Journal of Rock Mechanics and Geotechnical Engineering.

2011, 3 (4): 373–384

Advance of in-situ stress measurement in China

Meifeng Cai1, Hua Peng2
1
University of Science and Technology Beijing, Beijing, 100083, China
2
Institute of Geomechanics, Chinese Academy of Geological Sciences, Beijing, 100037, China
Received 12 April 2011; received in revised form 14 September 2011; accepted 28 September 2011

Abstract: In-situ stress is an essential parameter for design and construction of most engineering projects that involve
excavation in rocks. Progress in in-situ stress measurement from the 1950s in China is briefly introduced. Stress relief by
overcoring technique and hydraulic fracturing technique are the two main techniques for in-situ stress measurement in China at
present. To make them suitable for application at great depth and to increase their measuring reliability and accuracy, a series of
techniques have been developed. Applications and achievements of in-situ stress measurement in Chinese rock engineering,
including mining, geotechnical and hydropower engineering, and earthquake prediction, are introduced. Suggestions for further
development of in-situ stress measurement are also proposed.
Key words: in-situ stress measurement; new development of techniques; engineering application

dam site of the Three Gorges Project [3].

1 Introduction
 of geological science in China, emphasized the
The knowledge of in-situ stress state is
indispensable for design and construction of most
engineering projects that involve excavation in rocks,
such as mining engineering, geotechnical engineering,
hydropower engineering, petroleum engineering, railway
and road engineering. The original stress in rocks is a
kind of natural stress existing in the earth’s crust. Its
formation is firstly related to tectonic movement,
especially the mutual pushing and extruding of the
tectonic plates, and secondly to the earth’s gravity.
Before excavation, the stress in the earth’s crust is in
equilibrium. The excavations in rocks break the stress
equilibrium and make the stress released. Just the
released stress is the basic force causing deformation
and damage of rock engineering [1, 2].
In-situ stress measurement in China started at the
end of the 1950s when the flat jack and local stress
relief methods were used to measure stresses at the
rock surface of tunnels, slopes, pillars and cavities in
hydropower projects and mines. The typical work was
completed at Pingshanba, the originally pre-selected
Doi: 10.3724/SP.J.1235.2011.00373

Corresponding author. Tel: +86-13701252928;
E-mail: caimeifeng@ustb.edu.cn
Professor Siguang Li (J. S. Lee), the major founder
function of horizontal tectonic stress for the first time.
In the 1920s, he pointed out that at a certain depth of
the earth’s crust, the horizontal stress is much larger
than the vertical stress [4]. He initiated stress
measurements for earthquake prediction at the
beginning of the 1960s. The first practical in-situ
stress measurement in mines with stress relief method
at local rock surface was completed by Professor
Zongji Chen (Tjongkie Tan), the founder of rock
mechanics and rock engineering in China, at Daye
open-pit iron mine in 1964 [5].
Stress relief by overcoring technique was used for
in-situ stress measurement in the late 1960s. Various
types of piezomagnetic stress gauges were developed
for the measurement at the early stage. Then,
“doorstopper” type strain gauges, borehole diametrical
deformation gauges and triaxial strain cells were used
[6, 7]. The in-situ stress measurement with nationwide
repercussions was made in Jinchuan nickel mine using
overcoring technique, which was conducted by
Institute of Geomechnics, Chinese Academy of
Geological Sciences, and Institute of Geology,
Chinese Academy of Sciences in 1975–1980 [8]. The
measurement revealed that the horizontal principal
374
stress was larger than the vertical principal stress at a
certain depth. The measured results led to
modification of original design of tunnels and
roadways, in which the orientation and sectional
shapes of the tunnels and roadways were in good
agreement with distribution of in-situ stress based on
the theory of elasticity. It successfully solved the
problems caused by unexpected deformations and
damages of tunnels and roadways.
Hydraulic fracturing technique was first introduced
to China at the end of the 1970s by F. Li and J. Ding.
They completed the first in-situ stress measurement
with the hydraulic fracturing technique at Yi County,
Hebei Province, China, for oilfield exploration.
During 1983–1987, they made a joint work cooperated
with M. D. Zoback and B. C. Haimson from USA to
complete hydraulic fracturing experiments at a 800-m
deep borehole and three 500-m deep boreholes to
investigate the tectonic stress field in western Yunnan
Province. Since then, this technique was widely used
in hydropower, hydraulic and petroleum engineering
[9].
Besides the aforementioned two main techniques of
stress relief by overcoring and hydraulic fracturing,
there are also several other techniques, such as
acoustic emission technique, differential strain curve
analysis, sonic wave anisotropy measurement and
borehole wall breakout measurement techniques,
which are used as supplements to the two main
techniques for different engineering fields. They are
relatively simple and easy to use in usual conditions.
However, their measuring accuracy is lower than those
of the two main techniques and roughly suitable for
primary estimation of in-situ stress state.
The importance of in-situ stress measurement was
gradually recognized by more and more engineers and
administration workers from the beginning of the
1980s in China. Large-scale and elaborately organized
in-situ stress measurements were completed in all
major rock engineering projects in China, such as the
Three Gorges Project, South-to-North Water Transfer
Project, Qinghai-Tibet Plateau Railway Project. The
measured results provided solid basis for optimal
design and construction of the projects.
During the later 20 years, new ideas and techniques
of in-situ stress measurement were developed due to
the requirements of various engineering projects,
which were developed towards larger scale, deeper
position and more complicated geological and
environmental conditions.
Meifeng Cai et al. / J Rock Mech Geotech Eng. 2011, 3 (4): 373–384
2 New development of stress relief
by overcoring technique
2.1 Advantages and problems of stress relief by
overcoring technique
Stress relief by overcoring technique has been used
for in-situ stress measurement for a long period of
time since it was firstly adopted by N. Hast to measure
in-situ stress state in Scandinavian Peninsula in the
1950s. A rough estimation showed that 70%–80% of
in-situ stress measurement data in the world were
obtained by using this technique. Stress relief by
overcoring technique is especially suitable for in-situ
stress measurement in mining and geotechnical
engineering, because in such engineering projects,
many entrances, such as shafts, tunnels, inclines and
roadways can provide accessing ways to the position
where the stress measurement is conducted. Therefore,
a great cost can be saved for driving a special
thoroughfare to the measuring position [10].
In the 1980s, systematic tests were completed by
Cai and Thomas [11] to check the performance of the
main overcoring stress measurement devices,
including USBM deformation gauge, CSIR
doorstopper, CSIR triaxial strain cell, CSIRO hollow
inclusion strain gauge and UNSW solid including cell.
Among these devices, the hollow inclusion strain
gauge is most commonly used in various rock
conditions. There are two main problems influencing
the reliability and accuracy of the overcoring stress
measurement: (1) most of the overcoring stress
measurement devices use resistance strain gauges as
the sensing elements, and the measured strains
induced by stress relief during overcoring are
interpreted as the in-situ stress based on the known
relationship between stress and strain of the rocks. The
measured strains are recorded by a Wheatstone bridge
type indicator. Because the resistance strain gauges are
sensitive to temperature variation, the additional strain
induced by temperature variation will induce a large
error in the measured stresses. Therefore, the
reasonable temperature compensation is critical for
validity and accuracy of the measurement. (2)
Traditional methods to calculate rock stress according
to the measured strain values are based on the theory
of elasticity, which assumes that the rock is linearly
elastic, continuous and isotropic. However, field rock
masses have characteristics of nonlinearity,
discontinuity and anisotropy to some extent. The
difference between the practical and assumed rock
Meifeng Cai et al. / J Rock Mech Geotech Eng. 2011, 3 (4): 373–384
conditions will cause errors in the calculated stresses.
2.2 Full temperature compensation technique
Traditional temperature compensation methods use
dummy gauges as compensation elements, which are
not effective for devices bonded to rocks during the
measurement, such as the hollow inclusion strain
gauge. According to the principle of Wheatstone
bridge, for effective compensation, the strains
measured by the dummy gauges and the working
gauges during temperature change should be the same.
However, the working strain gauges are bonded to
rocks, and their deformation behaviors are mainly
controlled by the rocks. In order to obtain the same
strain, the dummy gauges must be kept at the same
position and also bonded to rocks as the working
gauges, which is not allowed because the dummy
gauges should be stress-free. To address this problem,
a full temperature compensation technique was
developed by Cai [12]. It consists of the following 4
main significances:
(1) Resistance elements in the Wheatstone bridge
type indicator are all of low temperature coefficient
except for the working strain gauges at the measuring
position, which ensures that no additional strain
induced by temperature change is recorded by the
indicator.
(2) Temperature changes at the measuring points are
continuously recorded by a thermistor during
overcoring.
(3) After the completion of the overcoring test, the
overcoring with the working strain gauges still inside
is put in a temperature-controllable oven to make a
temperature calibration test, conducted at temperature
difference of 1 C, from which the thermal strain rate
for each strain gauge is obtained.
(4) Lead wires of the strain gauges can also induce
remarkable additional strains due to temperature
change. To solve this problem, the lead wires with the
same length and same type coming from the same
measuring point as the working strain gauges are
connected to a neighboring arm of the strain gauges in
the bridge, which neglects the thermal effect of the
lead wires.
From the calibrated thermal strain rate and recorded
temperature change during overcoring, the additional
thermal strain values for each strain gauge can be
determined and then eliminated from the totally
measured strains to get the correct strain values for
stress calculation.
2.3 Techniques accounting for the effects of
nonlinearity, anisotropy and discontinuity of rock
masses on calculation of in-situ stress
375
To solve the second problem mentioned above,
some techniques have been developed to interpret the
measured strain caused by stress relief to original
stress state considering nonlinearity, discontinuity and
anisotropy of the practical rock masses [13].
(1) Interpreting the measured strain to in-situ stress
state needs the value of deformation modulus of the
rocks. For nonlinear elastic rocks, the value of the
deformation modulus depends on stress level. To
ensure correct interpretation, the value of deformation
modulus should exactly correspond to its stress level.
Because both stress and deformation modulus are
unknown, an iterative program is used for the
interpretation. Before calculation, the nonlinear
relationship between deformation modulus and stress
should be determined and inserted into the iterative
program.
(2) For orthotropic and transversely isotropic rocks,
a method to determine anisotropic parameters with
biaxial loading test of overcore has been developed,
which makes it possible to calculate rock stress from
the measured strains based on the constitutive
equations of the rocks.
(3) The results of the biaxial loading test of
overcoring can also be used to determine anisotropic
factor of strain measured by each strain gauge. The
ratio of strain measured by each strain gauge to the
average strain value in the same direction is defined as
the anisotropic factor of strain measured by each strain
gauge, which is used to modify the measured strain
values during overcoring to get correct strain values
for in-situ stress calculation. If the rocks are perfectly
isotropic, the strains measured by strain gauges in the
same direction should be equal under biaxial loading.
(4) The performance of rock stress measurement
devices in various rock conditions with different
distributions can be evaluated and extents of
anisotropy and discontinuity through systematical
laboratory modeling tests can be determined. Based on
the modeling test results, the in-situ stress
measurement result is modified according to the real
rock condition.
(5) Numerical modeling and iterative methods can
be used to account for the effects of nonlinearity,
anisotropy and discontinuity on the measured strains,
which make the calculated results of rock stress close
to the real situation step by step.
The full temperature compensation technique and
techniques accounting for the effects of nonlinearity,
anisotropy and discontinuity have greatly improved
the reliability and accuracy of the overcoring stress
measurement.
376
2.4 Techniques for effective and convenient use of
overcoring technique at great depth
2.4.1 Wireless automatic data transmission and
collection techniques
The measuring depth of commonly used overcoring
stress measurement devices, such as the hollow
inclusion strain cell, is limited by their wire-dependent
data transmission technique. To solve this problem, a
series of wireless automatic data collection techniques
with overcoring stress measurement devices were
developed to make them effective and convenient at
great depth [14, 15]. A typical device is shown in Fig.1.
The main body of the device is a hollow inclusion
strain sensor, which is connected with an automatic
data collection and recording unit. The unit is
wireless-remotely controlled, so the device can be
reliably and conveniently used at any depth for stress
measurement.
Fig.1 A hollow inclusion strain cell with automatic data
collection.
2.4.2 Some local stress relief techniques able to be
used at great depth without full overcoring
The process of stress relief by full overcoring is
somewhat complicated. To ensure successful
measurements, a relatively integrated rock condition is
needed, which is difficult to be satisfied in some
places. To solve this problem, some local stress relief
techniques without full overcoring were developed,
which could be used at great depth. Among them, the
local borehole-wall stress relief technique without full
overcoring developed by Ge and Hou [16] is
especially useful, which is shown in Fig.2. Its
measuring principle is similar to those of triaxial strain
cell and hollow inclusion cell. Three strain gauge
rosettes are cemented on the borehole wall distributed
in the same circumference with an interval of 120.
The rock mass around the strain gauge rosettes is
stress released by local overcoring one by one. From
the measured 9 strain values induced by stress relief,
the 3D in-situ stress state at the measuring point can
be determined.
The new technique has remarkable advantages of no
need of full overcoring and no limit of borehole depth.
Meifeng Cai et al. / J Rock Mech Geotech Eng. 2011, 3 (4): 373–384

45
 45
Fig.2 Measuring layout of the local borehole-wall stress relief
technique.
Therefore, it can be used in slightly fractured rock
mass condition and may provide an alternate approach
for in-situ stress measurement at great depth.
Application of drilling robot and new automatic
recording system will make it more popular.
3 New development of hydraulic
fracturing technique
3.1 Advantages and problems of hydraulic
fracturing technique
Hydraulic fracturing technique is an efficient
technique for in-situ stress measurement at depth. It
has been mainly used for in-situ stress measurement in
hydropower engineering, highway and railway
engineering, etc., but less used in mining engineering
in China. However, in recent years, it was used in
mining engineering for in-situ stress estimation at the
exploration stage of the mine. It is an important
progress for mine administrators to get the information
of in-situ stress state before design and construction of
the mine. At the exploration stage, there is no any
entrance to underground positions for stress
measurement with overcoring and other techniques.
Therefore, hydraulic fracturing technique is a
relatively economical and convenient technique for
detecting the underground stress state of the mine,
because it can use the readily drilled boreholes to
conduct the hydraulic fracturing tests, which saves
costs for drilling special entrances if using overcoring
and other techniques. Hydraulic fracturing technique
has also been effectively used in deep-concave
open-pit mines. The design of open-pit mines is
traditionally based on limit equilibrium analysis,
which neglects the influence of in-situ stress on
Meifeng Cai et al. / J Rock Mech Geotech Eng. 2011, 3 (4): 373–384
stability of the mine. But for design of deep-concave
open-pit mines, overall mechanical analysis with
numerical modeling or other techniques is necessary
and information of in-situ stress is indispensible.
The hydraulic fracturing technique has two basic
disadvantages: (1) It is a 2D measuring technique. In
order to get 3D in-situ stress state, measurement at
three boreholes that cross the same measuring point
and are not parallel to each other is necessary. This
condition is difficult to be satisfied. (2) It is assumed
that the fracture in the rock borehole wall is initiated at
the minimum tangential stress position. If there are
original fractures, the assumption may not be valid.
Although the hydraulic fracturing technique is
efficient for in-situ stress measurement at great depth,
the maximum measuring depth of this technique in
China was only 800 m before the measurement in
Wanfu coal mine in 2004. The traditional hydraulic
fracturing equipment used at great depth has the
following three bottlenecks:
(1) The pressure-enduring abilities of the sealing
packers and pressurizing system of the hydraulic
fracturing equipment are not good. Because at depth
over 1 000 m, the pressure provided by the pressurized
water should be high enough to make the borehole
wall fractured. However, the high pressurized water
will also make the sealing packers and water pipelines
damaged or malfunction.
(2) With the increase in the measuring depth, the
groundwater pressure is increased. In many
engineering fields, such as in Wanfu coal mine
situated in the alluvium of the Yellow River, the soil
layers with thickness greater than 700 m will make the
borehole filled with slurry at the measuring point. It
will strongly influence the sufficient pressure relief
and removal of the sealing packers after completion of
the fracturing test.
(3) The traditional hydraulic fracturing equipment
uses double-loop system. Because the borehole is
more than 1 000 m deep and will cross soil layers with
thickness more than 700 m, the hydraulic fracturing
equipment will suffer removal difficulty because it is
stuck on the borehole wall during its transferring from
the lower to upper positions.
3.2 Improvement of the hydraulic fracturing
technique at great depth
To increase the pressure-enduring ability, removal
flexibility and measuring accuracy of the traditional
hydraulic fracturing technique with equipment, the
following five techniques were developed and used in
the Wanfu coal mine by the authors in 2004 [17]:
377
(1) A new type of sealing packer with special
structure was developed, whose pressure-enduring
capacity is 70 MPa, enough for hydraulic fracturing
test at depth of 2 000 m.
(2) A single-loop hydraulic fracturing system was
developed, in which a push-pull switch with high
strength was used for transferring the pressurizing
lines to the sealing packers or the sealed section of the
borehole.
(3) An automatic valve for low-pressure relief of the
sealing packer was developed, which could
automatically make pressure relief of the sealing
packers after completion of the hydraulic fracturing
test.
(4) Two sets of pressure monitoring equipments
were used to detect water pressure in the pipeline,
ensuring the measuring accuracy of water pressure
applied to the sealed section of the borehole.
(5) The loading and unloading process of pressure
was automatically program-controlled, which
eliminated the influence of manual control in the
traditional hydraulic fracturing system on the
measuring results.
With the improved equipments and techniques,
in-situ stress measurement at depth more than 1 000 m
has been successful completed in the Wanfu coal mine.
The maximum depth of the measuring point was 1 105 m,
which created a new record of measuring depth for
in-situ stress measurement using the hydraulic
fracturing technique in China [18]. The old record hit
800 m in 1983.
3.3 Measuring results in the Wanfu coal mine
The Wanfu coal mine is a part of the Juye coal field
located in southwest of Shandong Province, China.
The Juye coal field is within the range of the alluvium
of the Yellow River, and the overburden soil layer is
726 m thick on average. The average overburden
depth of the coal seam is 1 104 m in the Wanfu coal
mine. The complicated geological conditions and high
level of rock stress at depth more than 1 000 m will
strongly influence the stability of mining structures
and safety of mining operation. To provide reliable
engineering-geological information for design and
construction of the mine, in-situ stress measurement
should be carried out before the design and
construction of the mine.
To save measuring cost, the hydraulic fracturing
stress measurement was performed in the exploration
boreholes. There were 29 exploration boreholes in the
mine, of which 7 boreholes were meticulously chosen
for hydraulic fracturing tests, as shown in Fig.3.
378 Meifeng Cai et al. / J Rock Mech Geotech Eng. 2011, 3 (4): 373–384

stress, are shown in Table 1, where Pi, Ps and Pr are

Fig.3 Location of measuring boreholes (unit: m).
Successful hydraulic fracturing tests were carried
out at 3–6 measuring points in each borehole and
altogether 37 hydraulic fracturing tests were
completed in 7 boreholes. The measured results,
including hydraulic fracturing parameters and in-situ
the fracture initial pressure, shut-in pressure and
fracture re-open pressure, respectively; t is the tensile
strength of rocks at measuring point; P0 is the
groundwater pressure at the measuring level; H and
h are the maximum and the minor horizontal
principal stresses, respectively, which are calculated
from the measured fracturing parameters in a set of Pi,
Ps, P0 and T or the other set of Pr, Ps and P0; and v is
the vertical principal stress, which equals the gravity
stress (0.02 MPa/m) in the Wanfu coal mine.
3.4 Further improved techniques and spot test
results in Xinjiang oilfield
After the measurement in the Wanfu coal mine, the
hydraulic fracturing equipments and techniques were
further improved [19]. A series of newly invented or
developed equipments were used. The diameter of the
sealing packers is increased from 130 to 230 mm,
made of specially developed high-strength material.
Several sealing packers can be serially connected to
increase their sealing pressures. An ultrasonic
scanning probe is used to detect fractures in the
borehole wall, which can determine fractures less than

Table 1 Results of in-situ stress measurement in Wanfu coal mine.

Measuring Fracturing parameters (MPa) Stresses (MPa) Fracture
Borehole
depth (m) Pi Pr Ps P0 t H h v orientation
891.2 27.55 21.74 20.40 8.91 5.81 30.55 20.40 18.64 NE65.3°
960.2 25.23 25.02 24.38 9.60 0.21 38.52 24.38 20.46 —
#1 1 011.3 26.29 26.07 24.91 10.11 0.22 38.55 24.91 21.81 —
1 046.2 28.61 25.33 24.28 10.46 3.28 37.05 24.28 22.74 NE78.7°
1 063.0 30.62 27.87 26.60 10.63 2.75 41.30 26.60 23.19 —
1 104.9 33.27 29.47 27.67 11.05 3.80 42.49 27.67 24.30 NE62.9°
890.3 26.93 24.92 23.55 8.90 2.01 36.83 23.55 18.47 NE76.8°
976.6 28.09 25.97 24.18 9.76 2.12 36.81 24.18 20.75 —
1 025.1 28.62 26.82 24.50 10.25 1.80 36.43 24.50 22.05 NE63.4°
#2
1 056.4 33.17 27.67 25.55 10.56 5.50 38.42 25.55 22.87 —
1 074.6 32.43 25.66 24.50 10.75 6.77 37.09 24.50 23.37 —
1 079.6 35.17 30.60 25.98 10.80 4.57 36.54 25.98 23.50 NE68.5°
791.0 33.00 28.40 21.60 7.91 4.60 28.49 21.60 15.75 NE46.1°
859.0 40.90 35.30 26.00 8.59 5.60 34.11 26.00 17.55 NE44.7°
#3 910.0 41.50 37.60 26.50 9.10 3.90 32.80 26.50 18.90 —
1 012.0 42.80 39.10 27.60 10.12 3.70 33.58 27.60 21.86 —
1 063.0 38.40 36.20 29.00 10.63 2.20 40.17 29.00 22.95 NE52.4°
799.6 40.80 38.30 32.80 8.00 2.50 52.10 32.80 16.16 NE36.0°
900.4 43.70 37.60 34.30 9.00 6.10 56.30 34.30 18.81 —
#4 988.6 43.10 37.70 35.20 9.89 5.40 58.01 35.20 21.16 —
1 026.4 42.70 33.80 32.70 10.26 8.90 54.04 32.70 22.14 NE39.2°
1 052.7 40.90 35.70 33.90 10.53 5.20 55.47 33.90 22.85 NE43.6°
874.7 36.20 31.80 29.10 8.75 4.40 46.75 29.10 18.08 NE45.5°
1 011.7 35.60 30.10 28.70 10.12 5.50 45.88 28.70 21.70 —
#5
1 027.9 37.20 30.70 29.70 10.28 6.50 48.12 29.70 22.12 NE67.9°
1 080.4 30.40 26.80 24.80 10.80 3.60 36.80 24.80 23.50 NE52.8°
813.8 29.62 26.92 20.47 8.14 2.70 26.35 20.47 16.53 NE98.1°
843.8 31.94 30.38 26.89 8.44 1.56 41.85 26.89 17.32 NE110.6°
#6 861.5 35.95 30.25 26.63 8.61 5.70 41.03 26.63 17.77 —
879.3 37.89 35.17 28.05 8.79 2.72 40.19 28.05 18.25 —
891.8 45.78 37.63 30.51 8.93 8.15 44.97 30.51 18.59 NE92.5°
891.2 24.97 22.10 20.34 8.55 2.87 30.37 20.34 18.57 NE38.5°
960.4 28.10 26.88 21.74 8.94 1.22 29.40 21.74 20.39 —
1 011.3 28.32 26.05 23.65 9.13 2.27 35.77 23.65 21.74 —
#7
1 046.2 27.84 26.40 24.85 10.42 1.44 37.73 24.85 22.67 NE32.6°
1 063.0 29.76 28.56 27.72 10.68 1.20 43.92 27.72 23.12 —
1 105.0 37.29 32.75 29.99 10.95 4.54 46.27 29.99 24.23 NE39.0°
Meifeng Cai et al. / J Rock Mech Geotech Eng. 2011, 3 (4): 373–384 379

0.1 mm in width. The high-pressurizing system with a
specially developed pump, the automatic controlling
with data collection system and the sealing packer are
integrally connected and sent to the measuring points
in the borehole with a armored cable, as shown in
Fig.4. The cable is also used as remote controlling
signal line to the automatic controlling with data
collection system at the measuring point.
Measuring section
hydraulic fracturing technique in a single borehole
As mentioned above, the traditional hydraulic
fracturing technique is basically a 2D measuring
technique. To get 3D in-situ stress state, measurement
in 3 non-parallel boreholes is necessary. There are
three problems to be solved for such measurement: (1)
Directional drilling of inclined boreholes is difficult
and drilling length is strictly limited. (2) It is
impossible or difficult to make the 3 non-parallel
boreholes cross the same point, i.e. the fracturing
measurement point. (3) High cost and long time are
needed. For deep in-situ stress measurement, such a
technique is even more useless.
To solve the problems, a 3D stress measurement
method with hydraulic fracturing technique in a single
borehole was developed by Liu [20]. In the method,
besides a normal hydraulic fracturing test, fracture
re-opening tests in at least 3 natural fractures in the
borehole wall near the measuring point were added.
Some other researchers also presented similar designs

Fig.4 A new type of hydraulic fracturing stress measurement
system.
With the further improvement of hydraulic fracturing
equipments and techniques, a field experiment was
successfully completed in an exploration borehole
named Caican #1 in the Xinjiang oilfield in 2010. The
experimental measuring results are shown in Table 2.
The maximum measuring depth reached 2 800 m,
which created a new record of measuring depth with
hydraulic fracturing technique again in China.
3.5 Techniques for 3D stress measurement with
[21].
The method is based on the principle that the 3D
stress state at a point is composed of 6 stress
components, which are taken as 6 unknowns. To get
the 6 unknowns, firstly, a normal hydraulic fracturing
test is performed, which can provide 3 independent
equations. Secondly, fracture re-opening tests in 3
natural fractures in the borehole wall are conducted.
One fracture re-opening test can provide one
independent equation, thus 3 natural fracture
re-opening tests can provide 3 independent equations.
Then, 6 independent equations are used to slove the 6
unknowns and the 3D stress state is finally determined.
This method enlarges the usability of the hydraulic
fracturing technique. To make it successful, finding 3
natural fractures closely near the normal hydraulic
fracturing section in the borehole is the key. If the
natural fractures are far from the fracturing section of
the normal hydraulic fracturing test, this method will
be less useful for calculation of stress state at the same
measuring point. Correct information of the natural
fractures’ attitude is also important. It has been found
that in many cases, especially in the field with high

Table 2 Measuring results in Caican #1 exploration borehole.

Fracturing parameters (MPa) Stresses (MPa) Fracture
Measuring section Depth (m)
Pi Pr Ps P0 t H h v orientation
#1 2 750 55.10 52.07 50.92 27.50 3.02 73.70 50.92 68.75 NE30
#2 2 765 56.52 52.08 49.82 27.65 4.44 69.73 49.82 69.13 NE21
#3 2 785 58.97 53.72 51.41 27.85 5.25 72.66 51.41 69.63 NE10
#4 2 800 59.26 52.78 51.69 28.00 6.48 74.29 51.69 70.00 NW12
380
in-situ stress, initial fracturing is not along the natural
fracture if its orientation is far apart the orientation of
the maximum principal stress, which means failure of
fracture re-opening test in the natural fracture.
4 Application of in-situ stress
measurement to rock engineering and
earthquake prediction
4.1 Application in mining, geotechnical and other
similar rock engineering
Mining, geotechnical and other similar rock
engineering, such as hydropower, railway and road
engineering projects, all involve rock excavation and
construction. Rock stress exists in the surrounding
rocks of the engineering structures prior to and during
the whole procedure of excavation and operation.
Therefore, it is necessary to understand in-situ stress
state for guiding design, construction and operation of
rock engineering, especially in the following 5 aspects
[22, 23].
4.1.1 Overall layout of engineering structures
In primary stage of engineering design, clear
understanding of in-situ stress condition is a
prerequisite for design of the overall layout of the
engineering. In the design, important engineering
facilities, such as main chambers, shafts, tunnels and
transport roadways, should be kept away from the
high stress areas. Orientation of the above-mentioned
excavation structures should be in or along the
direction of the maximum principal stress, because in
such layout, there are two smaller principal stresses in
the vertical section of the excavations, which is
favorable for stability of the excavation structures.
To make the best of overall layout of the
engineering structures, in-situ stress measurement
before engineering design is necessary. Since there is
not any existing underground way to access the
measuring position before construction of the
engineering, hydraulic fracturing is the relatively
suitable technique for the measurement. Stress
measurement with the improved hydraulic fracturing
technique in the Wanfu coal mine is the first time to
make large-scale in-situ stress measurement in the
early exploration stage in Chinese mines.
4.1.2 Selection of the optimal shape of underground
tunnels and roadways
According to the theory of elasticity, to minimize
stress concentration around the tunnels and roadways,
their ideal sectional shapes are an ellipse, in which the
ratio of horizontal to vertical axes matches the ratio of
horizontal to vertical principal stresses in the section.
Meifeng Cai et al. / J Rock Mech Geotech Eng. 2011, 3 (4): 373–384
In such a condition, the boundary of the tunnels and
roadways will be in an even-compressed stress state,
which means that the values of tangential compressive
stress at every point of the boundary are equal. It is
very favorable for stability of the tunnels, roadways
and rock excavations.
Jinchuan nickel mine is the largest nickel mine in
China, whose nickel production occupies 90% of the
nickel production of the whole country. This mine
began construction in the mid-1960s, but until the
mid-1970s, it could not be put into operation. The
main problem was that the transport tunnels and
mining roadways could not be stabilized because large
deformation or destabilization often appeared in a few
months after completion of excavation. In-situ stress
measurement was carried out for a few years in the
late 1970s. The results showed that the horizontal
stress was remarkably larger than the vertical stress in
the mine. The sectional shape of the original tunnels
and roadways is shown in Fig.5, with height greater
than width. Based on the stress measurement results,
the sectional shape of the tunnels and roadways was
changed. The changed section has a larger width than
its height. After that, with improved shotcrete-rock
bolts support, the stability problem of the tunnels and
roadways was addressed quickly and subsequently the
mine was put into production in a short period of time.
Original shape Changed shape
Fig.5 The original and changed shapes of the roadways in the
Jinchuan mine.
4.1.3 Optimization of excavation sequences and steps
of the engineering
Excavation in rock engineering is usually a
complicated procedure. Different excavation
sequences and steps will cause different mechanical
effects, i.e. different stability status of the engineering.
Because the structural shapes and excavation steps of
the most rock engineering projects are very
complicated, it is impossible to implement engineering
design, including selection of excavation methods and
arrangement of excavation steps by theoretical
calculation. Therefore, the traditional excavation
design mainly relies on the experts experience and is
less reasonable and reliable. The rapid development of
Meifeng Cai et al. / J Rock Mech Geotech Eng. 2011, 3 (4): 373–384
numerical modeling methods provides efficient and
powerful tools for quantitative calculation and optimal
design of mining, geotechnical and other similar rock
engineering excavations. All the calculation and
design should be performed in a condition of known
stress state in the engineering area.
4.1.4 Optimal design of high and steep slope in
deep-concave rock engineering
Traditional method for slope design is the limit
equilibrium analysis method. This method is based on
the gravity equilibrium principle, which neglects the
influence of horizontal tectonic stresses, rock mass
characters, faults, etc. However, these factors play key
roles in stability of the slope in deep-concave rock
engineering.
It is often assumed that the tectonic stress in hillside
area has been completely released. Stress
measurement with both overcoring and hydraulic
fracturing techniques was completed in the Ekou iron
mine in 1992. In the mine, the final vertical height of
the slope is 720 m and the excavation depth is 440 m.
It was the first time to carry out systematical in-situ
stress measurement in deep-concave open-pit mines in
China. The measured results showed that the
horizontal stress was significantly larger than the
vertical stress, in which the traditional assumption is
limited. For reliable and optimal design of the slopes
in deep-concave open-pit mines, the effect of the
horizontal tectonic stress should be considered [24].
The shiplock in the Three Gorges Project is a key
engineering structure for the project. It is located at the
left side of the Yangtze River with a length of 1 600 m.
Both sides of the shiplock are high and steep slopes
with the maximum vertical height of 170 m. The
shiplock walls at two sides are almost vertical with
height of 50–70 m. The shiplock of the Three Gorges
Project is the typical deep-concave rock engineering
with high steep slopes. To ensure long-term stability of
the shiplock, stability analysis and optimum design of
the high and steep slopes should be performed. To
provide sufficient information for the analysis and
design, in-situ stress measurement with the largest
scale in China has been carried out since 1984 [25, 26].
The two main techniques, i.e. hydraulic fracturing and
overcoring techniques, were used for the measurement.
Using the overcoring technique, stress measurement at
3 deep boreholes was performed with the maximum
depth of 304 m, from which the 3D stress state at 33
measuring points within 21 sections was determined.
Besides, overcoring stress measurement was also
conducted at 4 shallow horizontal boreholes within
adits, from which the 3D stress state at other 9
381
measuring points was obtained. Using the hydraulic
fracturing technique, stress measurement at 6
boreholes was performed with the maximum depth of
137.6 m, from which the 2D stress state at 39
measuring points was determined. The measured
results show that the stress state in the measured
region is dominated by horizontal tectonic stress with
lateral factor of 1.08–1.82, which provides solid basis
for scientific design and construction and safe
operation of the shiplock.
4.1.5 Prediction of rockburst and other dynamic
disasters induced by deep excavations
Rockburst and other dynamic disasters induced by
deep excavations are closely related to rock stress state.
They are all dynamic processes of energy
accumulation and stress release during excavation. To
quantitatively calculate magnitude and distribution of
underground energy accumulation as well as its
evolutionary process based on clear understanding of
in-situ stress state is the only way for making time-
space-strength prediction of dynamic disasters induced
by deep excavations. Quantitative calculation with
numerical modeling method can provide detailed
information on magnitude and distribution of stress
and strain in the surrounding rocks. From the stress
and strain, the magnitude and distribution of the
energy accumulation around the excavation face can
be obtained. Through energy analysis, evaluation and
prediction of rockburst and other dynamic disasters
induced by deep excavations in the engineering can be
made using the knowledge of earthquake theory.
Therefore, in-situ stress measurement is indispensable
for the evaluation and prediction.
Based on the above theory and technique, a ML =
3.7 mining seismicity (artificial earthquake like
rockburst in metal mine) induced by mining
excavation in the Laohutai coal mine, which is located
in Fushun City of Liaoning Province, was predicted a
month before its happening [27].
4.2 For earthquake monitoring and prediction
Professor Siguang Li believed that the earthquake
occurs as the stress exceeds the strength of rock mass.
It is the origin of the energy theory for earthquake
prediction. After the M6.1 earthquake happened at the
Xinfengjiang reservoir in Heyuan County of
Guangdong Province in 1962, Professor Siguang Li
made effort to study the inducing mechanism of
earthquake through in-situ stress and displacement
measurement. In 1964, through measurement using
overcoring technique with piezomagnetic stress
gauges, the magnitude of the maximum and minor
principal stresses and orientation of the maximum
382
stress were determined at the Xinfengjiang reservoir
area. On March 8, 1966, an earthquake of M6.7 with
death toll more than 8 000 occurred in Xingtai of
North China, 350 km northeast to Beijing. After this
earthquake, practical earthquake prediction study
started in China. Researchers were looking for various
precursors of the earthquake at that time, including
seismicity, deformation, electric resistivity change,
groundwater change, animal response, etc. In-situ
stress was also considered and expected as the most
promising one [28]. So stress measurement is a key
issue in earthquake prediction. Under organization of
Professor Siguang Li, several tens of in-situ stress
monitoring stations were established, which were
distributed in 21 provinces of China.
The in-situ stress monitoring station in Shandan
County of Gansu Province in Northwest China was
established in 2007. It is 844.3 km away from the
epicenter of the M=8.0 earthquake that happened in
Wenchuan on May 12, 2008. In the station, a high-
precision volumetric strain gauge system was installed
in a borehole drilled in the red coarse granite for
in-situ stress monitoring. During the Wenchuan
earthquake, the volumetric strain gauge system at the
Shandan monitoring station successfully observed the
whole evolution of the in-situ rock strain. Figure 6
shows the complete strain curve from May 10 to 13.
From the figure, it can be seen that before May 11, the
strain curve is smooth normally with only periodic
strain changes caused by solid earth tides. However,
quite strong strain wave activities were observed in
two periods, i.e. from 7:45 AM to 5:57 PM on May 11
and from 6:32 AM to 7:41 AM on May 12. The
duration of these strain wave activities was about 30
and 8 hours, respectively, before the Wenchuan
earthquake, which could be considered as the
Strain (1010)
Fig.6 Strain curve of 96 hours observed in Shandan station during Wenchuan earthquake.
Meifeng Cai et al. / J Rock Mech Geotech Eng. 2011, 3 (4): 373–384
remarkable precursor of the Wenchuan earthquake.
The above monitoring results indicate that the
remarkable warning signs before earthquake have
been captured by the strain/stress monitoring system,
which demonstrates that the strain/stress monitoring
technique is useful for prediction and forecast of
earthquakes. Strain/stress monitoring can provide
information on the strain energy accumulation in the
earth’s crust, which is the source to induce
earthquakes. In contrast, seismological monitoring can
only record seismic waves that reflect the energy
release. Therefore, strain/stress monitoring has
stronger early-warning capability than seismic
monitoring for earthquake forecasting. However,
strain/stress monitoring at only one or few points does
not achieve a practical prediction. It is necessary to
build a monitoring network over a large enough range
around active earthquake belts. The monitoring results
from such networks can provide prediction
information on the location, strength and time of
earthquakes [29].
In recent years, great progresses have been made for
stress measurement used for earthquake prediction
study in China. Repeated in-situ stress measurements
by overcoring technique found a stress drop of a few
MPa after the Kunlun earthquake. In-situ stress was
also measured with hydraulic fracturing technique in
the northeastern end of the active faults of the
Wenchuan earthquake in 2008, indicating that a high
maximum horizontal shear stress existed near the fault,
and it kept a high value even after the coseismic stress
drop. Long-term monitoring of stress variations in
China now can provide reliable records of earth tides,
which was found in Guza station, 140 km away from
the epicenter of the Wenchuan earthquake. Azimuthally
anisotropic responses to earth tides were found in
Meifeng Cai et al. / J Rock Mech Geotech Eng. 2011, 3 (4): 373–384
many stress stations. Variations of the responses were
observed in Yushu station before the Yushu earthquake
in 2010, while no such change was observed in stable
regions. These progresses are exciting, and a project
for construction of about 300 new stress stations is
underway [26].
5 Conclusions and suggestions
(1) The knowledge of in-situ stress state is
indispensable for design and construction of all
engineering projects that involve excavations in rocks,
such as mining engineering, geotechnical engineering,
hydropower engineering, railway and road engineering,
petroleum engineering. It is suggested that for reliable
design and construction of rock engineering,
information on in-situ stress state should be provided
before design and construction of the engineering,
which means that it is necessary to carry out primary
in-situ stress measurement in the early exploration
stage of the engineering. Progress and development of
in-situ stress measurement, since the end of the 1950s
in China, have made the scientific level of rock
engineering design and construction greatly improved
and enhanced. Application of in-situ stress
measurement to earthquake monitoring and prediction
has already made exciting progress.
(2) Stress relief by overcoring technique and
hydraulic fracturing technique are two mainly used
techniques used for in-situ stress measurement in
various rock engineering fields in China. Stress relief
by overcoring technique has been used for in-situ
stress measurement for a long time and is relatively
mature with a higher measuring accuracy than other
techniques. It is especially suitable for detailed in-situ
stress measurement in mining and other similar
engineering at their operation stage, in which there are
many entrances to the position where the stress
measurement is conducted. Hydraulic fracturing
technique is an efficient technique for in-situ stress
measurement at depth. It is suitable for stress
measurement at the early exploration stage of
underground mining and other similar engineering as
well as surface excavation such as open-pit mining
engineering where there is not any entrance to
underground positions for stress measurement with
overcoring and other techniques. For major
engineering projects, both overcoring and hydraulic
fracturing techniques are used to ensure the reliability
of the measured results by mutual comparison and
checking.
(3) During the last 20 years, new ideas and new
383
techniques for in-situ stress measurement were
developed based on the demand of various newly
developed engineering projects with larger scale,
deeper position, and more complicated geological and
environmental conditions. Especially, a series of
techniques and equipments have been developed for
the two main stress measurement techniques. Full
temperature compensation technique, techniques
accounting the effect of nonlinearity, anisotropy and
discontinuity of rocks on the measuring results,
wireless automatic data transmission and collection
techniques, local stress relief without full overcoring
techniques and so on, have been successfully used to
increase the measuring reliability, accuracy and
practical usability of the overcoring technique.
Application of new type of sealing packers with
high-pressurizing strength, automatic valve for
low-pressure relief of the sealing packers, ultrasonic
scanning probe for precise detecting of fractures on
the borehole wall, high-pressurizing system with a
specially developed pump and so on, has significantly
increased the pressure-enduring ability, removal
flexibility and measuring accuracy of the traditional
hydraulic fracturing technique. With the improved
techniques and equipments, the record of measuring
depth with hydraulic fracturing technique in China
was changed from 800 m in 1983 to 1 105 m in 2004,
and changed again to a new record of 2 800 m in 2010.
(4) Current commonly used techniques for in-situ
stress measurement, including overcoring and
hydraulic fracturing techniques, are all “point”
measuring techniques. Based on the measured results,
using fitting and back analysis methods to build an
in-situ stress model for an engineering area is useful.
In recent years, nonlinear plural regression, neural
network, numerical modeling, fuzzy comprehensive
appraisal, genetic algorithm and particle swarm
optimization methods were widely used for the
analysis [30, 31].
References
[1] Amadei B, Stephansson O. Rock stress and its measurement. London:
Chapman and Hall, 1997.
[2] Cai Meifeng, Qiao Lan, Li Huabin. Principles and techniques of
in-situ stress measurement. Beijing: Science Press, 1995 (in Chinese).
[3] Liu Yunfang. Geostress and engineering construction. Wuhan: Hubei
Science and Technology Press, 2000 (in Chinese).
[4] Li Siguang. Introduction to geomechanics. Beijing: Science Press,
1973 (in Chinese).
[5] Yao B. Rock stress study and its progress. In: Mechanical Problems
Related to Engineering-geology of Rock Mass. Beijing: Science
384
Press, 1987 (in Chinese).
[6] Wang Lianjie, Pan Lizhou, Liao Chunting, Ding Yuanchen, Ou
Mingyi. In-situ stress measurement and its engineering application.
Beijing: Geological Publishing House, 1991 (in Chinese).
[7] Li Guangyu, Bai Shiwei. In-situ study on stress in rock mass. Rock
and Soil Mechanics, 1979, 1 (1): 76–90 (in Chinese).
[8] Liao Chunting, Cui Mingduo, Ren Xifei. In-situ stress measurements
and tectonic stress field in the Jinchuan mine area. Beijing:
Geological Publishing House, 1985 (in Chinese).
[9] Li Fangquan, Zhai Qingshan, Bi Shangxu, Liu Peng, Zhang Jun,
Zhao Shiguang. In-situ stress measurement by hydraulic fracturing
and preliminary results. Acta Seismologica Sinica, 1986, 8 (4): 431–
438 (in Chinese).
[10] Cai Meifeng. Review of principles and methods for rock stress
measurement. Chinese Journal of Rock Mechanics and Engineering,
1993, 12 (3): 275–283 (in Chinese).
[11] Cai M, Thomas L J. Performance of overcoring stress measurement
devices in various rock types and conditions. Transactions of the
Institution and Metallurgy (Section A: Mining Industry), 1993, 102:
134–140.
[12] Cai Meifeng. Studies of temperature compensation techniques in rock
stress measurements. Chinese Journal of Rock Mechanics and
Engineering, 1991, 10 (3): 227–235 (in Chinese).
[13] Cai M, Qiao L, Yu J. Study and tests of techniques for increasing
overcoring stress measurement accuracy. International Journal of
Rock Mechanics and Mining Sciences and Geomechanics Abstracts,
1995, 32 (4): 375–384.
[14] Ding X, Li W, Li L. A new type of overcoring stress measurement
system with automatic data collection in deep boreholes. In:
Proceeding of the 3rd National Conference on Rock Stress
Measurement. Beijing: Earthquake Press, 1994 (in Chinese).
[15] Zhao Yajun, Zhang Fei. Multichannel remote measuring system for
system data gathering in borehole stress measurement. Nonferrous
Metals, 1995, 47 (2): 16–18 (in Chinese).
[16] Ge Xiurun, Hou Mingxun. New approach to measure geostress—
local borehole-wall complete stress relief method. Chinese Journal of
Rock Mechanics and Engineering, 2004, 23 (23): 2 923–2 927 (in
Chinese).
[17] Cai Meifeng, Peng Hua, Ji Hongguang. New development of
hydraulic fracturing technique for in-situ stress measurement at great
depth of mines. Journal of University of Science and Technology
Beijing, 2008, 15 (6): 665–670.
[18] Cai Meifeng, Chen Changzhen, Peng Hua, Ji Hongguang, Qiao Lan,
Tan Zhuoying. In-situ stress measurement by hydraulic fracturing
technique in deep position of Wanfu coal mine. Chinese Journal of
Rock Mechanics and Engineering, 2006, 25 (5): 1 069–1 074 (in
Chinese).
[19] Institute of Geomechanics, Chinese Academy of Geological Sciences.
Meifeng Cai et al. / J Rock Mech Geotech Eng. 2011, 3 (4): 373–384
Research of deep borehole hydraulic fracturing stress measurement
system. Beijing: Institute of Geomechanics, Chinese Academy of
Geological Sciences, 2010 (in Chinese).
[20] Liu Yunfang. 3-dimensional geostress measurement by hydraulic
fracturing technique in one borehole. Chinese Journal of Rock
Mechanics and Engineering, 1999, 18 (2): 192–196 (in Chinese).
[21] Peng H. 3D stress measurement with hydraulic fracturing technique.
In: Proceedings of the National Conference on In-situ Stress
Measurement and Its Application. [S.l.]: [s.n.], 1991: 27–32 (in
Chinese).
[22] Cai M, Qiao L, Li C, Yu B, Wang S. Results of in-situ stress
measurements and their application to mining design at five Chinese
metal mines. International Journal of Rock Mechanics and Mining
Sciences, 2000, 37 (3): 509–515.
[23] Cai M, Qiao L, Li C, Yu B, Chen G. Application of an improved
hollow inclusion technique for in-situ stress measurement in
Xincheng gold mine, China. International Journal of Rock Mechanics
and Mining Sciences and Geomechanics Abstracts, 1995, 32 (7):
735–739.
[24] Cai M, Yu B, Qiao L, Chen G, Li C. Experience of in-situ stress
measurement with hydro-fracturing and overcoring techniques in
Ekou mine, China. International Journal of Rock Mechanics and
Mining Sciences, 1997, 34 (2): 299–302.
[25] Liu Yunfang, Xiao Guoqiang, Xu Jing. Geostatic stress measurement
for TGP shiplock and its dynamic variation after excavation. Yangtze
River, 2008, 39 (5): 3–6 (in Chinese).
[26] Zhong Zuowu, Xu Chunmin, Liu Yuankun. Study on the variation of
geostresses before and after excavation of permanent shiplock of the
Three Gorges Project. Chinese Journal of Rock Mechanics and
Engineering, 2001, 20 (Supp.1): 1 831–1 834 (in Chinese).
[27] Cai M, Ji H, Wang J. Study of the time-space-strength relation for
mining seismicity at Laohutai coal mine and its prediction,
International Journal of Rock Mechanics and Mining Sciences, 2005,
42 (1): 145–151.
[28] Shi Y. Stress measurement and earthquake prediction in China. In:
Abstracts Collection of the 5th International Symposium on In-situ
Rock Stress. Beijing: [s.n.], 2010.
[29] Cai M, Peng H, Ma X, Jiang J. Evolution of the in situ rock strain
observed at Shandan monitoring station during the M8.0 earthquake
in Wenchuan, China. International Journal of Rock Mechanics and
Mining Sciences, 2009, 46 (5): 952–955.
[30] Fu Yuhua, Wang Xingming, Yuan Haiping. Finite element inverse
analysis of boundary load for tectonic stress field. Rock and Soil
Mechanics, 2009, 30 (6): 1 850–1 855 (in Chinese).
[31] Hu Bin, Feng Xiating, Huang Xiaohua, Su Guoshuo, Zhou Hui.
Regression analysis of initial geostress field for left bank high slope
region at Longtan hydropower station. Chinese Journal of Rock
Mechanics and Engineering, 2005, 24 (22): 4 055–4 064 (in Chinese).