Académique Documents
Professionnel Documents
Culture Documents
By
Hussain Osama Salah
I
II
ACKNOWLEDGEMENTS
I would like to express my deepest and warmest gratitude to all those who helped me
complete this research. I would especially like to give my deepest thanks to my supervisors;
Dr. Maher Omar, Prof. Abdullah Shanableh, and Eng. Ali Tahmaz, who spent so much
effort reviewing, making suggestions and helping me in carrying out this study.
I would also like to thank, Eng. Yousef Al-Soboh, Director of Baynunah Engineering lab,
Abu Dhabi, Eng. Mohammed Obaid, Director of Matrix lab, Dubai, and Eng. Imad Al-Sharif,
Director of the Arab Center of Engineering Studies (ACES) Dubai for their assistance and
contribution. Additionally, I would like to thank everyone who was involved, directly or
I would also like to thank the examiners; Dr. Radhi Al-Zubaidi, the internal examiner, and
Prof. Mousa Attom, the external examiner, for their reviews, remarks, and comments on this
work. Their input was of great help in producing this research as it is now.
Finally, I would like to express my warmest, deepest and heartfelt gratitude to my family for
III
Estimating Unconfined Compressive Strength of
Sedimentary Rocks in the United Arab Emirates
from Generated Empirical Correlations
By: Hussain Osama Salah
ABSTRACT
A laboratory study was conducted to develop a database and models for predicting the
unconfined compressive strength of sedimentary rocks. A large number of rock samples from
different sites in the United Arab Emirates were collected and tested for the development of
the database and evaluation of models. Reliable empirical relationships were developed for
estimating the unconfined compressive strength of UAE rocks based on results obtained from
the following mechanical and physical tests that were performed on rock samples: unconfined
compressive strength, point load strength index, Schmidt rebound, Brazilian splitting and
ultrasonic pulse velocity tests. These were conducted to determine the mechanical properties
of rock specimens, while the bulk specific weight and moisture content test was conducted to
Twenty nine relations were selected from more than a hundred and thirty generated relations
developed. Each relation was the result of a statistical analysis and the application of the
Mean Average (MA) data smoothing algorithm and Least Absolute Residuals (LAR) robust
regression wherever that was necessary. In addition, four general relationships were
developed relating unconfined compressive strength to moisture content, unit weight, point
Laboratory Testing.
IV
Table of Contents
Chapter 1 Introduction............................................................................................................ 1
1.1. Background ..................................................................................................................... 1
1.2. Problem Statement .......................................................................................................... 2
1.3. Objectives ........................................................................................................................ 2
1.4. Scope of Study ................................................................................................................ 3
1.5. Engineering Significance ................................................................................................ 4
1.6. Limitations of the Study .................................................................................................. 4
Chapter 2 Literature Review .................................................................................................. 5
2.1. Definition and History of Rock Mechanics and Engineering ......................................... 5
2.1.1. Definition of Rock Mechanics and Engineering ...................................................... 5
2.1.2. History of Rock Mechanics and Engineering ........................................................... 6
2.2. Main Areas of Interest in Rock Mechanics and Rock Engineering ................................ 7
2.2.1. Interests in Rock Slopes Stability ............................................................................. 7
2.2.2. Interests in Shafts, Tunnels, Caverns and Underground Mines ............................... 8
2.2.3. Interests in Rock Foundations .................................................................................. 9
2.3. Software Usage in Rock Mechanics and Engineering .................................................. 11
2.4. Brief about Sedimentary Rocks .................................................................................... 12
2.5. Geological History of the United Arab Emirates and Its Rocks ................................... 14
2.6. Rock Classification ....................................................................................................... 17
2.7. Unconfined Compressive strength (UCS) of Rocks ..................................................... 19
2.7.1 Background .............................................................................................................. 19
2.7.2. Relations between UCS and Mechanical Properties .............................................. 21
2.7.2.1 Relations between UCS and Point Load Strength Index (Is(50)) ..................... 21
2.7.2.2. Relations between UCS and Schmidt Hammer Rebound Number .................. 23
2.7.2.3 Relations between UCS and Brazilian Splitting Tensile Strength .................... 25
2.7.2.4 Relations between UCS and Ultrasonic Pulse Velocity ................................... 26
2.7.2.5 Relations between UCS and Modulus of Elasticity .......................................... 28
2.7.3. Relations between UCS and Physical Properties ................................................... 31
2.7.3.1 Relations between UCS and Bulk Specific Weight .......................................... 31
2.7.3.2. Relations between UCS and Moisture Content ............................................... 32
V
Chapter 3 Experimental Program ........................................................................................ 33
3.1. Introduction ................................................................................................................... 33
3.2. Sample Collection ......................................................................................................... 33
3.3. Sample Transporting and Storing.................................................................................. 36
3.4. Sample Identification .................................................................................................... 38
3.5. Sample Preparation ....................................................................................................... 39
3.6. Sample Testing .............................................................................................................. 40
3.6.1. Mechanical Tests Done .......................................................................................... 40
3.6.1.1. The Point Load Test. ........................................................................................ 40
3.6.1.2. The Schmidt Hammer Test .............................................................................. 42
3.6.1.3. Brazilian Splitting Test .................................................................................... 42
3.6.1.4. Ultrasonic Pulse Velocity Test......................................................................... 43
3.6.1.5. UCS and E Test ................................................................................................ 43
3.6.2. Physical Tests Done................................................................................................ 45
3.6.2.1. Moisture Content Test...................................................................................... 45
3.6.2.2 Specific Weight Test ......................................................................................... 45
3.6.3. Order of Testing...................................................................................................... 45
Chapter 4 Results, Analysis and Discussion ........................................................................ 47
4.1. Introduction ................................................................................................................... 47
4.2. Test Results ................................................................................................................... 47
4.2.1. Mechanical Test Results ......................................................................................... 47
4.2.1.1 The Schmidt Hammer Test ............................................................................... 47
4.2.1.2 The Point Load Strength Index Test. ................................................................ 48
4.2.1.3. The Ultrasonic Pulse Velocity Test ................................................................. 48
4.2.1.4. The Brazilian Splitting Tensile Test ................................................................ 49
4.2.1.5. The UCS Test ................................................................................................... 49
4.2.1.6. The Young’s Modulus of Elasticity Test ......................................................... 50
4.2.2. Physical Tests Results ............................................................................................ 53
4.2.2.1. The Moisture Content Test. ............................................................................. 53
4.2.2.2 The Unit Weight Test........................................................................................ 53
4.3 Generated Relations ....................................................................................................... 54
4.3.1. UCS Relations with Mechanical Properties ........................................................... 55
4.3.1.1. UCS vs. HR....................................................................................................... 55
VI
4.3.1.3 UCS vs. Vp ........................................................................................................ 63
4.3.1.5. E vs. UCS ......................................................................................................... 68
4.3.2 Multiple Regression Models .................................................................................... 71
Chapter 5 Summary, Conclusion and Final Recommendations ....................................... 76
5.1 Summary ........................................................................................................................ 76
5.2 Conclusion...................................................................................................................... 77
5.3 Recommendations for Future Research ......................................................................... 78
References ................................................................................................................................ 79
Appendixes.............................................................................. Error! Bookmark not defined.
VII
List of Tables
Table 2.1: Relation For Different Is(50) Classes 22
Table 2.2: Recent BST Relations Collected By Nazir 25
Table 2.3: Correlations Found By Yassar and Erdogan 27
Table 3.1: Summary of Borehole Locations and Data 34
Table 4.1: Confidence Interval Estimate of Expected Values of HR 47
Table 4.2: Confidence Interval Estimate of Expected Values of Is(50) 48
Table 4.3: Confidence Interval Estimate of Expected Values of VP 48
Table 4.4: Confidence Interval Estimate of Expected Values of St 49
Table 4.5: Confidence Interval Estimate of Expected Values of UCS 50
Table 4.6: Confidence Interval Estimate of Expected Values of E 50
Table 4.7: Confidence Interval Estimate of Expected Values of ω 53
Table 4.8: Confidence Interval Estimate of Expected Values of γ 53
Table 4.9: Relation Cases Codes 55
Table 4.10: Relations Between UCS and HR, General Case 56
Table 4.11: Other Relations Summary Between UCS and HR 56
Table 4.12: Relations Between UCS and Is(50), General Case 58
Table 4.13: Other Relations Summary Between UCS and Is(50) 59
Table 4.14: Relations Between UCS and VP, General Case 63
Table 4.15: Relations Between UCS and St, General Case 64
Table 4.16: Other Relations Summary Between UCS and St, 65
Table 4.17: Relations Between E and UCS, General Case 68
Table 4.18: Other Relations Summary Between E and UCS 69
Table 5.1: Summary of Direct Relations 77
Table 5.2: Summary of Multiple Regression Relations 77
Table A.1: Schmidt Hammer Test, All Results 84
Table A.2: Point Load Strength Test, All Results 86
Table A.3: Ultrasonic Pulse Velocity Test, All Results 90
Table A.4: Brazilian Test, All Results 92
Table A.5: Unconfined Compression Strength Test, All Results 94
Table A.6: Modulus of Elasticity Test, All Results 98
Table A.7: Moisture Content Test, All Results 99
Table A.8: Unit Weight Test, All Results 103
VIII
List of Figures
Figure 2.1: Rock Slope Stability 7
Figure 2.2: Felton Quarry Granite Slope Failure 8
Figure 2.3: Copiapó Mining Accident 2010 9
Figure 2.4: Illustration of Rock Foundation of Burj Khalifa 10
Figure 2.5: House Poarch Foundation Stones 10
Figure 2.6: Clastic Rock Classification Scheme 13
Figure 2.7: Different Types of Sedimentary Rocks 13
Figure 2.8: Arabian Plate and UAE Position 14
Figure 2.9: Summary Stratagraphic of the UAE Foreland Basin 16
Figure 2.10: Different Compression States of Rocks 20
Figure 2.11: UCS – Is(50) Linear and Power Relation By Tsiambaos and Sabatakakis 22
Figure 2.12: UCS – Is(50) Stratified Relation By Tsiambaos and Sabatakakis 22
Figure 2.13: UCS vs. HR Relation By Shalabi etal. 24
Figure 2.14: UCS vs. HR Relation By Yilmaz and Sendir 24
Figure 2.15: UCS vs. BST Relation By Nazir 26
Figure 2.16: Comparison of Different Studies With Lab Work By Nazir 26
Figure 2.17: Vp vs. UCS Relation By Yassar and Erdogan 27
Figure 2.18: Vp vs. E Relation By Yassar and Erdogan 28
Figure 2.19: Average Modulus of Elasticity Calculation 28
Figure 2.20: 3D Visualization of Equation 2.7 30
Figure 2.21: Nomograph Constructed for Equation 2.7 30
Figure 2.22: UCS vs. ρ Relation By Moh’d 31
Figure 3.1: General Work Plan For Thesis 33
Figure 3.2: Sample Contribution By Place and Type 35
Figure 3.3: Envelop Borehole Locations 35
Figure 3.4: Flow Chart for Preserving and Transporting Rock Core Samples 36
Figure 3.5: Sample ID Illustration 38
Figure 3.6: Sample Ready for the Point Load Test 41
Figure 3.7: Point Load Test Different Failure Patterns 41
Figure 3.8: Failed Sample After Brazilian Test 43
Figure 3.9: Sample Ready for the UCS and E Test 44
Figure 3.10: Sample Testing Stages Flowchart 46
IX
Figure 4.1: Sample of Stress-Strain Diagrams of Samples 50
Figure 4.2: UCS vs. HR Relation, General Case 56
Figure 4.3: UCS vs. HR Relation, AV Case 57
Figure 4.4: UCS vs. HR Relation, AW Case 57
Figure 4.5: UCS vs. Is(50) Relation, General Case 58
Figure 4.6: UCS vs. Is(50) Relation, AV Case 59
Figure 4.7: UCS vs. Is(50) Relation, AW Case 59
Figure 4.8: UCS vs. Is(50) Relation, CA Case 60
Figure 4.9: UCS vs. Is(50) Relation, CW Case 60
Figure 4.10: UCS vs. Is(50) Relation, MA Case 61
Figure 4.11: UCS vs. Is(50) Relation, MV Case 61
Figure 4.12: UCS vs. Is(50) Relation, SA Case 62
Figure 4.13: UCS vs. Is(50) Relation, SV Case 62
Figure 4.14: UCS vs. Is(50) Relation, SW Case 63
Figure 4.15: UCS vs. VP Relation, General Case 64
Figure 4.16: UCS vs. St Relation, General Case 65
Figure 4.17: UCS vs. St Relation, AV Case 65
Figure 4.18: UCS vs. St Relation, AW Case 66
Figure 4.19: UCS vs. St Relation, CA Case 66
Figure 4.20: UCS vs. St Relation, MA Case 67
Figure 4.21: UCS vs. St Relation, SA Case 67
Figure 4.22: E vs. UCS Relation, General Case 68
Figure 4.23: E vs. UCS Relation, AV Case 69
Figure 4.24: E vs. UCS Relation, AW Case 69
Figure 4.25: E vs. UCS Relation, CA Case 70
Figure 4.26: E vs. UCS Relation, MA Case 70
Figure 4.27: Comparison of Works for Rebound Number HR 72
Figure 4.28: Comparison of Works for Point Load Strength Index Is(50) 73
Figure 4.29: Comparison of Works for Ultrasonic Pulse Velocity 74
Figure 4.30: Comparison of Works for Brazilian Strength St 74
Figure 4.31: Comparison of Works for Modulus of Elasticity E 75
X
Chapter 1
Introduction
1.1. Background
A closer look at the development projects and construction boom that occurred in the last
decade in the United Arab Emirates gives thoughtful considerations about the construction of
such major projects. In every construction project, geotechnical investigations are carried out
to determine how the components of the project that interact with the soil should proceed.
Geotechnical investigations vary in complexity and prices; some of them require days of
Therefore, geotechnical engineers thought about devising easier, less sophisticated and
such parameters is also needed to overcome sampling and handling problems. Estimation of
One of the most important rock parameters is the unconfined compressive strength (UCS) test
of rocks, since it is used widely in rock classifications like Rock Mass Rating (RMR),
analysis and design of rock related structures. Here, special sample preparation is involved. In
the case of sedimentary rocks, UCS testing becomes harder due to the fact that the recovered
rocks are sometimes of such geometric parameters that they are not allowed by the code to
have the test performed on them, or some rocks fail in the preparation stage before
performing the UCS test. Therefore, the need of a way to determine this important parameter
arises.
1
Moreover, due to the lack of information on local rocks, the main purpose of this work is to
generate empirical relations between UCS of sedimentary rocks in the UAE and other
Rock’s UCS is a very important test to be done on rock cores to give full understanding of the
rocks’ capabilities to accommodate proposed project loads and to do the RMR classification
analysis for rocks. Sometimes, Rock Quality Designation (RQD) of a certain core specimen
is of a level so low that it is hard to find a core piece to perform the UCS test on, since codes
require a special length to diameter ratio of (2:1) and received rocks’ condition usually
doesn’t meet this requirement. On the other hand, other tests can be done on these rocks' core
specimens like the point load strength index, rebound hammer and many other tests.
Therefore, it is assumed that there is a need for a simpler way to determine the UCS of rocks.
As a result of all of the above, the researcher decided to write a thesis on relating the UCS of
sedimentary rocks of the UAE to other mechanical parameters like point load strength index,
Brazilian splitting strength, modulus of elasticity, rebound number and ultrasonic pulse
velocity, as well as physical properties like bulk specific weight and moisture content. All of
this in order to simplify the approach of estimating the UCS for sedimentary rocks in the
UAE.
1.3. Objectives
The main objective of this research paper is to develop empirical relations between the UCS
of sedimentary rocks in the UAE and other physical and mechanical properties of rocks. The
specific objectives of the study are to relate the rock’s UCS to mechanical parameters like
point load strength index, Brazilian splitting strength, modulus of elasticity, rebound number
2
and ultrasonic pulse velocity, as well as physical properties like bulk specific weight and
moisture content. All tests were performed in accordance with the American Society of
Testing and Materials procedure codes. All math works were done using the MATLAB
software.
The following steps were done to fulfill the work. First, previous studies regarding the same
topic were reviewed. Next, a database with means of identification was created. Then, lab
environment was prepared to receive, store and retrieve specimens. After that, sedimentary
rock samples from different types (mudstone, crystalline gypsum, sandstone and calcarenite),
from different areas in the UAE (western region, central Abu Dhabi City and Dubai) were
acquired. Subsequently, lab tests on acquired samples were conducted. These tests include;
unconfined compressive strength, modulus of elasticity point load strength index, Schmidt
rebound, Brazilian splitting and ultrasonic pulse velocity tests, and were performed to
determine the mechanical properties of rock specimens, while the bulk specific weight and
moisture content test were performed to determine the physical properties of rock specimens.
The created database was then filled with the worked test results ready for relation generation
between data from different performed tests. Afterwards, the correctness and integrity of
found relations was checked versus previously done work regarding the same subject.
This thesis has 5 chapters; chapter 1 presents study background, problem statement,
objectives, significance of the study and study limitations. Chapter 2 presents the literature
review done for the thesis. Chapter 3 discusses methodologies of testing, lab environment and
worked data digestion. Chapter 4 presents and discusses results found in the lab. And, chapter
3
1.5. Engineering Significance
This thesis provides new and advanced knowledge about rocks in the UAE, as it brings
together analysis of strength and different physical properties and rock types of the UAE with
standardized international studies. This work is about developing relations between UCS of
UAE sedimentary rocks with other different, easier to find parameters and widely used tests
due to the lack of information about such properties. It is hoped that this work will provide a
good tool to predict rock’s UCS from other mechanical and physical parameters.
All investigations were done on sedimentary rocks of the following types; sandstone,
mudstone and crystalline gypsum in the UAE only. Therefore, careful generalization of
generated correlations for any different rock types than the aforementioned ones and other
regions than the UAE would be much recommended. Also, this study was conducted on
4
Chapter 2
Literature Review
2.1. Definition and History of Rock Mechanics and Engineering
After many studies in the field of rock mechanics and engineering (RME), researchers in this
field have agreed to use Judd’s definition of RME, which was stated in 1964 and amended in
1974 to become as follows “Rock mechanics is the theoretical and applied science of the
mechanical behavior of rock and rock masses; it is that branch of mechanics concerned with
the response of rock and rock masses to the force fields of their physical environment.” This
definition is more convenient in mining works as mine excavation changes force fields that
rock masses encounter. Additionally, Brady, B. and Brown, E. (2005) noted that rock
mechanics is a diverse science based on the type of rocks considered in the study. For
example, if fragmented or weathered rocks are considered, then rock mechanics approaches
soil mechanics. On the other hand, if rocks are at inaccessible depths for mining and drilling,
rock mechanics approaches mechanical aspects of structural geology. The definition of rock
engineering can be rewritten to conform with engineering definition by Smith (2012) which
combination; or to construct or operate the same with full cognizance of their design; or to
forecast their behavior under specific operating conditions; all as respects an intended
function, economics of operation and safety to life and property,” to be the following “The
related structures and to forecast and monitor their behavior under recommended operating
5
conditions with respects of intended function, economics of operation and safety to lives and
properties.”
Jaeger (1979) and Hoek (2006) have diligently studied the history and development of RME.
They disagreed on the date of the establishment of RME as a modern discipline. Jaeger gave
examples of construction of tunnels in the Alps Mountains in the late 19th century. He
thought that the first notice of residual stresses on rocks was back in 1874 when the German
tunnel expert, Rziha, noted the bursts and squeezing in the tunnels and galleries of the Alps
Mountains. He further stated that Heim, a professor at Zurich University and Zurich Federal
Institute of Technology, concurred with Rziha’s observation. Heim suggested that the order
of magnitude of horizontal forces acting on rocks in these mountains had to be the same as
the vertical forces acting on them. Jaeger observed that the first attempt at rock mechanics
was in 1926 when Schmidt conducted a thesis of which he related what Heim suggested
about residual stresses in rocks to the newly formulated ideas about rock elasticity.
Alternatively, Hoek had a different date of the beginning of rock mechanics and engineering.
He believed that rock engineering was considered a modern discipline as early as 1773 when
Coulomb had included results of Bordeaux rocks testing results in a thesis that was read
before the French academy in Paris. Hoek also gave the construction of the Panama Canal as
an example of the development of RME. He stated that 60 slides occurred in the cuts along
the Panama Canal during its construction and its operation (1910 – 1964). Hoek stated that
Lutton and others have concluded in 1979 that slides have occurred because of the structural
discontinuity of rocks. Hoek also reiterated a part of Karl Terzaghi’s presidential speech in
the first international conference on Soil Mechanics and Foundation Engineering in 1936
about the Panama Canal slides “The catastrophic descent of the slopes of the deepest cut of
6
the Panama Canal issued a warning that we were overstepping the limits of our ability to
One can confidently claim that RME is a science by itself that has its own areas of interest.
The following is a demonstration of some of the main areas of interest which are classified
Slope stability is a branch of geotechnical engineering which deals with the “assessment of
static and dynamic effects on different kinds of slopes; earth and rock-fill dams, slopes of
other types of embankments, excavated slopes and natural slopes in soil and soft rocks.” (US
Army Corps of Engineers, 2003) Based on the previous definition, it could be conclude that
rock slope stability is a conjoined science between slope stability and RME as shown in
Rock
Slope Mechanics
Stability and
Engineering
many areas in the world which are settled next to a rock slope. This becomes more important
in the case of open mines and quarries, since any rock slope failure would result in not only
7
working site hazards but also possible economic consequences due to the subsequent fixing
and rehabilitation required so they could be used again as reported by Rogers (1995).
Figure 2.2 shows the failure of Felton quarry, Santa Cruz, California on November 20, 1992.
It is easy to note the variation in sizes and applications among all main areas of interest in
RME. However, all of them share the main concept of analysis and design. As stated
previously in the definition of RME, it is concerned with the stress state in rocks before and
after the construction of rock related structures, especially in shafts and similar types of
structures where the main construction concept of them, briefly speaking, is to analyze the
stress state of the rock containing them in order to provide the design solution for the project.
The UCS of the rock is very much needed here. Any mistake in this step might be the
deathblow of the whole structure. An example from memory is the Copiapó mining accident
in Chile on August 8th, 2010 reported in Wikipedia (2012). Figure 2.3 illustrates the accident.
8
Figure 2.3: Copiapó Mining Accident on 2010, Acquired from Wikipedia (2012)
Depending on the geotechnical conditions of the site, a geotechnical engineer has only two
options: option one, to construct the project on or through soil which has its own design
considerations, or, option two, to construct the project on or through rocks which have their
previously. A project constructed on a rock bearing stratum is the one considered here.
Different design criteria are needed since the foundation supporting condition is different
than soil. A good designing manual is the US army corps of engineers' manual number EM
1110-1-2908, (1994). This manual gives a good idea about design considerations for large
military and civil engineering structures in terms of design considerations, site investigation,
9
Also, depending on rock depth and the proposed project the design parameters would change
either to construct a small project on a shallow rock bearing stratum, or to construct a major
project on a deep rock bearing stratum like the case of Burj Khalifa in Dubai, UAE, where
piles where constructed through soil to reach the rock stratum which is about 50 meters deep.
The following Figure 2.4 illustrates how the rock foundations for Burj Khalifa look like.
Here, the rocks’ UCS played an important role in the design of this major project.
Figure 2.4: Illustration of Rock Foundation of Burj Khalifa (Business Week, 2007)
One must not mix between rock foundations and the base stone that is found under porches of
some houses in America. These stones serve as a “raft” for the house porch structural wise.
See Figure 2.5. These stones might be rock foundation if the bearing stratum beneath them
was rock.
Figure 2.5: House Porch Foundation Stones (Brooks Stone Inc. 2011)
10
2.3. Software Usage in Rock Mechanics and Engineering
To facilitate the job for rock engineers and mechanics, many types of software have been
developed in the area of rock mechanics and engineering. These software programs, in a
personal point of view, are categorized in terms of the purpose of their usage. Some are used
as lab analysis software; others are used for specific types of rock-structural analysis, and
RocLab®, offered by RocScience Inc. (2013), is a simple lab analysis software program in
which small numbers of data are needed to calculate different parameters like Hoek-Brown
compressive strength, global strength and deformation modulus along with major and minor
Other programs offered by RocScience Inc. offer specialized analysis software for various
analysis software for rock structures like tunnels, caverns and other underground structures.
A more advanced finite element method software like TNO DIANA® offered by TNO BV
(2012) and Abacus® offered by Simulia Inc. (2012) are very advanced and general finite
element analysis software that can be utilized to solve different and complex rock structural
11
2.4. Brief about Sedimentary Rocks
In nature, rocks are categorized into three different types: igneous, metamorphic and
sedimentary rocks. It might be believed that the most dominant rock type is the sedimentary
rock as most people encounter this type in nature. But the fact is stated by Buchner, K and R.
Grapes (2011) which is; sedimentary rocks are only 8% of the total volume of the earth crust,
The best definition of sedimentary rocks is stated by Hamblin, K. and Christiansen, E.,
(2009) as follows “Sedimentary rocks are rocks that form from fragments derived from other
rocks and by precipitation from water.” These rocks are usually classified based on their
texture and composition into two categories, clastic rocks and chemical and biochemical
rocks.
A closer look at a clastic rock reveals the composition. Clastic rocks are formed from gravel,
sand and mud fragments. The word clastic comes from the Greek word “klastos”, which
means broken, and that implies the process of weathering and erosion of rocks, transportation
of fragments to deposition sites and finally the precipitation process where fragments fuse
The next Figure 2.6 shows clastic rocks' classification scheme which illustrates that clastic
rocks are classified into three categories based on their grain size. Conglomerates, shown in
Figure 2.7 (a), are those rocks with a grain size larger than 2 mm Rocks with grain sizes
varying between (1/16 – 2) mm are called sandstone, shown in Figure 2.7 (b). Lastly, rocks
with a grain size bellow 1/16 mm are named mudstones, shown in Figure 2.7 (c).
12
Figure 2.6: Clastic Rocks' Classification Scheme after Hamblin, K. and Christiansen, E., (2009)
On the other hand, chemical and biochemical rocks are formed from the chemical
precipitation or evaporation of salty lakes and shallow seas, from the growth process of some
organisms like coral and some types of algae, or from the decay of hydrocarbons in deep
sedimentary strata. Gypsum rock (CaSO4·2H2O), shown in Figure 2.7 (d), is a type of
chemical sedimentary rocks that results from the evaporation of shallow seas. An example of
chemical sedimentary rocks resulting from precipitation is Limestone, shown in Figure 2.7
(e). Coal is a sedimentary rock resulting from the decay of hydrocarbon content of organisms,
a b c
d e f
(a) Conglomerate (b) Sandstone (c) Mudstone (d) Gypsum rock (e) Limestone (f) Coal
Figure 2.7: Different Types Sedimentary of Rocks
13
2.5. Geological History of the United Arab Emirates and Its Rocks
Understanding the geological history of UAE is of utmost importance here, because it makes
the country, down from sabkhas to mid high elevation sand deserts to the Hajar mountains in
the eastern region of the country. Feulner (2005) conducted a research on the geological
history of the UAE and Hajar mountains. He concluded that the UAE is located in the corner
Figure 2.8: Arabian Plate and UAE Position (Dark Red) after Pierce (2002)
The UAE is highlighted in dark red. The Arabian plate is considered relatively sTable since
the Cambrian system of the Paleozoic era, about 520 million years ago, time scale-wise. The
Arabian plate includes, besides the Arabian Peninsula, the not true ocean basin, shallow
Arabian Gulf and the Zagros mountains. In the Chattian stage, about 25 million years ago, the
Arabian plate was disjointed from the Afro – Arabian continent to form the Red Sea.
Feulner also stated that the Precambrian history of the UAE can be known by reading the
Precambrian sediments of Saudi Arabia and Oman. It shows that the UAE has participated in
14
the late Precambrian glaciations. The UAE was often covered with shallow sea throughout its
history.
Feulner also stated that the movements of the Afro-Arabian plate during the Paleozoic caused
it to pass near the southern pole. In the Mid-Paleozoic, the Afro-Arabian continent was joined
to other continents to form the super continent of Gondwana, which began to breakup in the
Permian and Triassic periods. The UAE attained tropical and sub-tropical latitudes since the
Feulner also found that despite all the movements the UAE has made in its history; it still
appears to be tectonically sTable. The only exception is the formation of Al-Hajar Mountains
in the eastern region of the country. In general, the geological history of the UAE is just a
record of the advance and retreat of the sea in response to tectonic and climate changes
through time.
Ali, M.Y. etal. (2013) carried out valuable research on the seismic stratigraphy and
subsidence history of the UAE. In their work they presented a summary stratigraphic column
of the geological rock sequence in the UAE foreland basin. Although their research was
focused on oil bearing strata, the presentation way of their summary column is of great
interest. They found that UAE’s rocks are mainly carbonate rocks with small intercepts of
In their study they used locally used names for formations of the UAE with tectonic
interpretation regarding geological conditions and events that contributed to making the
subsurface history of the UAE in its known form. The following Figure 2.9 shows their work
15
Figure 2.9: Summary Stratigraphic of the UAE Foreland Basin , Ali, M.Y. etal (2013)
16
2.6. Rock Classification
design. It can be used as a check list for information or as an idea developer of strength and
deformation characteristics of rocks. Hoek (2006) did a good job in collecting different used
rock classification schemes in his book. The following is a brief exhibition of encountered
1946 uses descriptive bases to classify rocks. In this scheme, rocks are
seamy, crushed, squeezing and swelling rock. This scheme is used to estimate
rock loads which are carried by steel sets for tunnel designing purposes.
b. Classifications involving stand-up time: The stand-up time is the time for a
rock span being unsupported. The first researcher to propose the idea that rock
quality is related to stand-up time was Lauffer in 1958. For tunnels, the
unsupported span is the rock over the tunnel or between two supports. Some
modifications have been done on what Lauffer proposed by Pacher and others
in 1974 and are now part of the New Austrian Tunneling Method.
scheme developed by Deere and others in 1967. The RQD is defined as the
percentage of length of intact rock pieces longer than 100 mm in the total
length of the core. The drill bit for this classification should be of size NX (Φ
17
d. Rock structure rating (RSR): a quasi-quantitative scheme developed by
Wickham in 1972 to describe the quality of rock mass and to select the
quality of the rock mass which are the geological considerations, geometry of
proposed structure and effect of ground water inflow and joint conditions yield
1989. This classification deals with many factors affecting the rock mass and
criteria that contributes to the classification process. This scheme uses the
developed by Barton in 1974 for rock mass characteristics and tunnel support
scale from 10-3 to 103. This scheme, unlike others, is a very advanced
classification system that takes into consideration the effects of rock quality,
One has to be cautious when using rock classifications as there might be shortcomings of the
used scheme. Palmstorm and Broch (2006) worked on a thesis on the uses and misuses of the
18
Q rock classification scheme. They concluded that the shortcomings in the Q classification
scheme make it not recommended for use when it comes to calculating the penetration rate
(PR) and the advance rate (AR) for tunnel boring machines (TBM). They also strongly
equation. Furthermore, they quoted Terzaghi's statement in his last years “The geotechnical
engineer should apply theory and experimentation but temper them by putting them into the
2.7.1 Background
There are three types of compressive strength tests of rocks. The first is the unconfined
compressive strength (UCS) where only the axial load is applied to a rock sample and no
lateral loads of any type are applied, mathematically speaking (UCS1 > 0, UCS2 = UCS3 =
0). The second is the triaxial loading where not only axial loading is applied on the rock
sample, but also equal lateral loading is applied on the other two dimensions, mathematically
speaking (UCS1 > UCS2 = UCS3). The third is the true triaxial loading, similar to triaxial
loading but the difference being that lateral loads are not equal, mathematically speaking
(UCS1 > UCS2 > UCS3). The true triaxial loading is done using cubical load sample (Jaeger,
etal. (2007)). The following Figure (2.10) illustrates different compression types. (a) Implies
UCS, (b) implies triaxial loading and (c) implies true triaxial loading
19
Figure 2.10: Different Compression States of Rock by Jaegar, etal. (2007)
The most commonly used test is the UCS test as it is the easiest and less sophisticated among
all three compression test types. Other tests are needed if further understanding of rock
failure in semi-natural cases is required. But in general, rock triaxial and true triaxial are
seldom performed in the UAE. Another advantage of rock UCS test is the UCS value that is
This test is performed in accordance with the American Society of Testing and Materials
(ASTM) code number D2938 (2002) requirements. Although this code was withdrawn in the
year 2005 by the ASTM, the replacement code number ASTM-D7012 (2010), which is the
unconfined compressive strength (UCS) and modulus of elasticity (E) testing procedures,
specifies that the details of the testing procedure is acquired from the withdrawn code and
𝑃
𝜎 = 𝜋 𝑢2 ...................................................... (2.1)
4
𝑑
Where (Pu) is the ultimate load the sample can take and (d) is the sample’s diameter
20
2.7.2. Relations between UCS and Mechanical Properties
2.7.2.1 Relations between UCS and Point Load Strength Index (Is(50))
The point load strength index test is performed in accordance with the ASTM-D5731 (2008)
procedure and is meant to measure the rock’s point load strength index, which is a very
The following equation 2.2 was used to calculate the point load strength index Where (Pf) is
𝑃𝑓
𝐼𝑠 (50) = ………………………………. (2.2)
𝑑2
An interesting study was conducted by G. Tsiambaos and N. Sabatakakis (2004). The study
was about considerations on strength of intact sedimentary rocks, which aimed to find
correlations between point load strength index (Is(50)) with UCS (UCS) and Hoek-Brown
material constant (mi). In their study, sedimentary rocks from Greece where used. They
compared their work to the previous work of Bieniawski and the International Society of
Based on their data, they concluded that the relation between point load index and UCS could
be presented through three different models. The linear model was the first one shown in
Figure 2.11 as the bold line. This model gave an accepTable value of R2 of 0.75. They
concluded that their result is similar to the one found by Bieniawski and the International
Society of Rock Mechanics (ISRM). The power model was the second one shown also in
Figure 2.11 as the dashed curve. This model showed a better relationship as R2 was 0.82. The
classified linear model was the third one. They have observed that the point load index could
be categorized into three different classes (I, II and III). For each class a conversion factor
21
was assigned to multiply with the point load strength index value in order to get the UCS
value. The next Table 2.1 shows these classes which are shown in Figure 2.12.
Table 2.1: Relations for Different (Is(50)) Classes Tsiambaos and Sabatakakis (2004)
Class Is(50) Conversion Factor (UCS = Is(50) × ↓ )
I <2 13
II 2–5 20
III 5< 28
Figure 2.11: UCS – Is(50) Linear and Power Correlation by Tsiambaos and Sabatakakis (2004)
Figure 2.12: UCS – Is(50) Classified Correlation Tsiambaos, G. & Sabatakakis, N. (2004)
22
2.7.2.2. Relations between UCS and Schmidt Hammer Rebound Number
The Schmidt rebound hammer test is performed in accordance with the ASTM-D5873 (2013)
procedure and is meant to measure the rock’s surface hardness either in situ or in lab to give a
rapid indication about the rock’s strength. This test is best suited for rocks with UCS value
Faisal Shalabi and his colleagues carried out research about estimation of rock engineering
properties using hardness tests. The main idea was to estimate some important rock properties
such as UCS, modulus of elasticity and Poisson’s ratio using easier and cheaper methods
such as Schmidt hammer, shore scleroscope, abrasion, total hardness and unit weight. They
used dolomite, dolomitic limestone, shale, dolomitic marble, deopside and anhydrite from
different locations in California and New York as subject rocks for their study. All samples
were of size NX (54 mm diameter), which is the minimum requirement to perform a Schmidt
hammer test on rocks as per ASTM-D5873 (2013). They used standard practices to perform
required tests. They concluded that linear model could be used to estimate the UCS of
sedimentary rocks (Dolomite) from other properties such as Schmidt hammer rebound
23
Figure 2.13: UCS vs HR Relation by Shalabi, F. et al. (2007)
Another interesting paper was about correlation of Schmidt hardness with unconfined
compressive strength for gypsum from Sivas (Turkey) by Yilmaz and Sendir (2002). They
used the exponential model to express this relation. The value of R2 was as high as 0.96. The
24
2.7.2.3 Relations between UCS and Brazilian Splitting Tensile Strength
The Brazilian splitting tensile strength test is performed in accordance with the ASTM-
D3967 (2008) code and is meant to measure the rock’s splitting tensile strength. The code
states that rock engineers require the determination of complicated stress fields where a
combination of both compressive and tensile stresses are available. Furthermore, doing pure
tensile strength test is theoretically applicable but very hard to do on a practical level. This
test serves as an easy alternative to find this mechanical property of rocks. The following is
the equation used to calculate the strength, where (Pu) is the failure load, (L) is sample’s
2𝑃𝑢
𝑆𝑡 = ………………………………… (2.3)
𝜋𝐿𝐷
Nazir, R. etal. (2013) conducted a research regarding correlating UCS to the Brazilian
splitting strength of lime stone samples. Firstly, they collected different relations from recent
studies. The following Table 2.2 summarizes these relations. Here BST stands for St.
They also stated that one of the most agreed upon correlations is the one done by Sheorey,
where UCS equals 10 times the Brazilian splitting strength. They concluded that there is a
relation between St and UCS which is presented in Figure 2.15. They also compared
different previous work presented in Table 2.2. The following Figure 2.16 shows this
comparison.
25
Figure 2.15: UCS vs. St (BTS) found relation by Nazir, R. et al. (2013)
Figure 2.16: Comparison of Different Studies with lab Work by Nazir, R. et al. (2013)
This test is performed according to the ASTM-D2845 (2008) code and is meant to measure
the rock’s ultrasonic pulse velocity, which can be correlated to different important rock
properties like UCS, E and Poisson’s ratio (µ). It is important to mention that this test is not
meant to measure stress wave attenuation. The sound velocity (VP) is found by the following
equation;
𝑙
𝑉𝑝 = ………………………………….... (2.4)
𝑡
Where l = the length of specimen and t = is the time for a sound wave to move from the
26
A paper was written by Yasar and Erdogan (2004) regarding correlating sound velocity with
the density, compressive strength and Young’s modulus of carbonate rocks. The scope of the
work was to correlate density, UCS and Young’s Modulus (E) for carbonate rocks in
different areas in Middle Turkey (Adana). The linear model was used to present these
relations. In their study, three types of rocks were used; Dolomite, Marble and Limestone.
Correlating different rock types with each other is an interesting idea since Marble is a
metamorphic rock and other rocks were sedimentary. But from their point of view, they
considered all as carbonate rocks. They concluded that there is a good linear relation between
mean P-wave sound velocity (Vp) with UCS and E. The following Table 2.3 shows their
results and the following Figures 2.17 and 2.18 show these relations.
27
Figure 2.18: Vp – E Correlation by Yasar, E. & Erdogan, Y. (2004)
As stated previously, the ASTM merged the determination of the UCS and the modulus of
elasticity of rocks into one code starting from 2005. The code ASTM D-7012 (2010) is the
standardized procedure now to perform the modulus of elasticity E test. In this thesis, the
average modulus method was used to calculate the modulus of elasticity; which is the average
slope of the apparently straight line of the stress strain diagram. This is shown in the next
Figure 2.19.
28
Strain (ε) is calculated as per the following equation, where (δ) is the instantaneous
𝛿
𝜖= ……………………………………. (2.5)
𝐿
This test is done simultaneously with the UCS test. In fact, this test can be considered a
byproduct of the UCS test. Therefore it is logical to find the modulus of elasticity as a
function of the UCS. Tziallas, etal. (2009) did good research in correlating the UCS to E
through different models. They concluded that E can be determined as a function of UCS
with high R2 value equals 0.95. The following equation 2.6 is their concluded correlation;
They also concluded that E can be determined as a function of both UCS and the longitudinal
sound velocity Vp. The following equation 2.7 is their other concluded equation; where E is
0.541
𝐸 = 0.051(𝑈𝐶𝑆 ∗ 𝑉𝑝 ) − 0.27𝑈𝐶𝑆 …………………… (2.7)
They also visualized this correlation as a 3D diagram shown in Figure 2.20 and as a
29
Figure 2.20: 3D visualization of equation 2.7 by Tziallas (2009)
30
2.7.3. Relations between UCS and Physical Properties
The specific weight of rocks is determined according to the South Carolina department of
transportation code number SC-T-39 (2008). This procedure is considered one of the easiest
𝑊
𝛾= ……………………………………. (2.8)
𝑉
Where W is the sample weight and V is the sample volume. The methodology is to measure
the total weight and divide it by the total volume which counts for all voids in the specimen.
In any case, it is assumed that the denser the specimen, the stronger it will be. A study
confirms that was carried out by Moh’d, B. (2009). He also concluded that there is a good
relation between specific weight and UCS of a regression coefficient R2 as high as 0.9666.
Here he correlated density to UCS. In fact, the specific weight is nothing but the density
times the acceleration of gravity (g) therefore this fixed value can be included easily in the
correlation without jeopardizing the accuracy of the results. The following Figure 2.22 shows
31
2.7.3.2. Relations between UCS and Moisture Content
The moisture content is determined as per the ASTM-D2216 (2010) code. The procedure to
perform this test is described in the following chapter. The following equation 2.9 describes
𝑀𝑖 − 𝑀𝑜𝑣
𝑀𝐶 = × 100% …………………………. (2.9)
𝑀𝑜𝑣
Where Mi is the as received mass and Mov is the oven dry mass. The methodology is
summarized as measuring the masses of the rock specimen before and after placing it in the
oven for a certain time and temperature. It is believed that moisture content influences UCS.
32
Chapter 3
Experimental Program
3.1. Introduction
To achieve the proposed objectives of the study, lab and office works were conducted on a
set of samples from different locations in UAE. All Samples were acquired from the UAE's
coastal zone; the western region (W.R.), Abu Dhabi region (A.D.) and northern Emirates
region (N.E.). These samples were prepared and tested in accordance with the ASTM for all
tests except for the unit weight test where the code of South Carolina Department of
Transportation was used. The next Figure 3.1 shows the work plan flow. Every step of this
Sample Testing
Sample • Mechanical Tests
Transporting and • Physical Tests
Storing
Sample Identification
• Define Type and Depth of Sample
Each Sample
• Number The samples for Preperation
Ease of Access
In order to produce consistent and relaTable results, it was decided to have the sample span in
the core be of 1 meter length in order to ensure the uniformity of the tested samples. All cores
33
were of size HX (76.2 mm) to avoid misleading results in mechanical tests. The following
Table 3.1 shows in detail the core locations in UTM coordinates with a summary about them.
34
Figure 3.2 shows the contribution of samples by location and type where Figure 3.3 shows
12%
21%
34%
15%
73%
45%
(a) (b)
Figure 3.2: Sample Contributions (a) by Place, and (b) by Rock Type
Figure 3.3: Zone of Borehole Locations, Acquired from Google Maps (2013)
35
3.3. Sample Transporting and Storing
To ensure the integrity of the acquired cores, the ASTM-D5079 (2008) practice code was
used in transporting and preserving rock core samples. This code gives very detailed
procedures for samples from the time they are recovered from cores till the deposition of the
samples or storing for a certain period of time for future testing. The code provides a nice
flow chart for personnel in charge of all stages of core recovery, transporting, testing and
storage. The following Figure 3.4 shows this flow chart which is quoted with alteration from
the code.
1. Sample Recovery
2. Handling 3. Core Photography
From Drill
5. Sample Protection
•Routine Care 6. Preparation of
4. Initial Logging •Special Care Storage and Shipping
•Critical Care Containers
•Soil Like Care
9. Specimen
7. Transportation 8. Storage
Preparation
10. Testing
Figure 3.4: ASTM Flow Chart For Preserving and Transporting Rock Core Samples (2008)
36
Some geotechnical laboratories cooperated in granting all studied rock core specimens.
Therefore, due to their expertise, the chart steps 1 through 6 were assumed to be
professionally done. Steps 7 through 10 were done again in transporting samples from
laboratories to the university through testing of samples. Thankfully, all samples didn’t
require special transportation therefore a normal truck was used to transport samples.
Regarding storing of samples, the code specifies special measurements for storing if test
results are affected by storing conditions. In this study, only the moisture content test required
special storing measurements whereas all other tests required the normal rock core box
storing measurements.
The next step was specimen preparation for tests. This step was very important as it defined
which test came before the other. Before that, all samples were identified for ease of use and
access of data collected from various tests. Identification and preparation of samples are
Another important step is sample protection. Although this step can be considered as sample
required to determine how the sample processing should proceed. This identification means
extra protection measures for every class. Class one is the routine care which is for cores of
1.5 meters run and larger. If cores are less than 3 meters run, they are stored in structurally
sound core boxes. If they are longer than 3 meters run, they are placed in slightly wider and
longer PVC pipes. Here, core runs of 1 meter were considered the base for core box storing.
The second class is the special care class. It is necessary if the moisture condition is needed.
recommended. The third class of care is the critical care which is needed for sample
protection against shock, vibration and variations in temperature. The fourth and last class of
37
care is soil like care. Logically and by the code one can consider dealing with these cores as
soil. Thankfully, only class two of care was needed as a result of the as received situation of
the cores.
For ease of access, an identification system was established. Samples were given an ID of the
ID: XYZ-a-b
For example; sample ID (BY-5-08) means this sample is acquired with Baynunah labs, taken
from the 5th borehole acquired with that lab and 08 means this is the 8th sample acquired from
that borehole. It is good to mention here that the greater the sample number, the deeper the
sample. For tabulation purposes, the ID itself represents three columns of data; XY, a and b
data columns. This made it easy to store the sample depth and type in the database
established for this purpose. Here it is good to mention that the code assigned for the three
rock types is the following; CRGP is crystalline gypsum, MUDS is mudstone and SANDS is
sandstone.
38
3.5. Sample Preparation
Depending on the test, one can decide if the specimen requires preprocessing or not. Some
tests like the UCS and E tests require special preparation, whereas other tests don’t need any
preparation, such as the point load test. All samples were cut using the diamond rock core
cutter for the UCS and E, Brazilian splitting, and ultrasonic pulse velocity where the
geometry of the sample is required to perform these tests as explained in the next section.
In fact, only the UCS and E test needed special sample preparation. Since the height to
diameter ratio of the sample has to be 2:1 sharp with a very low margin of error. Moreover,
the two surfaces of the sample had to be parallel to each other with a very low margin of error
and vertical to the sample with, likewise, a very low margin of error. These margins and
checking methods are stated in the ASTM D-4543-08 code. This method involves three
checks of the sample, and these are; the deviation from straightness, flatness and
For straightness check, the sample was placed on a horizontal surface prepared for this check.
The sample was then rolled on that surface to check the straightness. Any sample that had a
gap more than 0.5 mm didn’t meet the straightness check. The check of flatness was done as
per procedure 5.2 B of the code. In this procedure, the sample was placed on the horizontal
surface. Then a dial gage of precision 2.5 µm was set in contact with the specimen. Readings
of three diameters were taken for the specimen. If the difference between maximum and
minimum readings of the diameter was less than 38 µm, the sample was accepted.
For perpendicularity check, the sample was placed on the horizontal surface again with a true
square being in contact with the specimen. Then the sample is rotated to find the maximum
gap between the specimen and the square. If the gap to the length ratio was less than 1:230
ASTM-D4543 (2008), section 5.2 requirements even with “best effort” as the code stated,
which was using a very sharp diamond cutter and dry cutting methodology operated by a very
experienced person. The code here gave a concession for this case as the code directed to cut
The following is a demonstration of the detailed procedures used to perform tests. Tests are
classified into mechanical tests; point load, Schmidt hammer, Brazilian splitting, ultrasonic
pulse velocity and UCS and E tests; as well as physical tests; moisture content, and specific
weight tests. Tests are presented in that order to make it easy for the reader to navigate
As stated previously in chapter 2, this test was done in accordance with ASTM D-5731-08.
This test was performed by subjecting a rock specimen to an increasing concentrated load
until splitting of the specimen. The concentrated load was applied through coaxial conical
platens. The failure load was used to calculate the point load strength index by equation 2.2
which was used to estimate the UCS. The Figure 3.6 shows a sample ready for the test. This
A qualifying sample was of length to diameter ratio of 1:1 or more (no preparation
needed)
The diameter of the sample was then recorded
Sample was then inserted into the machine and platens were closed to form contact
with the diameter of the specimen. Here the contact point had to be in the middle if
the sample
Sample was then subject to steadily increasing load until failure occurred
40
The failure load and pattern were recorded
If the failure pattern was the same as Figure 3.7 (a) the test was conducted perfectly
If the failure pattern was the same as Figure 3.7 (b) the test was rejected.
(a) (b)
Figure 3.7: Point Load Test Different Failure Patterns as Per ASTM-D5731 (2008)
(a) Accepted (b) Rejected
41
3.6.1.2. The Schmidt Hammer Test
As stated previously in chapter 2, this test was done in accordance with ASTM D-5873-13.
This test was performed by subjecting a rock specimen to shock load without resulting in the
failure of the specimen. The shock load was applied through a rebound hammer (Schmidt
hammer). The height of the plunger after the shock was then recorded to calculate the
rebound number which was used to estimate UCS. This test was done as follows;
As stated previously in chapter 2, this test was done in accordance with ASTM D-3967-08.
This test was performed by subjecting a rock specimen to an increasing concentrated load
until splitting of the specimen. The concentrated load was applied through coaxial flat
platens. The failure load was used to calculate the tensile strength of the sample by equation
2.3 which could be used to estimate the UCS. This test was done as follows;
Samples were prepared as per ASTM D-4543-08. The thickness to diameter ratio was
to be between 0.2 and 0.75. Thickness and diameter were recorded
Samples were then marked on their diameter to ensure proper positioning in the
loading machine
Then, samples were positioned in the loading frame
After that, samples were loaded until failure of samples as shown in Figure 3.9
The failure load was recorded for all samples and the strength was calculated for all
samples as per equation 2.3
42
Figure 3.8: Failed Sample after Brazilian Test
As stated previously in chapter 2, this test was done in accordance with ASTM D-2845-08.
This test was performed by subjecting a rock specimen to an ultrasonic pulse. This pulse was
applied by a transducer and received by a receiver. The machine records the time a pulse
needed to travel from the transducer to the receiver. Then the pulse velocity was calculated
by equation 2.4 which was then used to determine some parameters for the rock of which
Samples finely prepared as per ASTM D-4543-08. Length to diameter ratio was
recommended not to exceed 5 and at least 10 times the larger grain size. The length of
the sample was recorded
Samples were then marked for the place of transducer and receiver placement
Then, grease was applied to surface of samples, transducer and receiver to ensure no
air is entrapped between the apparatus and the sample
The sample was then subjected to ultrasonic pulse. The time of travel was recorded
for each sample
The pulse velocity was then calculated by equation 2.4
3.6.1.5. UCS and E Test
As stated previously in chapter 2, UCS and E are usually done simultaneously. This test
was done in accordance with ASTM D-7012-10 methods C and D. A rock core specimen was
cut to achieve an aspect ratio of 2:1. The ends were engineered. The specimen was placed in
a loading machine. Axial load was applied gradually and increasingly on the specimen.
Deformation was measured as a function of load until peak load and failure happened. Then
43
the UCS was calculated by equation 2.1 and E was calculated as per equation 2.5 and Figure
2.18. Figure 3.10 shows a sample ready for the test. This test was done as follows;
Samples finely prepared as per ASTM D-4543-08. Length to diameter ratio was to be
2:1 The length and diameter of sample were recorded
Samples were then placed in the loading machine connected to a computer to record
load and its corresponding deflection to construct the stress strain diagram
After that, samples were loaded until their failure
The extreme load was used to determine the UCS as per equation 2.1
The whole record was converted into stress strain diagram as per equations 2.1 and
2.5 and Figure 2.19. (E) was then calculated from the stress strain diagram
44
3.6.2. Physical Tests Done
As stated previously in chapter 2, this test was done in accordance with ASTM D-2216-10. A
test specimen was dried in an oven at a temperature of 110° ± 5°C to a constant mass. The
loss of mass due to drying is considered to be water. The water content is calculated using the
mass of water and the mass of the dry specimen. The moisture content is then calculated by
As stated previously in chapter 2, this test was done in accordance with SC T 39-08.
Specimens’ dimensions and weight were recorded and the specific weight was calculated as
A closer look at the specifications of performed tests and its influence on the sample in terms
of destructivity reveals many facts. To ensure the correctness of moisture content test results,
it was decided to perform it first, just after removing the sample cover. Therefore the stage
As mentioned earlier, sample preparation for other tests required accurate measurements and
cautious cutting of samples depending on their condition. This processing was required for
geometry related tests like unit weight, ultrasonic pulse velocity, UCS and E, and Brazilian
45
splitting strength tests. Also, the unit load test was the simplest test done so it was decided to
combine the unit weight test with sample preparation in one stage (stage two).
For the ultrasonic pulse velocity and The Schmidt hammer tests, it was obvious that no
sample destruction resulted. Therefore it was decided to group both tests in stage three. Stage
four was the last stage; all destructive tests were performed in this stage. It is good to mention
that the desired sample span of 1 meter facilitated the conduction of all previously mentioned
tests. Therefore, a sample segment was only tested once for stage three and four. The
Moisture
Stage One
Content
Schmidt
Sample Stage Three USPV
Hammer
Sample Unit
Stage Two Point Load
Preparation Weight
Brazilian
Splitting
46
Chapter 4
Results, Analysis and Discussion
4.1. Introduction
This chapter is dedicated to presenting, analyzing and discussing lab test results. Section 4.2
presents all lab tests results with some statistical analysis and relations to rock types that were
included in the study. Section 4.3 is dedicated to the correlation analysis, in order to meet
goals and objectives set for the thesis. Section 4.4 holds a comparison between this thesis’s
This section presents a summary of results for all experiments done, where statistical analysis
was performed to define 95% confidence intervals estimate of the expected value from each
Since the capacity of the used rebound hammer varies between 10 and 90, nearly 12% of
samples failed before the fulfillment of the test. Here, 210 samples passed this test. From a
statistical analysis to determine the confidence interval of the expected value based on 95 %
confidence, the results are given in the following Table 4.1. All data is in appendix Table A.1
47
4.2.1.2 The Point Load Strength Index Test.
In this test, the apparatus used was more sensitive, although apparatus’s gauges were
analogue clock type gauges, and knowing that this test doesn’t require further sample
processing, 419 samples were tested according to the point load test. Some samples just failed
This test is considered one of the most important tests as it is correlated directly to the UCS
of rock and is widely used for its ease of application. Nonetheless, the following Table 4.2
shows the result of the 95% confidence interval analysis for the expected value of test results.
Table 4.2: 95% Confidence Intervals Estimates of Expected Values Is(50) (MPa)
Rock Type Minimum Expected Maximum Expected
Any 0.468 0.582
Mudstone (MUDS) 0.469 0.633
Sandstone (SANDS) 0.419 0.608
Crystalline Gypsum (CRGP) 0.314 0.636
Although this test is simple to perform, the requirement of sample processing to obtain an
engineered surface cut the amount of samples tested to 183 samples. The next Table 4.3
displays the result of confidence interval analysis and appendix Table A.3 shows all test
results
48
4.2.1.4. The Brazilian Splitting Tensile Test
In this test, the geometric constraint of length to diameter ratio is smaller and less strict. But
this doesn’t mean that no sample pre-processing is required. In fact, all tests which take into
account the geometry of the sample have a specific sample processing. At any rate, 195
samples were tested according to the Brazilian splitting tensile strength test. The next Table
4.4 shows the 95% confidence intervals and appendix Table A.4 shows all results.
This test was very accurate, since the testing machine and sensors were all computer
controlled. However, the geometrical requirements for samples were very strict; like strict
length to diameter ratio, strict verticality of the sample and strict parallelism of ends
constraint. Therefore, the amount of samples tested was 419 samples. If this suggests
something, it implies that the weathering conditions of United Arab Emirates have affected
the rocks; therefore it was hard to recover samples for the unconfined test.
As per the BS, EN, ISO 14689-1:2003, the rock classification of results is of two categories.
Very weak rocks are rocks that have their UCS less than (5 MPa) whereas weak rocks have
their UCS between (5 MPa) and (25 MPa). Therefore it is obvious that all results are either
weak or very weak. The next Table 4.5 shows the result of the confidence interval analysis
49
Table 4.5: 95% Confidence Intervals Estimates of Expected Values of UCS (MPa)
Rock Type Minimum Expected Maximum Expected
Any 6.246 7.976
Mudstone (MUDS) 5.269 8.106
Sandstone (SANDS) 6.009 8.197
Crystalline Gypsum (CRGP) 6.053 11.051
Since UCS and E were done simultaneously as one test. The same strict geometrical
constraints apply here too. The following Table 4.6 shows the confidence interval analysis
results while appendix Table A.6 shows all test results. The Figure 4.1 shows a sample of
stress-strain diagrams of samples where a and b are sandstones, c is crystalline gypsum and d
a) Sandstone
50
b) Sandstone
c) Crystalline Gypsum
51
d) Mudstone
e) Mudstone
Figure 4.1: A Sample of Stress-Strain Diagrams of Samples
52
4.2.2. Physical Tests Results
The time needed to perform this test is approximately 24 hours if we consider the time the
samples spent in the oven. It doesn’t require any sample processing, but because of the long
time it takes to be done it was scheduled to be performed at the end of the testing. In any
case, 419 samples were tested here. The next Table 4.7 shows the result of the confidence
interval analysis while appendix Table A.7 shows all test results.
A total of 419 samples were tested here. The next Table 4.8 shows the results of the
confidence interval analysis while Table A.8 shows all test results.
53
4.3 Generated Relations
The enormous amount of data that resulted from conducting nearly 1630 tests made it easy to
do correlation analysis. It also made it easy to understand some special characteristics and
relationships between rock properties in the UAE, which was the main objective of this study.
The following is a demonstration of the generated correlations between UCS and other rock
properties.
In statistical analysis, especially in raw data as this work’s data, it is very helpful to do data
smoothing before starting the correlation analysis. Data smoothing helps in dealing with
outliers that might affect the goodness of fit of the generated model. Also, the least square
regression analysis sometimes gives very low value of R2, which seems to be a not good fit,
because this method counts for all records without distinguishing outliers from other data.
This becomes worse in equally weighted records case. Therefore it was decided to use the
moving average (MA) smoothing method, with least absolute residuals (LAR) robust fit
Since three rock types were encountered, if possible, a relation for each rock type was
generated in addition to the general relation. Therefore, based on rock type, relations can be
categorized into four categories. Also, based on BS EN ISO 14689-1 (2003) rocks have been
classified into weak rocks, for UCS between (5 MPa) to (25 MPa), and very weak rocks, for
UCS less than (5 MPa). Therefore, based on strength, rocks and relations can be classified
into three categories. Therefore, the result will be twelve different categories by combining
everything together. The next Table 4.9 shows the Cartesian product used to generate cases.
54
Table 4.9: Cartesian Product for Codes of Different Cases
All Strengths (A) Very Weak (V) Weak (W)
All Types (A) AA AV AW
Crystalline Gypsum (C) CA CV CW
Mudstone (M) MA MV MW
Sandstone (S) SA SV SW
Moreover, for the first relation, the AA case, different generated relations were presented,
and then the best relations were selected based on R2 value. Other generated relations would
be the result of this analysis only. This means, for any selected relation, 5 relations were
generated. This does not mean other equations for the general case are not useable, it means
the selected equation is highly recommended unless users prefer to use a simpler equation for
Based on data retrieved from both UCS and Schmidt hammer tests, 3 relations had been
generated. The presented relations are the best possible relations that could be found either
with or without the application of MA data smoothing and LAR robust regression.
For the AA case (the general case), five different possibilities were generated. It was found
that the exponential model was the best for the general case (AA) and the general weak case
(AW) whereas the power model was the best for the general very weak case. The following
Table 4.10 presents relation analysis for the general case (AA) while Table 4.11 shows the
other two found relations. Figures 4.2 to 4.4 show these found relations.
55
Table 4.10: Relations Between UCS and HR, General Case
Figure Case Equation: UCS = R2
0.295 HR + 3.009 0.49
0.4872 HR 0.27
AA 1.939 sqrt(HR) 0.45
1.428 (HR)0.612 0.47
4.2 4.06 exp (0.04 HR) 0.52
56
Figure 4.3: UCS vs. HR Relation, AV Case
57
4.3.1.2. UCS vs. Is(50)
Based on data retrieved from both UCS and point load tests, 10 relations had been generated.
The presented relations are the best possible relations that could be found either with or
without the application of MA data smoothing and LAR robust regression. The following
Table 4.12 presents the relation analysis for the general case (AA) while Table 4.13 shows
the other best found relations. Figures 4.5 to 4.14 show these relations.
Table 4.12: Relations Between UCS and Is(50) (MPa) General Case
Figure Case Equation: UCS = R2
7.977 Is(50) + 1.151 0.67
9.459 Is(50) 0.63
AA 8.306 sqrt(Is(50)) 0.62
4.5 9.191 (Is(50))0.75 0.68
2.67 exp(1.085 Is(50)) 0.6
For the simpler linear model (UCS = 9.459 Is(50)), the large difference between this model
and the proposed ISRM model of (UCS = 24 Is(50)) can be noted. It is obvious how the
58
Table 4.13: Other Relations Summary Between UCS and Is(50) (MPa)
Figure Case Equation: UCS = R2
4.6 AV 5.833 √Is(50) 0.73
4.7 AW 5.414 exp(0.57 Is(50)) 0.69
4.8 CA 11.08 Is(50) 0.69
4.9 CW 11.24 Is(50) 0.55
4.1 MA 6.050 √Is(50) 0.71
4.11 MV 5.953 √Is(50) 0.7
4.12 SA 7.701 Is(50) 0.72
4.13 SV 5.679 √Is(50) 0.8
4.14 SW 8.170 √Is(50) 0.44
R2 = 0.80
Based on data retrieved from both UCS and ultrasonic pulse velocity tests, only 1 relation had
been generated. Here, it was obvious that the power model gave the best fit. The following
Table 4.14 presents the relation analysis for the general case while Figure 4.15 shows this
relation.
63
Figure 4.15: UCS vs Vp Relation, General Case
Based on data retrieved from both UCS and Brazilian splitting tests, 6 relations had been
generated. Here, the power model and the exponential model dominated the relationship
between UCS and St. The following Table 4.15 presents the relation analysis for the general
case while Table 4.16 presents other found relation. Figures 4.16 to 4.21 show these relations.
64
Figure 4.16: UCS vs St Relation, General Case
An important set of relations that was generated is between modulus of elasticity and the
UCS. Here, 5 relations had been generated. In this case, the power and exponential models
proved to be the best choice for all generated relations. The following Table 4.17 presents
these found relations while Table 4.18 presents other generated relations. Figures 4.22 to 4.26
68
Table 4.18: Relations Summary Between E and UCS (MPa)
Figure Case Equation: E = R2
4.23 AV 285.2 exp(0.3641 UCS) 0.65
0.5741
4.24 AW 355.2 UCS 0.49
4.25 CA 609.9 exp(0.09 UCS) 0.88
4.26 MA 411.6 sqrt(UCS) 0.65
69
Figure 4.25: E vs. UCS Relation, CA Case
Other relations was found to be very useful in the field where simple tests were correlated
together to give the following relation. The UCS can be related to the point load strength
index, moisture content, specific weight and type of rock. Here, the type of rock was given a
sandstone and 3 is assigned to crystalline gypsum. The following relation is the UCS as a
function of point load strength index, specific weight and type of rocks. It was concluded that
a good relation between UCS and the independent variables. The value of R2 was equal to
The following relation is the UCS as a function of point load strength index, specific weight,
moisture content and type of rocks. It was concluded that a good relation between UCS and
the independent variables. Here, the value of R2 was 0.75.The next equation is the found
relation.
Here it was found that the type of rock has small influence on the result of the equation unlike
the previous equation. Therefore, the following equation is between UCS and the same
Another good relation was found between UCS and point load strength index, moisture
71
4.4 Comparison of Work with Previous Literature
sedimentation scenarios and in chemical and physical compositions of one rock type to
another. Even so, designers keep using relations from different places to determine some
required parameters. The following Figures present comparisons between different relations
plus those which were presented earlier with lab determined results. This comparison aims to
show how far some widely used relations are from the real rock properties it is to determine.
The next Figure 4.27 compares between the Schmidt hardness relations.
It is obvious that there are two competing models, the one presented in this thesis and another
which was done by Yilmaz et.al (2002). The proposed model by Shalabi (2006) doesn’t work
72
The next comparison between point load strength index and UCS is of great importance, as
many designers rely on the ASTM relation to predict the UCS. This comparison is to show
how far this relation is from reality about UCS of UAE rocks and promotes a better solution.
Here, it is clear that the ASTM relation diverges from the real value of the UCS where one
might claim that the work of T. and S. (2004) seems more suitable to be used. The answer to
this claim is: it is a fact that the presented model gives lower values for the UCS, but, having
lower values for a capacity factor (UCS) gives more conservative designs.
The next comparison is between works for the ultrasonic pulse velocity (USPV). The
following Figure 4.29 shows this comparison. This very sensitive parameter is not only
affected by the type of rock, but also the physical condition (cracks, voids, integrity, … etc.)
73
Figure 4.29: Comparison of Works for USPV
Here, the proposed model by Yassar and Erdogan (2004) fails to give true UCS results, since
UCS can never be a negative value. Again, the proposed relation gave conservative results for
The next relation compares works done by Nazir (2013) and the presented relation for the
Brazilian splitting tensile strength. Once more, the presented work maintained the trend and
gave more conservative results. The following Figure 4.30 shows that
74
The last comparison to be presented is between relations generated for the modulus of
elasticity. The model presented by Tziallas (2009) failed to give conservative results. The
importance of this parameter was to calculate the allowable settlement of structures resting on
rock. Therefore a more conservative model would be a good choice to perform settlement
Although all models are confidently conforming, this doesn’t mean the end of scientific
research in the field of rock engineering. These results are subjected to correctness and
falseness. The duty of the coming researchers is to prove the correctness or the falseness of
the proposed work. To insure that all proposed equations are used in the correct way they are
crated for.
75
Chapter 5
Summary, Conclusion and Final Recommendations
5.1 Summary
A laboratory study was conducted to develop a database for predicting the unconfined
compressive strength of sedimentary rocks. Large numbers of rock samples from different
sites in United Arab Emirates (UAE) were collected and tested for the development of this
database and evaluation of models. Reliable empirical relationships were developed for UCS
1. Schmidt Rebound
4. Brazilian Test
5. Modulus of Elasticity
6. Moisture Content
7. Unit Weight
Sixteen hundred and thirty tests were performed on various sedimentary rock types, from
various depths and areas in the UAE. Lab experiments were conducted to investigate and
discover possible relations between parameters of these rocks to predict the UCS of them.
Twenty nine relations were selected from more than 130 generated relations to relate all
rocks' parameters to each other and if applicable, subsets of samples based on their type and
strength were also related. Furthermore, general relationships were developed relating
unconfined compressive strength with unit weight, moisture content point load strength index
and type of rock. It is thus hoped that this research will serve civil and geotechnical engineers
in the UAE in making practical decisions at the stage of the preliminary site investigation
76
works and for determining the unconfined compressive strength of UAE rocks from quick
and cheap tests. The following Tables 5.1 and 5.2 show the general case relations
5.2 Conclusion
From different tests performed, including the Schmidt’s rebound hammer, point load strength
index, ultrasonic pulse velocity, UCS, E, Brazilian splitting strength, moisture content, and
bulk specific weight tests, all with conformance to the ASTM standard, it was concluded that
there are valid relations between UCS and rock physical and mechanical parameters of
previous literature and more reliable and conservative. For the sake of developing these
relations, different statistical methods were used in addition to the use of the least square
regression wherever that was required. These methods were the mean average data smoothing
MA to produce better relatable data sets, and least absolute residuals LAR to optimize found
relations. After that, a simple analysis was done based on the highest R2 value to determine
77
the best model to be used in representing the relation between UCS and the subject physical
or mechanical parameter.
As recommendations for future research and in light of the results reported in this thesis,
further research could be done to investigate the correctness and falseness of the work
presented here. Future researchers might be interested in studying a specific region in the
UAE. For although this country's size is small, the geotechnical and geological diversity
makes it a very interesting subject for research in this field, especially in rock engineering
which has its own influences on the petroleum wealth of the country.
78
References
1. Ali, M. Y. e. (2013). Seismic stratigraphy and subsidence history of the United Arab
Emirates (UAE) rifted margin and overlying foreland basins. In K. Al-Hosani etal
(Ed.), Lithosphere dynamics and sedimentary basins: The Arabian plate and
30609-9_6
2. Standard test method for unconfined compressive strength of intact rock core
3. Standard practice for preparing rock core specimens and determining dimensional and
4. Standard test method for determination of the point load strength index of rock and
5. Standard test method for laboratory determination of pulse velocities and ultrasonic
6. Standard test method for splitting tensile strength of intact rock core specimens,
7. Standard test method for compressive strength and elastic moduli of intact rock core
specimens under varying states of stress and temperatures, ASTM. D7012-10 (2010).
8. Standard test method for determination of pore volume and pore volume distribution
9. Standard test method for specific gravity and absorption of rock for erosion control,
10. Standard test methods for laboratory determination of water (moisture) content of soil
79
11. Standard test method for determination of rock hardness by rebound hammer method,
12. Brady, B. B., E. (2005). Rock mechanics and mining engineering, general concepts.
Rock mechanics for underground mining (Third Edition ed., pp. p. 1). United States of
13. Brooks Stone Inc. (2011). House porch foundation stones. Retrieved from
http://www.brooks-stone.com/photos/medium/4722.PNG
http://images.businessweek.com/ss/07/05/0501_makingof_burj/image/1_piles.jpg
17. Feulner, G. (2005). Geology. The emirates, A natural history (pp. 41-46). South
18. Hamblin, K. and Christiansen, E. (2009). Sedimentary rocks. Earth's dynamic systems
19. Hoek, E. (2006). The development of rock engineering. Practical rock engineering
20. Hoek, E. (2006). Rock mass classification. Practical rock engineering (pp. 3.1-3.23).
Ontario: RocScience.
21. Jaeger, J. (1979). The historical development of rock mechanics. Rock mechanics and
engineering (Second Edition ed., pp. 1-4). United Kingdom: Cambridge University
Press. doi:http://dx.doi.org/10.1017/CBO9780511735349.003
80
22. Jaeger, J. e. a. (2007). Deformation and failure of rocks. Fundumentals of rock
their and sound propagation . Jordan Journal of Earth and Environmental Sciences,
2(1), 18-25.
24. Nazir. R. etal. (2013). Correlation between unconfined compressive strength and
25. Palmstorm, A. and Broch, E. (2006). Use and misuse of rock mass classification
http://www.rocscience.com/products/14/RocLab
http://www.rocscience.com/usage/use/1/Tunnels
29. Rogers, J. (1995). Felton quarry granite slope failure, santa cruz, california.
http://www.3ds.com/products/simulia/portfolio/abaqus/latest-release/
33. South Carolina Department of Transportation. (2008). Method for determining the
81
35. Tsiambaos, G. & Sabatakakis, N. (2004). Considerations on strength of intact
36. Tziallas, G. P., etal. (2009). Determination of rock strength and deformability of intact
37. US. Army Corps of Engineers. (2003). Slope stability. (Manual No. EM 1110-2-
http://en.wikipedia.org/wiki/2010_Copiap%C3%B3_mining_accident
39. Yasar, E. & Erdogan, Y. (2004). Correlating sound velocity with the density,
40. Yilmaz, I. & Sendir, H. (2002). Correlation of schmidt hardness with unconfined
82