oil from machining, which can 
potentially interfere with bonding.  The adhesive used for bonding is 3M Scotch-weld 
instant adhesive, model CA8.  This is a great choice for high pressure sealing of plastics. 
The 3M adhesive acts more similarly to a plastic weld than a traditional glue.  The 
process of applying begins with first coating each surface with the adhesive via paint 
brush.  This method allows the adhesive to seep into the roughed up surfaces and improve 
bonding.  After, the plug is coated in the adhesive and inserted into the borehole.  
Although the adhesive sets quickly, it was given 12 hours to gain strength prior to testing.  
Additionally, the 7.9 mm (5/16 in) hole allowing the wires to run through the plug is 
sealed with epoxy at the top surface.  A completely assembled and ready to test model is 
shown in Figure 19. 
 
Figure 18: Engineering drawing of first stage plug  
 32 
 
 
Figure 19: Assembled model prepared for first stage of testing  
 
 33 
 
 After the bonding adhesive has had sufficient time to cure, the model is ready to 
be tested.  The assembled block is inserted into the hydraulic press with the borehole 
oriented in the horizontal plane.  In this orientation, the press is loaded and 80 kN (18,000 
lbf) of compressive force is applied in the vertical direction, perpendicular to the 
horizontal plane.  The charge is detonated and the results are analyzed.   
 
3.2.2 Second Stage of Testing 
 After the results of the first stage are measured and recorded, a second stage of 
testing is conducted to expand the initial fractures.  In order to prepare for the second 
stage of testing, the plug and device from the first stage must be removed.  The plug 
occasionally dislodged from the borehole when testing, but typically was drilled out.  
Significant changes in borehole size from the first stage does not occur, but to remove 
unwanted adhesive or surface abnormalities, the borehole is drilled again.  After removal 
of the plug, the rod, pin, and wedges are all removed from the borehole.  A new charge 
with 100 mg PETN and 10 mg lead azide is then prepared.  The charge is inserted in the 
borehole, positioned around the horizontal fractures initiated in the first stage of testing.  
The 3M scotch-weld is painted on the inside of the borehole to seep into the prepared 
surface and cover any unwanted cracks.  The borehole is then plugged with a new, longer 
PMMA or PC plug.  The additional length is useful in providing additional strength and 
plugging any unwanted cracks that developed at the bottom of the first stage plug.  Aside 
from the length, the new plug is dimensioned and prepared in the same manner as the 
first stage plug.  An engineering drawing for the second stage plug is shown in Figure 20.  
 34 
 
The dimensions shown are in inches.  The surface preparation method is repeated, 
sanding and degreasing the bonding surfaces of the plug and the borehole.  The 3M 
Scotch-weld instant adhesive is applied and given at least 12 hours to cure.  After the 3M 
scotch-weld has set, the borehole is filled with water through the hole in the plug and 
then sealed with epoxy.  The water serves to efficiently pressurize the borehole while 
minimizing unwanted crushing.  The model block is then reinserted into the hydraulic 
press with the same orientation for a second detonation.  The constant load of 80 kN 
(18,000 lbf) is applied vertically.  The charge is detonated and the results from the second 
stage are analyzed and recorded.   
 
Figure 20: Engineering drawing of second stage plug 
 35 
 
3.2.3 Third Stage of Testing  
 To prepare for the third and final stage of testing, the second stage plug is 
removed and the borehole is drilled fresh again.  This time, a steel plug with internal 3/8 
inch NPT threads is used.  A schematic of the plug can be seen in Figure 21, the 
dimensions provided are in inches.  The use of steel as the material for this component is 
important, as plastic threads are prone to leakage and failure for this high pressure 
application.  The plug’s bonding surface is prepared by etching with a steel file and 
degreasing with a cleaner.  The borehole wall is prepared by using sand paper and 
isopropanol alcohol, in the same manner as the previous stages.  Once again, the 3M 
scotch-weld is applied to the inside of the borehole and the plug is inserted.  After the 
plug has set, the borehole is filled with hydraulic oil, identical to the type used in a 
hydraulic pump.  Then, Teflon tape is applied to the matching male 3/8 NPT threads of 
the hydraulic hand pump.  The threads are inserted into the steel plug protruding out of 
the assembled model block and secured with a wrench.  The block is then reinserted onto 
the loading plate of the hydraulic press and the constant load of 80 kN (18,000 lbf) is 
reapplied.  The Phantom v12.1 high speed camera is positioned to record both crack 
growth and the pressure gauge located on the hydraulic pump.  This allows for a pressure 
reading at the time of failure.  The hydraulic hand pump is then loaded in a quasi-static 
manner.  Load is slowly applied in small increments until a crack is extended to a free 
surface, losing pressure within the borehole. 
 36 
 
 
Figure 21: Engineering drawing of third stage plug 
  
 37 
 
Chapter 4: Experimental Results 
4.1 Preliminary Testing and Design Iteration 
 In order to develop with the current functioning device, some basic preliminary 
testing was conducted.  These tests used different mechanisms for attempting to mimic 
the results of notched boreholes and did not have as much success controlling crack 
initiation.  Although the mechanisms were not directly relatable, the results gave insight 
into sizing of explosives, sizing of borehole diameter, method of sealing, and selecting 
the appropriate detonation medium for each testing stage.   
 The first tests done started with a deformation mechanism of creating flaws in 
ductile rods, inserting explosives, and detonating.  Hoping that the flaws could control the 
deformation of the rods, resulting in an impact on the borehole wall and crack growth.  In 
order for this method to work, the rod needed to be a similar diameter to the borehole and 
an excessive coupling ratio was achieved.  A picture of the resulting crack formation is 
shown in Figure 22.  Learning from the mistakes of this mechanism, the borehole 
diameter was increased.  A new mechanism was attempted, placing steel blades inside an 
alignment rod and driving them outward with explosives detonated from within the rod.  
This mechanism initially used O-rings as a method for containing borehole pressure, but 
this was unsuccessful.  The next iteration in design involved incorporating threaded plugs 
at the top of the borehole.  A CAD drawing of this is shown in Figure 23.  Note that the 
borehole diameter was increased again (to the current size of 15/16 in) to reduce 
unwanted cracking and to provide larger threads for better structural support.  This design 
iteration proved to be an improvement, but still had problems with thread failure and 
 38 
 
consistent ejecting of blades. It was at this point that reinforcing pins introduced as an 
aide of restraining the plug within the block.  The pins were inserted through the 
thickness of the block (102 mm or 4 in) and into a groove machined circumferentially 
through half the plug thickness.  Additionally, the wedge and pin design of the current 
device, shown in Figure 13 and Figure 16 were implemented.  These changes showed 
improvement, however, with the successful containment of the pressure there was an 
increase in crushing and the blades still failed to consistently eject into the borehole wall.  
It was determined that using a water filled borehole allowed the pressure to transmit 
around the pin at detonation, pressurizing the borehole to failure instead of ejecting the 
wedges.  Subsequent tests used air filled boreholes which helped in driving down the pin 
and ejecting the wedges.  The only problem remaining with the device was the stress 
concentration due to the restraining pins.  The holes drilled through the block that the 
pins were inserted into were a weak point and caused unwanted fractures to grow.  The 
pins and threads were removed from the design and adhesive secured plugs were 
introduced, leading to the current device being tested. 
 39 
 
 
Figure 22: Unsuccessful Preliminary Test 
 
Figure 23: CAD Drawing of Section Design Iteration 
 
 
 40 
 
4.2 Establishing Planes of Reference 
Prior to presenting results, it is convenient to introduce a coordinate system to 
make referencing orientation simpler.  In Figure 24, a photograph of a block is presented 
looking in the normal direction to the front surface.  Each edge is labeled with the name 
of the block side it represents. 
 
Figure 24: Front view of a block prior to testing 
  
 Note that we are looking onto the 102 mm x 152 mm (4 in x 6 in) cross section 
and that the borehole is machined into the front plane.  The surface 203 mm (8 in) in 
distance, but parallel to the front surface is referred to as the back surface, denoted by the 
blue square and white text.  When the device is inserted into the borehole, the wedges are 
 41 
 
oriented in the left and right direction.  The block is placed in the hydraulic press and 
pressurized in the top and bottom direction, which is parallel to gravity.  This means the 
left and right directions represent the horizontal plane in the model.  Crack growth left or 
right may be referred to as horizontal.  While growth in the top and bottom directions will 
be referred to as vertical.  Growth towards the front and back will be referred to as 
longitudinal. 
 
4.3 Resulting Crack Growth from Test 1 
4.3.1 Test 1, Stage 1 
 Test 1 was conducted with the standard component geometry presented in the 
material and methods section.  This included a 102 mm x 152 mm x 203 mm (4 in x 6 in 
x 8in) PMMA block and tapered wedges that feature blade angles of 30° and a length of 
25.4 mm (1 in).  The resulting crack growth from the first stage of testing is shown from 
a top view in Figure 25.  This picture provides a good overview of the developed cracks, 
displaying both desired cracks in the horizontal direction and other undesired cracks.  
 42 
 
 
Figure 25: Top view of Test 1, Stage 1 
 43 
 
 
Figure 26: Top view, close up of horizontal cracks at bottom of borehole  
Figure 26 shows a close up of the bottom of the borehole, offering a detailed view 
of the horizontal crack development.  The rectangular crack extending from the borehole 
wall is a direct result from the wedges ejecting and embedding within the borehole wall.  
The initial cracks caused by this deformation are extended both horizontally and 
longitudinally by the pressure within the borehole.  The left side crack reaches a 
maximum distance from the borehole wall of 9.7 mm (0.38 in).  The longitudinal length 
is 36.8 mm (1.45 in).  The horizontal crack on the right side of the borehole features a 
slightly smaller peak of 8.3 mm (0.325 in), but has a similar length of 36.3 mm (1.43 in).  
 44 
 
The right side crack is centered about the embedded wedge, while the left side crack 
favored growth towards the front plane.  An important characteristic of both cracks are 
the roughness they have developed.  Observed in Figure 26, the indicators of roughness 
are noted.  In addition, the cracks feature a ripple pattern extending outwards horizontally 
which is also indicative of a high energy and high roughness crack.  As previously stated, 
the roughness is an indicator of a high driving energy and high crack speed, a 
characteristic of explosively driven cracks. This is a desirable crack characteristic and 
will aide in extraction from shale [21]. 
 
Figure 27: Right view, detailing side profile and unwanted vertical cracks 
 45 
 
For analyzing vertical crack development and localized cracking around the 
detonation point, a side view is presented in Figure 27.  The bottom of the plug is located 
44.5 mm (1.75 in) from the front surface.  The charge is detonated immediately below 
this point.  The area in close vicinity to the charge experiences a high pressure during 
detonation.  This effect combined with the shockwave reflection at the plug bottom 
causes unwanted cracking in this area.  The majority of the cracking is a high number of 
short, ill-defined cracks extending less than 5 mm (0.2 in) outward.  However, there are 
two fractures in the vertical direction that are undesirable and noted.  The largest vertical 
crack extends 15 mm (0.6 in) outward and has a longitudinal length of 19.6 mm (0.77 in).  
This crack is substantial, but is not problematic.  In stage two it will be completely 
covered by the plug preventing any future growth.  It does however give insight into the 
effect of the in situ stress.  The fact that there is zero significant horizontal fractures 
around the point of detonation, but two vertical ones, clearly indicates the effect the 
compressive stress has on crack orientation. 
 
4.3.2 Test 1, Stage 2 
 A bottom view displaying the horizontal plane can be seen in Figure 28.  Note 
that this is the bottom view instead of the top view from stage 1, the right and left side are 
reversed.  Recall that stage 2 is performed with a fluid filled borehole and a smaller 
charge mass.  The use of water increased coupling, generating a higher ratio of borehole 
pressure to expand the existing horizontal fractures [26].  The horizontal cracks resulting 
from stage 1 are greatly expanded in both the longitudinal and horizontal directions. The 
 46 
 
left and ride side cracks each span the entire effective borehole length (63.5 mm or 2.5 
in), reaching from the bottom of the stage 2 plug, to the bottom of the borehole.  The 
right side crack shows the most significant growth in the horizontal direction.  It expands 
outward symmetrically, reaching a peak distance from the borehole wall of 28.7 mm 
(1.13 in). The left side experienced a similar magnitude of crack propagation.  The 
maximum peak reached 27.7 mm (1.09 in) from the borehole wall.  The outline of the 
cracks developed in stage 1 are identified in Figure 28, so the relative size of the cracks 
can be compared.  The results show that reapplying pressure to the borehole is an 
effective way to enhance crack growth in the desired orientation.  It is also important to 
analyze the effect of stage 2 on the undesired crack growth. 
 The plug inserted and bonded to the front most 76 mm (3 in) of the borehole was 
effective at preventing additional growth of the undesired cracks developed in stage 1.  
The two significant cracks developed in the vertical direction did not experience any 
further growth.  Additionally, the cracks that resulted from local effects in stage 1 are not 
pressurized to a noteworthy degree during stage 2.  This is important, because if the 
vertical cracks are allowed to grow to a significant degree and reach the outer surface, the 
horizontal cracks will not be able to be pressurized and grow further.  In the second stage 
of testing, additional non-horizontal cracks developed.  Figure 29 shows a view from the 
backside of the block, detailing radial fracture growth around the borehole.  The two 
largest cracks in the horizontal direction can be used as a scale to analyze the growth of 
other cracks.  In the vertical direction, there are cracks developed that reach slightly 
below 12.7 mm (0.5 in) in length.  For the purpose of model testing, it is desired to 
minimize this growth to show directional control and ensure that the vertical cracks do 
 47 
 
not interfere with further expansion of the cracks during stage 3.  However, when testing 
in the field, minor cracking in the vertical direction would also contribute to shale access 
and would not be considered an undesired result.   
 The results from stage 2 also highlight the effects of discontinuities in the 
borehole.  Two significant circumferential cracks have developed around the borehole.  
One originating at the bottom of the plug, and one at the bottom of the borehole, where 
there is a change in diameter due to drilling.  As previously discussed, this is due to the 
reflection of the pressure wave at the boundaries.  It is a noteworthy result, but not 
problematic for controlling fracture orientation. 
 48 
 
 
Figure 28: Bottom view of Test 1, Stage 2 
 
 49 
 
 
Figure 29: Back view of Test 1, Stage 2 
 
 50 
 
4.3.3 Test 1, Stage 3
 
Figure 30: Top view of Test 1, Stage 3 results 
 The third and final stage results are presented in Figure 30.  When the quasi-static 
loading reached 9 MPa (1,300 psi), the stress in the right horizontal crack reached a 
critical value and crack propagation began.  This shows the well-defined horizontal crack 
beat out the many ill-defined cracks in the vertical direction. The horizontal crack 
expanded outward until it reached the outside right surface of the block, losing pressure.  
The resulting crack expanded significantly in both the longitudinal and horizontal 
directions.  The longitudinal expansion reached the bottom circumferential crack before 
stopping, and extended up the borehole reaching 27.9 mm (1.1 in) from the front surface 
of the block.  The longitudinal distance peak to peak is 118.5 mm (4.67 in).  After 
pressure is removed from the borehole, the outer edges of the statically driven cracks 
close up and are very thin.  This presents challenges for viewing in pictures.  The outside 
 51 
 
profile of the crack is marked with red arrows on the block.  The smoothness of the 
statically driven cracks can also be observed in Figure 30.  The cracks resulting from 
stage 3 testing do not have the ripples or other small branch shaped cracks that indicate 
roughness.  They appear to be thin and mostly transparent, as they do not alter the 
structure and optical qualities of the block as much as rough cracks do.  
 
Figure 31: Back view of Test 1, Stage 3 results 
 The results from stage 3 also showcase the effects of the in situ stress on crack 
growth.  In Figure 31, the curvature of the statically driven crack can be noted when 
comparing its orientation to that of the horizontal, dynamically driven crack.  The stress 
concentration from the device in stage one, combined with the high pressure from 
 52 
 
dynamic loading dominates the crack behavior in stage one and two.  However, for stage 
3 the energy and pressure of quasi-static loading is lower and making the in situ stress a 
more dominate factor as the fracture propagates away from the dynamically initiated 
cracks.  
 
4.4 Model Repeatability 
 With test 1 showing great success at producing controlled crack growth in the 
desired orientation, proving these results could be repeatable is desired.  Test 2 is 
conducted under the same process and the results are compared to test 1 to analyze model 
consistency. 
 
4.4.1 Test 2, Stage 1 
 The resulting horizontal cracks from test 2 viewed from the bottom surface are 
shown in Figure 32.  The horizontal crack development is very similar to that seen in test 
1. The cracks begin with the embedded blades, with further crack growth extending 
outward from the borehole wall.  The right side features a crack with a peak of 8.5 mm 
(0.334 in) and a longitudinal length of 34.5 mm (1.36 in).  The left side crack is slightly 
larger in both length and extension.  It features a peak extension of 8.9 mm (0.35 in) from 
the borehole wall and a longitudinal length of 5.8 mm (1.41 in) 
 
 53 
 
 
Figure 32: Bottom view of Test 2, Stage 1 horizontal cracks 
 The unwanted crack development for test 2, stage 1 is limited to mostly crushing 
around the detonation point.  There are a large number of small radial fractures located at 
the bottom of the plug, where the charge was located.  They are all limited to 5 mm (0.2 
in) or less and will be insignificant for later stages of testing.  It is worth noting there is 
an increased roughness of the borehole for this test.  This is due to machining with a 
single flute drill bit that was not very sharp.  This is most notable around the plugged 
section in Figure 33.  This roughness may have helped in the surface bonding of test 2, as 
the plug remained bonded throughout the duration of testing.  Only the epoxy that was 
used to plug the through hole for running wires was blown out.  This may have had a 
 54 
 
minor effect on increasing pressure build up and driving the pin.  The pedal shaped 
cracks detailed in Figure 32 and Figure 33 are due to the impact of the pin hitting the rod, 
which in turn impacts the bottom of the borehole wall.   
 
Figure 33: Right view of Test 2, Stage 1 detonation location 
 
4.4.2 Test 2, Stage 2 
 The results from test 2, stage 2 are shown from a top view in Figure 34.  There is 
substantial crack growth on both sides of the borehole.  Both sides of the borehole have 
an arcing crack that extends from the flat end (above the beginning of the drill bit taper), 
to the circumferential crack originating at the end of the stage 2 plug.  The right side of 
 55 
 
the borehole has a horizontal crack with a longitudinal length of 59.2 mm (2.33 in).  The 
shape of this crack is an almost symmetric arc, centered about the embedded blades.  The 
maximum extension out from the borehole wall is 38.1 mm (1.5 in).  The left horizontal 
crack has an antisymmetric arcing shape that favors the front surface of the block.  The 
longitudinal length of this major crack is also 59.2mm (2.33 in). However, the maximum 
extension from the borehole has a slightly smaller peak of 36.6 mm (1.44 in).  A much 
smaller minor horizontal crack also formed on the left side.  It originates at the 
circumferential crack at the base of the stage 2 plug and extends towards the front surface 
23.6 mm (0.93 in).  The peak extension of this minor crack is 11.2 mm (0.44 in).  A 
noteworthy result regarding the orientation of the horizontal cracks is shown in Figure 35.  
The crack on the left side is oriented roughly 15° off of the horizontal plane.  This is due 
a misalignment within the borehole.  Confirmation that this is not an effect from the in 
situ stress is that the crack does not show curvature, it instead deforms in a straight line 
outward from the misaligned blade. 
 56 
 
 
Figure 34: Top view of Test 2, Stage 2 
 The unwanted crack development from stage two is showcased best in Figure 35.  
There are significant radial cracks around the detonation point.  The largest of these 
cracks extend outwards to 18.5 mm (0.73 in).  The increase in magnitude of these 
unwanted cracks is likely due to the wall roughness within the borehole.  The pedal 
shaped cracks originating in stage 1 are further extended, but not to a significant degree.  
As expected, there is a circumferential crack extending from the discontinuities at the 
bottom of the borehole and bottom of the stage 2 plug. 
 57 
 
 
Figure 35: Back view of Test 2, Stage 2 
 
4.4.3 Test 2, Stage 3 
 The quasi-static loading of test 2 reached 6.9 MPa (1,000 psi), causing crack 
growth on both sides of the borehole, before the left horizontal fracture reached the outer 
surface.  The major crack growth of the left horizontal fracture is marked with red arrows 
on Figure 36.  The fracture started expanding from around the peak distance and 
midpoint of the horizontal crack developed in stage 2.  It extends longitudinally up the 
borehole towards the front surface, reaching 31.2 mm (1.22 in) from the front surface.  
This gives a total longitudinal length of 81.8 mm (3.22 in).  The minor extension of the 
right side crack is labeled in Figure 36 as well. The fact this crack grew at a similar 
pressure implies that within an infinite medium, when enough resistance is provided for 
the left horizontal crack to arrest, the crack propagation is likely to switch to the right 
 58 
 
side crack.  The smoothness and curvature of the left side fracture is displayed in Figure 
37.  Although the dynamically driven crack was already oriented 15° off horizontal, there 
is a significant change in curvature once the crack is driven statically.  
 
Figure 36: Bottom view of Test 2, Stage 3 
 59 
 
 
Figure 37: Front view of Test 2, Stage 3 detailing curvature 
 
4.4.4 Discussion of Repeatability 
 Comparing test 1 and test 2, it is apparent that the mechanism of fracture initiation 
is repeatable.  The cracks formed directly from the function of the device in stage one are 
extremely similar in extension and length.  When detonating a second charge during stage 
2, some deviations in crack growth are observed, but this is expected.  Crack patterns 
between specimens are expected to show variations due to randomly distributed flaws in 
materials [24].  This effect could be enhanced by the variations in borehole conditions, 
 60 
 
specifically the change in roughness due to machining methods.  Instead, if comparing 
the broad patterns and behaviors, tests 1 and 2 are very similar.  The statically grown 
fractures are not identical in size and shape, but exhibit a similar trend.  These fractures 
grow outward until reaching the surface, then lose pressure.  Both fractures display 
smooth surfaces and an influence due to the in situ stress.  The successful repetition and 
consistent behavior of crack growth show that the function of this device is repeatable. 
 
4.5 No-Load Condition 
 Test 3 was conducted without the vertical load simulating in situ stress for two 
purposes.  One, without the restriction of placing the model in the hydraulic press, high 
speed video of the device could be taken from a better angle.  Allowing a top view to 
observe the wedge ejection and crack formation. Two, to observe the crack behavior 
without the in situ stress imposed.  The geometry of the wedges, rod, and block remained 
the same, however, a new tapered pin was made.  This pin retained all the same 
dimensions from the original design except an increased thickness of the 15/16 inch head.  
The thickness of the head was changed from 1.59 mm (1/16 in) to 6.35 mm (1/4 in).  This 
was done to increase durability and allow multiple uses of the pin.  The drill bit used to 
machine the borehole of this test caused significant roughness. 
 
  
 61 
 
4.5.1 Top View of Functioning Device 
 The high speed camera was positioned to record form a top view during the test 3, 
stage 1.  Selected images are presented in Figure 38, showing key moments in the 
function of the device.  The first image presented takes place 44 microseconds after the 
time of detonation.  The borehole is still saturated by light, but the outline of some 
deformation around the detonation point is evident.  By 67 microseconds, the light 
produced by detonation has reduced enough to see the crushing formed around the 
borehole.  The pin has yet to change position.  The third image at 100 microseconds is the 
moment the pin begins to move down the borehole.  At this point in time, the epoxy has 
failed at the top of the plug and pressure is no longer contained within the borehole.  The 
frame taken at 144 microseconds shows progression in the failure at the top of the plug.  
The top, larger diameter section of the plug fractures and breaks off, however, the portion 
of the plug bonded to the borehole wall remains intact.  Image 5 marks the moment when 
although no cracks have formed, the pin has moved enough to displace the wedges 
against the borehole wall.  At 418 microseconds after detonation is triggered, the wedges 
have displaced outwards and horizontal crack growth has been initiated, favoring the 
right side of the block.  In the seventh image, both cracks continue to grow, with the left 
crack beginning to catch up in size.  At 605 microseconds, the pin and rod slam into the 
bottom of the borehole and initiate a circumferential crack.  Both horizontal cracks and 
the circumferential cracks continue to grow until 680 microseconds, where the pin and 
rod are observed to bounce and begin to move back up the borehole.  The last image at 
726 microseconds is when all cracks arrest and the pin and rod have moved a noticeable 
 62 
 
amount up the borehole.  It is important to note during the last two images, the 
circumferential continued to see small amounts of growth.  
 
Figure 38: High speed images of functioning device 
 63 
 
4.5.2 Test 3, Stage 1 
 Looking at the overview of crack growth presented from the top view in Figure 
39, the overall crack growth is significantly larger.  The largest horizontal crack develops 
on the right side of the block.  The crack initiated around the embedded blade and 
propagated down to the beginning of the tapered section of the borehole.  The total 
longitudinal length of the right crack is 43.2 mm (1.7 in).  The crack reaches a maximum 
horizontal distance of 14.2 mm (0.56in) from the borehole wall.  The left horizontal crack 
extends down the borehole toward the back surface, but does not reach the end of the 
borehole.  The total length of the left crack is 40.9 mm (1.61 in).  The crack’s peak 
reaches 9.9 mm (0.39 in) outward from the borehole wall.  The left horizontal crack is 
composed of three smaller cracks that line up in the same plane, whereas the right crack 
is a single continuous crack. 
 Looking at the detonation point, there is significant crushing around the borehole.  
Most of the cracks formed are randomly distributed radial cracks less than 3.2 mm (0.125 
in) in length.  There are a few radial cracks that exceed 3.2 mm (0.125 in), with the 
largest being oriented in the horizontal plane.  The length of this crack is 9.1 mm (0.36 
in).  There is also a circumferential crack with a diameter of 55.9 mm (2.2 in) at the 
bottom of the borehole caused by the impact of the pin.   
 
 64 
 
 
Figure 39: Top view of Test 3, Stage 1 
 
4.5.3 Test 3, Stage 2 
 The results of the stage 2 are shown in Figure 40.  The left horizontal crack has 
the most extensive crack growth resulting from the second stage of pressurization.  The 
horizontal crack extends down the borehole to the circumferential crack closest to the 
back surface of the block.  It also extends up the borehole and past the second stage plug, 
 65 
 
merging with other smaller horizontal cracks.  The combination of these horizontal 
cracks create a total longitudinal distance of 118 mm (4.67 in) in the horizontal plane.  
The maximum distance from the borehole of 34.5 mm (1.36 in) is reached at 
approximately the midpoint of where the blades were ejected.  Although the left side has 
the largest over all crack growth, the right horizontal crack experienced the largest crack 
extension outwards from the borehole.  The maximum distance from the borehole was 
38.1 mm (1.5 in) and was also centered about the point of wedge ejection.  Similarly to 
the left side, the right side extends down the borehole to the circumferential crack closest 
the back surface of the block.  The horizontal crack on the right side extends up the 
borehole reaching the bottom of the stage 2 plug and merging with the circumferential 
crack formed at the boundary.  There are multiple smaller horizontal cracks that formed 
on the other side of the stage 2 plug, but they did not connect with the main crack formed 
from the ejected wedges.  The total longitudinal length of the right horizontal crack is 110 
mm (4.34 in).  
 Despite the severe roughness of the borehole walls, it does not appear to affect the 
crack formation a significant amount.  Around the detonation point some radial cracks 
have formed that seem to originate from visually rough areas of the borehole, but they are 
all ill-defined and short in length.  The largest of these radial cracks in the vertical 
direction extends 8.4 mm (0.33 in) from the borehole wall.  There are two significant 
circumferential cracks that have formed at the bottom of the borehole.  The crack formed 
in stage 1 due to the impact of the pin has experienced additional growth.  The diameter 
has grown slightly to 58.4 mm (2.3 in).  The longitudinal size has grown considerably, 
extending towards the back surface of the block 7.4 mm (0.29 in).  The second 
 66 
 
circumferential crack initiated in stage 2 of testing due to the reflection at the 
discontinuity.  It features a diameter of 45 mm (1.77 in).  
 
Figure 40: Top view of Test 3, Stage 2 
 
4.5.4 Test 3, Stage 3 
 The crack growth resulting from stage 3 is best seen from the front view, as 
presented in Figure 41.  The static pressure reached approximately 6.2 MPa (900 psi) 
before the right horizontal crack expanded to the outer right surface and lost pressure.  
 67 
 
The outline of the major crack is marked with red arrows in the figure.  The fracture 
growth in both the longitudinal direction and horizontal direction was drastic.  The 
fracture growth extended longitudinally downward, past that of stage 2, extending past 
the largest circumferential crack by 7.4 mm (0.29 in).  The crack also grew longitudinally 
up the borehole past the stage 2 plug and connected to the smaller horizontal cracks 
developed in stage 2.  The total longitudinal length of the static crack peak to peak is 118 
mm (4.65 in).  When expanding outward horizontally the crack remained at a significant 
longitudinal crack length.  When it reached the outside surface the crack was still 86 mm 
(3.4 in) in longitudinal length. 
 
Figure 41: Front view of Test 3, Stage 3 
 
  
 68 
 
4.5.5 No-Load Condition Discussion 
 The no-load condition provides insight into the magnitude in which the in situ 
stress affects crack development.  In stage 1, the cracks developed around the detonation 
point are no longer largest in the vertical direction. This shows that without the influence 
of the in situ stress, the crack growth extends in whichever direction has the most 
significant surface flaws, likely randomized depending on manufacturing methods.  
Additionally, in all stages of testing, the crack development was to a greater magnitude in 
the horizontal direction, showing the device had better control over the orientation of 
growth.   
In stage 2, the relationship between vertical radial fractures and in situ stress can 
be observed.  The radial crack development in the non-horizontal directions around the 
detonation point were lower than that of the previous loaded tests, despite the increased 
surface roughness.  This shows that without the constraints on horizontal crack growth 
provided by the large compressive stress, the driving energy from detonation is more 
efficiently used to expand the existing cracks.  Stage 2 of testing also showed that cracks 
in the horizontal plane are more easily extended up the borehole, past the boundary of the 
plug.  This is likely due to the randomly distributed horizontal cracks farther up the 
borehole, which connected to the wedge driven horizontal cracks.  Other loaded tests 
have not seen any horizontal crack development, aside from the ones initiated by the 
device.   
Stage 3 confirmed the earlier hypothesis that the curvature of the statically driven 
cracks is due to the in situ stress.  Also, without the restraint on horizontal crack 
 69 
 
development, the width and overall surface area of the crack was much larger than the 
loaded tests.  The statically formed crack in stage 3 was 86.4 mm (3.4 in) when reaching 
the outer surface, compared to loaded tests that were closer to 25.4 (1 in) in longitudinal 
length at the outer surface. 
 
4.6 Variations to Wedge Geometry 
 After establishing a functional device, three tests with varying wedge geometries 
were conducted to study the effect on crack growth.  Test 4 features a 20° wedge angle 
and a length of 25.4 mm (1 in).  Test 5 features a 10° wedge angle and a length of 25.4 
mm (1 in).  Test 6 remained at the same wedge angle as Test 1-3, but instead has a 
longitudinal length of 12.7 mm (0.5 in).  All three tests used the original block and slotted 
rod geometry.  The pins used in this test have the increased thickness of ¼ inch, as 
described in Test 3.  Test 4 and Test 5 were machined with the same drill bit and surface 
roughness as Test 3.  For Test 6, a new bit was used that resulted in significantly reduced 
roughness.  The impact of roughness and other borehole characteristics will be discussed 
in Section 4.8. 
 
4.6.1 Test 4, Stage 1  
 The resulting cracks from stage 1 are best seen by looking at both Figure 42 and 
Figure 43.  The horizontal cracks developed are difficult to see from the top view, so they 
have been traced with red arrows.  On both the right and left sides of the borehole two 
 70 
 
cracks developed.  Each side has a small vertical crack oriented 10° from the vertical 
plane pointed towards the top surface. Extending a maximum distance of 3.2 mm (0.125 
in) from the side of the borehole wall and running a longitudinal distance of 25.4 mm (1 
in).  The vertical oriented cracks are both aligned about the embedded wedges.  The 
horizontal cracks that developed on each side are equal or greater in magnitude than the 
vertical cracks.  The horizontal crack on the right side of the borehole is the largest.  It 
extends 5.6 mm (0.22 in) outward from the borehole and has a longitudinal length of 28 
mm (1.1 in). Looking at Figure 43, it is evident that the horizontal cracks are not strictly 
in the horizontal plane.  The right crack is approximately 30° offset in the direction of the 
top surface.  While the left side crack is oriented approximately 40° offset in the direction 
of the top surface.  The horizontal crack on the left side of the borehole is also smaller in 
size than the right side. The crack is composed of two smaller cracks that are aligned in 
the same plane to form a single larger crack. This extends a maximum distance of 3.2 mm 
(0.125 in) from the borehole wall and has a combined longitudinal length of 28 mm (1.1 
in).   
Figure 43 also highlights the roughness of the borehole and some unwanted crack 
development.  There are a high number of very small ill-defined cracks that have grown 
off the borehole, largely in the vertical direction.  Small pedal shaped cracks have also 
developed at the bottom of the borehole due to the impact of the pin.  Figure 44 shows a 
close up from the right side of the block, detailing the deformation around the detonation 
point.  There is some crushing around the borehole and one significant vertical crack.  
The large vertical crack extends 14 mm (0.55 in) outward from the borehole.  The 
longitudinal length of the vertical crack is 26.2 mm (1.03 in) 
 71 
 
 
Figure 42: Top view of Test 4, Stage 1 
 72 
 
 
Figure 43: Back view of Test 4, Stage 1 
 
Figure 44: Right view close-up of detonation point, of Test 4, Stage 1 
 73 
 
4.6.2 Test 4, Stage 2 
 The results of stage 2 can be seen from a top view in Figure 45 and a back view in 
Figure 46.  There was significant growth in the horizontal and vertical direction.  An 
important result shown in Figure 46, is the device initiated cracks have grown in the 
horizontal direction, whereas the resulting horizontal cracks in stage 1 were out of plane a 
notable amount.  The horizontal cracks on the left side of the borehole experienced the 
most growth.  As said in stage 1, there are two separate cracks, each experiencing 
extension.  The maximum horizontal extension from this coalescence of cracks is 32.3 
mm (1.27 in) outward from the borehole.  The combined longitudinal length is 65 mm 
(2.56 in), reaching from the circumferential crack at the bottom of the borehole to the 
bottom of the stage 2 plug.  The right side crack is smaller in both extension and length.  
The maximum extension in the horizontal direction is 22.6 mm (0.89 in).  The crack does 
not extend as far down the borehole, stopping where the borehole begins to change 
diameter due to the drill bit, instead of extending all the way to the circumferential crack.  
It does extend all the way up the borehole to the second stage plug, totaling in a peak to 
peak longitudinal length of 55.2 mm (2.17 in). 
 The vertical crack development is best seen in Figure 46.  There are numerous 
vertical cracks developed, almost reaching the maximum extension of the device initiated 
horizontal crack.  The largest crack in the vertical direction reaches an extension from the 
borehole of 30.2 mm (1.19 in).  These cracks are likely to be caused by the surface 
roughness, and are grown from the small cracks displayed in Figure 43.  
 74 
 
 
 Figure 45: Top view of Test 4, Stage 2 
 75 
 
 
Figure 46: Back view of Test 4, Stage 2 
 
4.6.3 Test 4, Stage 3 
 The results from stage 3 are displayed from a top view in Figure 47.  The pressure 
in the borehole reached 7.58 MPa (1,100 psi) prior to a fracture on the left side reaching 
the outer left surface of the block and losing pressure.  The crack does not extend down 
the borehole longitudinally, but it does increase up the borehole to a total longitudinal 
length of 85.9 mm (3.38 in).  The shape of the crack as it extends horizontally is marked 
with red arrows in Figure 47.  The longitudinal length of the crack when it reaches the 
outside surface is 26.4 mm (1.04 in).  The curvature of the statically driven crack is 
 76 
 
shown in detail in Figure 48.  Note that the curvature is opposite of the original offset of 
the blades in stage 1. 
 
Figure 47: Top view of Test 4, Stage 3 
 77 
 
 
Figure 48: Front view of Test 4, Stage 3 
 
4.6.4 Test 5, Stage 1 
 The results show the 10° blade to be an ineffective wedge at initiating crack 
growth.  The blades buckled during the test and have negligible penetration of the 
borehole wall.  The small cracks developed are shown in Figure 49.  Both the left and 
right sides of the borehole have crack development that extends less than 1 mm (0.04 in) 
from the borehole.  The left side has a series of very small cracks that cumulatively make 
up around 19 mm (0.75 in) in total longitudinal length.  The right side has crack 
development of negligible size. Figure 50 shows a left side view.  It can be seen that no 
significant crack is developed from the device, but the borehole has been scored where 
the wedges impacted and buckled. 
 78 
 
 
Figure 49: Top view of Test 5, Stage 1 
 
Figure 50: Left side view of Test 5, Stage 1 
 79 
 
4.6.5 Test 5, Stage 2 
 Referring to Figure 51 and Figure 52, it is evident that horizontal cracks did not 
form in stage 2 of testing.  The scored borehole wall from stage 1 was left unchanged and 
only vertical cracks experienced significant growth.  Stage three was not conducted, as 
there were no horizontal cracks capable of being expanded by quasi-static loading. 
 
Figure 51: Left side view of Test 5, Stage 2 
 80 
 
 
Figure 52: Right side view of Test 5, Stage 2 
 
4.6.7 Test 6, Stage 1 
  The horizontal crack development from the half inch, 30° wedges used in Test 6 
can be seen in Figure 53.  It is also worth noting the significant reduction in surface 
roughness in Test 6, due to the use of a new drill bit during machining.  The bottom tip of 
the borehole is now at an angle of approximately 80° instead of the standard 118°.  The 
two wedges have ejected and formed significant horizontal cracks on each side of the 
borehole.  The right side of the borehole displays the largest crack formation, composed 
of one large crack, and a smaller crack superimposed to form a single arcing fracture.  
The total longitudinal length formed is 24.6 mm (0.97 in).  The maximum extension 
outward from the borehole wall is 8.6 mm (0.24 in).  The left side crack is composed of 
 81 
 
five smaller cracks, failing to coalesce to the same degree as the right horizontal crack.  
The maximum horizontal distance reached on the left side of the borehole is 6.9 mm 
(0.27 in).  The longitudinal length spanned by the series of small cracks reaches 22.6 mm 
(0.89 in).   
 The impact of the pin on the bottom of the borehole creates large pedal shaped 
cracks as shown in Figure 53.  The quantity and magnitude of these pedals is far more 
significant than the other loaded tests.  The cracks together span a horizontal distance of 
46.4 mm (1.83 in) peak to peak. The right most pedal also has a longitudinal length of 
13.7 mm (0.54 in).  A single large vertical crack is shown in Figure 54.  After reviewing 
high speed video footage, it was determined that this crack was formed by the pin 
dragging along the borehole wall.  The longitudinal length is 33 mm (1.3 in) and the peak 
extension from the borehole in the vertical direction reached 14.7 mm (0.58 in).  As 
expected, there is also some minor crack development and crushing local to the 
detonation point, limited to radial fracture length of 5 mm (0.2 in).  It is worth noting that 
when the epoxy blew out of the top of the plug, the plug head (section with 25.4 mm (1 
in) diameter) fractured radial and blew out the top of the block.  Due to the bonding 
between the plug and borehole wall, some cracking forming in the block as well, the 
effect of this is labeled in Figure 54. 
 82 
 
 
Figure 53: Bottom view of Test 6, Stage 1 
 
Figure 54: Left side view of Test 6, Stage 1 
 
 83 
 
4.6.8 Test 6, Stage 2 
 Note that the stage 2 plug used for Test 6 was inserted an extra 5mm (0.2 in) in an 
effort to plug the vertical crack developed in stage 1.  The cracks growth due to stage 2 is 
presented from a top view in Figure 55.  The well-defined arcing horizontal crack from 
stage 1 on the right side undergoes the largest expansion.  This crack extends the full 
length of the borehole until it reaches the circumferential crack at each discontinuity, 
reaching a total longitudinal length of 54.9 mm (2.16 in).  The two cracks from stage 1 
fully merged and extend outward horizontally to a peak of 34.5 mm (1.36 in).  The 
horizontal cracks on the left side of the borehole experience a much smaller degree of 
growth.  The series of small cracks formed in stage 1 extend individually in stage 2, 
instead of merging.  This is labeled in Figure 56.  The lack of coalescence is detrimental 
to the cracks ability to propagate outward.  The maximum distance from the borehole 
achieved is 18.2 mm (0.72 in).  The cracks do not span the entire longitudinal length of 
the borehole.  They reach the circumferential crack at the bottom of the stage 2 plug, but 
only extend downward a total of 43.9 mm (1.73). 
 Highlighted in both Figure 55 and Figure 56, the large vertical crack initiated in 
stage 1 merges with newly formed vertical cracks from stage 2.  It is likely that the pre-
existing flaw aides in the vertical development of new flaws in stage 2, as the other 
locations experience relatively uniform radial crack growth, not exceeding 7.6mm (0.3 
in).  The vertical cracks do not reach farther than the 14.7 mm (0.58 in) developed in 
stage 1, but the longitudinal length of the series of vertical cracks reaches a total of 66 
mm (2.6 in).  A new circumferential crack forms at the bottom of the borehole and the 
 84 
 
bottom of the stage 2 plug.  The pedal shape cracks formed in stage 1 expanded and 
merged into a large cone shaped circumferential crack.  The new diameter reaches 55.9 
mm (2.2 in) and the longitudinal length expanded to a peak distance of 19.3 mm (0.76 in) 
towards the back surface.  The cracking due to the fracture of the stage one plug remains 
unchanged. 
 
Figure 55: Bottom view of Test 6, Stage 2 
 85 
 
 
Figure 56: Left side view of Test 6, Stage 2 
 
4.6.9 Test 6, Stage 3 
 Test 6 was pressurized quasi-statically to 8.27 MPa (1,200 psi) before two cracks 
formed and extended to the outer bottom surface.  One of the cracks originated at the 
right horizontal crack, extending down the borehole before merging with the large 
circumferential cracks initiated by the pin impact in stage 1.  The interaction with the 
circumferential crack directs the crack downward sharply to the bottom surface.  
Although the right crack down not expand all the way outward to the right surface, it 
experiences a significant amount of horizontal extension.  The maximum distance 
reached from the borehole is 53 mm (2.09 in). Additionally, the right crack has a large 
width of 31 mm (1.22 in) when reaching the bottom surface.  This is larger than the other 
 86 
 
loaded condition tests, most likely because it is propagating in the vertical direction and 
not experiencing as much resistance from the in situ stress.  The second crack is a 
circumferential expansion of the back most circumferential crack.  The crack growth 
originates at the left most tip, expanding outwards circumferentially then downwards 
towards the bottom surface.  Both cracks are outlined in red arrows in Figure 57.  
Additionally, the side view of each crack is presented in Figure 58. 
 
Figure 57: Back view of Test 6, Stage 3 
 87 
 
 
Figure 58: Right side view of Test 6, Stage 3 
 
4.6.10 Discussion of Variations in Wedge Geometry 
 The variation in wedge blade angle appears to have an effect on the device’s 
capability of ejecting the wedges and initiating crack growth in stage 1.  Both Test 4 
(20°) and Test 5 (10°) had a lesser degree of wedge ejection, compared to all tests 
conducted with a 30° wedge.  This is likely due to the difference in mass and structure of 
the portion of the blade deforming the wellbore wall.  The extreme of this is seen with 
Test 5, where the blade completely buckled and was unable to produce any significant 
horizontal cracks.  The effect on Test 5 was less pronounced.  The 20° wedges were 
capable of penetrating the borehole wall, but did not extend out from the borehole nearly 
 88 
 
as far.  One of the horizontal cracks developed was much smaller than average.  
However, when stage 2 was conducted on Test 4, the crack growth was comparable to 
those resulting from the 30° wedges, even the crack extending out from the small 
horizontal crack in stage 1. This observation shows that while blade angle impacts the 
ability to drive out the wedges, as long as the blades are capable of penetrating the 
borehole walls without buckling, the results will be comparable.  However, due to the 
small cracks developed in stage 1 of Test 4, the recommended wedge angle for future 
tests is 30°, as it is shown to produce the best results at all stages of testing. 
 The most notable effect of wedge length is the ability to consistently initiate a 
single well-defined crack.  As seen in Test 6, the blades successfully ejected to the same 
extent as the other 30°, however, the resulting cracks do not have the same 
characteristics.  The left side crack is composed of a series of less defined cracks that do 
not coalesce during stage 2.  The inability to merge and create a large defined crack is 
problematic when attempting to grow horizontally and overcome the in situ stress.  For 
this reason, the 25.4 mm (1 in) blade is the recommended wedge length. 
 
4.7 Sensitivity to borehole characteristics 
 The crack growth behavior displayed in the model testing described above shows 
a dependency on the borehole conditions.  Specifically, the roughness of the borehole 
wall and the various discontinuities due to plugging and machining. 
 
 89 
 
4.7.1 Borehole Wall Roughness 
 The borehole wall roughness does not seem to directly impact the ability of the 
device to eject the wedges.  However, it does provide a potential flaw for non-horizontal 
cracks to grow from.  These flaws can compete for crack growth with the desired 
horizontal cracks.  This is undesirable, as it uses energy that could otherwise drive the 
horizontal cracks farther.  Cracks growing from roughness flaws also are problematic if 
they become more well-defined than the horizontal cracks.  If this occurs, it has the 
potential to interfere with the quasi-static expansion of the horizontal fractures.  The 
borehole roughness seems to be most relevant during stage 2 of testing.  When using 
water, the coupling is increased for the purpose of driving out the horizontal cracks.  
However, coupling also increases the tendency of the roughness flaws to initiate crack 
growth.  For the tests that had successful blade ejection, the surface roughness is ordered 
most rough to least rough as follows: Test 4, Test 3, Test 2, Test 6, and Test 1. This is 
shown visually in Figure 59.  Note that Test 2 roughness is difficult to see due to the 
machining oil occupying the small cracks on the surface. The reason for this observed 
roughness pattern is that the drill bit gradually deteriorated, then was sharpened new for 
Test 6.  The amount of vertical crack development in stage 2 directly follows the pattern 
of roughness except for Test 3, the no-load condition.  This shows that under the same 
stress field, surface roughness will dictate the extent of radial crack development in the 
undesired orientation. However, a noticeable decrease in vertical crack development is 
seen when the vertical compressive stress is absent. 
 90 
 
 
Figure 59: Roughness Comparison 
 
4.7.2 Borehole Discontinuities 
 After reviewing the test results, the relationship between circumferential crack 
growth and discontinuities is obvious.  When the pressure reaches a change in borehole 
diameter or plug bottom it increases and creates a circumferential crack.  These cracks 
are mostly insignificant when attempting to grown horizontal fractures, but as seen in 
Test 6, they can change behavior if the cracks coalesce.  In full-scale testing, plugging the 
 91 
 
bottom of the borehole so that future circumferential crack growth does not occur is an 
option.  Additionally, in a full-size wellbore there is more flexibility with longitudinal 
spacing.  Placing the device farther away from the bottom of the borehole could avoid the 
problems observed in Test 6. 
The circumferential crack development seen from the pin impacting the bottom of 
the borehole can also be mitigated.  Comparing Test 6 with the models machined with a 
less steep drill bit shows that the angle affects the crack development from this impact.  It 
is likely that a flat borehole end would best support the impact of the pin and cause less 
resulting crack growth.  An alternative method of minimizing this growth is reducing the 
mass of the pin.  For the purpose of model testing, a large pin head diameter was chosen 
for durability and the convenience of reusing components.  However, the thin headed pin 
used for Test 1 and Test 2 was equally effective, while producing less of a crack due to 
impact.  
 
4.8 Implications and Recommendations for Full-Size Testing 
 As mentioned in the discussion of wedge geometry, the best consistent results 
have been using the 30° wedge angle and 25.4 mm (1 in) length.  The cross section of the 
wedges are scaled by relating the model borehole diameter to the borehole diameter of 
full-scale.  Typical oil and gas borehole diameters range between 120 mm (4.7 in) and 
400 mm (15.7 in) [30].  Assuming a medium sized borehole diameter of 200 mm (7.9 in), 
the length scaling factor will be 8.4.    An engineering drawing and relevant dimensions 
of the components are shown in Figure 60, Figure 61, and Figure 62.  
 92 
 
 
Figure 60: Engineering Drawing of Scaled Wedge 
 
Figure 61: Engineering Drawing of Scaled Pin 
 93 
 
 
Figure 62: Engineering Drawing of Scaled Rod 
The charge size will use cube root scaling. Meaning for a length scaling factor of 
8.4, the charge mass will be increased by a factor of 592.7.  Including the initiating 
explosive, the total charge masses at model scale are 260 mg and 110 mg, for stage 1 and 
stage 2 respectively.  Scaling to full-scale results in a first stage explosive mass of 154 
grams, to be detonated in an aired filled borehole and a second stage explosive mass of 
65 grams, to be detonated in a water filled borehole. 
From the results observed during model testing, it is advised that the outside 
diameter of the pin head be lubricated prior to testing.  This will ensure that it does not 
drag significantly against the borehole surface causing unwanted crack development, as 
seen in Test 6.  It is also important that a sharp drill bit is used when creating the 
wellbore.  Minimizing surface roughness will increase the effectiveness of the device in 
regards to controlling crack growth orientation.   
 94 
 
Chapter 5: Conclusions and Future Work 
5.1 Conclusion 
The purpose of this research was to develop a mechanical device capable of initiating 
fracture growth in a horizontal wellbore, providing a more environmentally friendly 
method of accessing unconventional reservoirs.  This problem was modeled by 
controlling the orientation of crack growth in a borehole loaded by an unfavorable in situ 
stress field.  Preliminary testing was performed to explore mechanisms of initiating 
fractures, determine an appropriate size scale of explosives and device components, and 
assess optimal fluid medium for various stages of testing.  The results from preliminary 
testing were used in an iterative process to make changes to the developing mechanical 
device until an acceptable design was determined. 
 After determining an initial design, six detailed tests were conducted.  The 
repeatability of the device was proven and acceptable variation in the results was 
discussed.  A no-load test was performed to study the effect of the applied in situ stress 
and observe the successfully functioning device with a high speed camera.  Three 
variations in component geometry were manufactured and tested to determine which 
geometry produced optimal fracture growth.  The sensitivity to changes in borehole 
conditions was discussed.  Finally, the best performing device geometry was scaled and a 
schematic is prevented for full scale manufacturing and testing.   
The results presented in this paper prove that using an insertable device is a 
promising method for the purpose of controlled fracture initiation.  The ability to produce 
well-defined horizontal cracks in the presence of a vertical compressive stress 
 95 
 
demonstrates the appeal of using this technique.  The method presented is a potential 
alternative to the current methods used in hydraulic fracturing, but additional testing is 
recommended to assess the performance at a larger scale.  
 
5.2 Suggested Future Work 
 With the successful development of a small-scale model device, the next step 
would be increasing the scale or complexity of the testing.  This could include using the 
recommended scaled device to see if the results can be replicated in the real environment.  
A less aggressive and more economically feasible option might be to conduct a new 
series of model tests in a non-homogenous material.  Using a lamina structure to mimic 
the joints found in rock would give insight into how interfaces and boundaries would 
affect crack propagation.  Increasing the complexity of the model would undoubtedly be 
beneficial in further developing an insertable device for controlling crack growth 
orientation. 
 Another opportunity for continuing this work would be variation in material 
choices and pressurization methods of the current device, specifically manipulating the 
mechanical characteristics of the tapered wedge.  Implementing a destructive wedge 
capable of initiating cracks before losing its structure would be very beneficial.  
Accomplishing this would alleviate the complications of removing the ejected wedges in 
full-scale modeling.  Finally, considering implementing a fourth pressurization stage is an 
alternative method of improving the current process.  An investigation into using 
 96 
 
dynamic loading via explosives could be performed as a fourth stage to increase the 
roughness and crack quality of the cracks previously formed by quasi-static loading. 
 
 
  
 97 
 
Bibliography 
[1]  “U.S. Energy Facts Explained: Consumption & Production,” U.S. Energy  
Information Administration. [Online]. Available: 
https://www.eia.gov/energyexplained/?page=us_energy_home. [Accessed: 28-
Dec-2017]. 
 
[2]   “Annual Energy Outlook 2016: Marked Trends: Natural Gas,” U.S. Energy  
Information Administration, 2016. [Online]. Available: 
https://www.eia.gov/outlooks/archive/aeo16/MT_naturalgas.cfm#natgasprod_exp. 
[Accessed: 28-Dec-2017]. 
 
[3]   National Energy Technology Laboratory. “Modern Shale Gas Development in the  
United States: A Primer,” U.S. Department of Energy, 2009. [Online]. Available: 
https://energy.gov/sites/prod/files/2013/03/f0/ShaleGasPrimer_Online_4-
2009.pdf. [Accessed: 28-Dec-2017]. 
 
[4]   House, Review of hydraulic fracturing technology and practices: hearing before  
the Committee on Science, Space, and Technology, House of Representatives, 
One Hundred Twelfth Congress, first session, Wednesday, May 11, 2011. 
Washington: U.S. G.P.O., 2011. 
 
[5]   Veatch, R.W., Jr., Moschovidis, Z.A., and Fast, C. R. 1989. An Overview of  
Hydraulic Fracturing. Recent Advances in Hydraulic Fracturing, ed J.L. Gidley, 
S.A. Holditch, D.E. Nierode, and R.W. Veatch Jr., Vol. 12. Richardson, Texas: 
Henry L. Doherty Monograph Series, SPE.  
 
[6]   Jennings, A.R., Jr., and Darden W. G. 1979. Gas Well Stimulation in the Eastern  
United States. SPE 7914. 
 
[7]    Stephenson, M.H. Shale gas and fracking: The science behind the controversy.  
Amsterdam (Holanda): Elsevier Science, 2015. 
 
[8]    Overbey, W.K., A.B. Yost, and D.A. Wilkins. 1988. Inducing Multiple Hydraulic  
Fractures from a Horizontal Wellbore. SPE 18249. 
 
[9]    Chesapeake Energy Corporation. Components of Hydraulic Fracturing, Presented  
to the NY DEC. October 2008. 
 
[10] Wood, P. “Maryland General Assembly approves fracking ban,” Baltimore Sun,  
28-Mar-2017. [Online] Available: 
http://www.baltimoresun.com/news/maryland/politics/bs-md- 
fracking-ban-passes-20170327-story.html. [Accessed: 28-Dec-2017]. 
 
[11] Evensen, D., Jacquet, J.B., Clarke, C.E., and Stedman, R.C, ‘What’s the  
 98 
 
‘fracking’ problem? One word can’t say it all’, The Extractive Industries and 
Society, vol. 1, no. 2, pp. 130-136, 2014. 
 
[12] Zhang, D., and Yang, T, ‘Environmental impacts of hydraulic fracturing in shale  
gas development in the United States’, Petroleum Exploration and Development, 
vol. 42, no. 6, pp. 876-883, 2015. 
 
[13] Howarth, R.W., Ingraffea, A., and Engelder, T., ‘Natural gas: Should fracking  
stop?’, Nature, vol. 477, no. 7364, pp. 271-275, 2011. 
 
[14] Ellsworth, W L., ‘Injection- induced earthquakes’, Science, vol. 341, no. 6142,  
pp. 1225942-1–1225942-7, 2013. 
 
[15] King, G E. 2012. Hydraulic fracturing 101: What every representative,  
environmentalist, regulator, reporter, investor, university researcher, neighbor and 
engineer should know about estimating frac risk and improving frac performance 
in unconventional gas and oil wells. SPE 152596. 
 
[16] Osborn, S.G., Vengosh, A., Warner, N.R., et al., ‘Methane contamination of  
drinking water accompanying gas-well drilling and hydraulic fracturing’, 
Proceedings of the National Academy of Sciences, vol. 108, no. 20, pp. 8172-
8176, 2011. 
 
[17] Jackson, R.B., Vengosh, A., Darrah, T.H., et al., ‘Increased stray gas abundance  
in a subset of drinking water wells near Marcellus shale gas extraction’, 
Proceedings of the National Academy of Sciences, vol. 110, no. 28, pp. 11250-
11255, 2013.  
 
[18] Warner, N.R., Christie, C.A., Jackson, R.B., et al., ‘Impacts of shale ga  
wastewater disposal on water quality in western Pennsylvania’, Environmental 
Science & Technology, vol. 47, no. 20, pp. 11849-11857, 2013. 
 
[19] Barker, D.B., Fourney W.L., and Dally, J.W., ‘Fracture control in tunnel blasting:  
In tunneling and underground structures’, Transportation Research Record, no. 
684, 1978. 
 
[20] McKown, A. F., Fourney, W.L., Sperry P. E., and Thompson, D.E., ‘Field  
evaluation of fracture control in tunnel blasting,’ Report No. 3060, U.S. 
Department of Transportation, Sep 1978. 
 
[21] Fourney, W.L., and Leiste, U., ‘Controlled explosive fracturing’, Blasting &  
Fragmentation, vol. 9, no. 1, 2015. 
 
[22] Fourney, W.L., and Barker, D.B., ‘Characteristics of a crack driven by explosive  
loading’, UMD Report to Department of Energy, Aug 1979. 
 99 
 
 
[23] Fourney, W.L., Barker, D.B., and Holloway, D.C., ‘Model studies of explosive  
well stimulation techniques’, International Journal of Rock Mechanics and 
Mining Sciences & Geomechanics Abstracts, vol. 18, no. 2, pp. 113-127, 1981. 
 
[24] Bieniawski, Z.T., ‘Mechanism of brittle fracture of rock: Part I-theory of the  
fracture process’, International Journal of Rock Mechanics and Mining Sciences 
& Geomechanics Abstracts, vol. 4, no. 4, pp. 395-404, 1967. 
 
[25] Li, Q., Xing, H., Liu, J., and Liu, X., ‘A review of hydraulic fracturing of  
unconventional reservoir’, Petroleum, vol. 1, no. 1, pp. 8-15, 2015. 
 
[26] Saharan, M.R., Mitri, H.S., and Jethwa, J.L., ‘Rock fracturing by explosive  
energy: Review of state-of-the-art,’ Fragblast The International Journal for 
Blasting and Fragmentation, vol. 10, no. 1-2, March-June 2006. 
 
[27] Baker, W.E., Westine, P.S., and Dodge, F.T., Similarity Methods in Engineering  
Dynamics: Theory and Practice of Scale Modeling, Spartan Books, Haydon Book 
Company, Inc., Rochelle Park, New Jersey, 1973. 
 
[28] Swisdak, M.M., ‘Explosion Effects and Properties Part I - Explosion Effects in  
Air’, NSWC-TR 75-116, Naval Surface Warfare Center, White Oak, MD, 1975. 
 
[29] Fu, R., Lindfors, A., and Davis, J., ‘Scaling for Internal Blast’, AIP Conference.  
Proceedings, vol. 706, pp. 1440-1443, 2004.  
 
[30] Lecampion, B., ‘Modeling size effects associated with tensile fracture initiation  
from a wellbore’, International Journal of Rock Mechanics and Mining Sciences, 
vol. 56, pp. 67-76, Dec 2012.  
 
 100