You are on page 1of 29

International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989

An overview of rock stress measurement methods
3. Ljunggrena,*, Yanting Changa, T. Jansonb, R. Christianssonc
0c 0a SwedPower AB, Sweden 0b Golder Associates AB, Sweden Svensk Karnbr.anslehantering. AB, Sweden Accepted 20 July 2003

Abstract This paper presents an overview of methods that have been used to estimate the state of stress in rock masses, with the emphasis on methods applicable to hard rocks and Scandinavia. Rock stress is a difficult quantity to estimate because the rock stress measuring techniques consist of perturbing the rock, measuring displacements or hydraulic parameters, and converting the measured quantities into rock stresses. There are two main types of method: those that disturb the in situ rock conditions, i.e. by inducing strains, deformations or crack opening pressures; and those that are based on observation of rock behaviour without any major influence from the measuring method. The most common methods are briefly described including their application areas and limiting factors. r 2003 Elsevier Ltd. All rights reserved.

1. Introduction

This paper focuses on the methods for rock stress measurement with the emphasis on hard rock. It is known that the reliability of rock stress measurements/ estimations is partially dependent on the measuring technique and equipment, and partially dependent on the nature of rock masses; a large amount of literature exists on the subject of rock stress and these factors. One relatively recent compilation of information is the 1997 book by Amadei and Stephansson [1]. The so-called ‘scale effects’ have been studied intensively in recent years and a review of this aspect and a statistical study have been presented by Ljunggren et al. [2]. However, it is beyond the scope of this paper to discuss scale effects and other similar issues in detail; instead, the most common methods are highlighted and discussed in the It must be emphasised that context of the rock volume included in the measure-ment. Also included are the conditions for which the different methods are appropriate. Factors such as, inter engineering and safety issues alia, the purpose of the measurements, borehole loca-tions, borehole orientations, should govern the rock stress geological circumstances and water conditions will have an impact on the decision measurement process. The rock

on which method to apply. The paper discusses both direct measurement methods, e.g. hydraulic fracturing and overcoring, as well as indicative methods, such as core discing. The paper distinguishes between methods applicable from the ground surface and methods to be used when there is underground access, recognizing that some methods can be practised from both the ground surface and via underground access.

*Corresponding author. Tel.: +46-920-77-335; fax: +46-920-77-375.
E-mail address: christer.ljunggren@swedpower.com (C. Ljunggren).

stress information may be required for the following engineering aspects, either directly or as input to numerical models:
0* long- and short-term stability of underground struc-tures (tunnels, caverns, shafts and other openings);

1365-1609/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijrmms.2003.07.003

0* determination of excavation methods (drill-and-blast or TBM and raise-boring);

0* 0* 0* 0* 0* 0*

design of rock support systems; prediction of rock bursts; thermo-hydro-mechanical behaviour of the rock; design of grout methodology; fluid flow and contaminant transport; fracturing and fracture propagation.

Rock stress measurements are required as input information for the above engineering issues and design. Hence, for each measurement campaign, the aim of the

the in situ stresses. Further measurements. 0* geological observational methods and 0* earthquake focal mechanisms. Furthermore. (direct and/or indirect) methods at different locations. inducing strains. depending on at what stage an * hydraulic methods. Ljunggren et al. The first consists ofvolume involved in their use methods that disturb the in situ rock conditions.1.e. 2. when the stress state at a location is to be determined using the overcoring technique. the procedure is to conduct a series of tests to obtain accurate and reliable results at a given location or within a predetermined depth interval. byprovided. The data obtained from eachmethods method should be analysed separately and checked to see if observation of rock behaviour the simplifying assumptions associated with each method are without any major influence from met. methods may be included in this The objective in determining rock stresses may in manycategory: cases be seen as an interactive process. The following a suitable approach strategy to the estimation programme. Methods 2.976 C. deformations or The rock volumes presented in Table 1 above indicate the typical volumes involved in a test using the different methods. The methods may also be classified by their opera-tional Methods for the determination of in situ rock stress can betype and an indication of the rock classified into two main categories. and For a given project. stress 0* statistics of measured measurements can be conducted in several stages with one or data (database). Also. This approach is recommended since it will provide a measure of second consists of confidence through considerations of the consistency and The based on the reliability of the information. Types of stress measurement methods 0* strain recovery methods. different methods for a given project. 0* borehole relief methods that further advance of the project should be questioned. stresses can be determined using several 0* surface relief methods. The details of the stress measurement programme will depend on what questions are to be solved in any specific project. given that the . several methods. When conducting stress measurements. The combination of data is also useful when a limited number of tests from each method is available. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 measurements must be fully understood in order to developcrack opening. it is known from experience that at least five successful separate tests should be obtained close to each other in order to obtain adequate results on the stress state at that specific location. A more thorough discussion can be found in 0* acoustic methods (Kaiser Brudy [3]. The order to impose more rigorous constraints on the in situfollowing methods belong to this category: stresses. including underground project is in. The type of data needed may change. The idea is to use the best attributes of the 0* core-discing. preliminary hydraulic fracturing and hydraulic tests on pre-existing fractures measurements may reveal information on the stress state (HTPF). see Table 1. effect). Combining several 0* borehole breakouts. i. methods (hybrid measure-ment) based on their respective 0* relief of large rock attributes can help in obtaining a more reliable assessment of volumes (back analysis). The data from different methods may also be combined in the measuring method. For example.

geology does not change.5–2 1–2 109 108 102–103 .5–50 10₃3–10₃2 1–10 10₃2–100 10₃3 10₃3 Category Methods performed in boreholes Method Methods performed using drill cores Methods performed on rock surfaces Analysis of large-scale geological structures Other 10₃3 0. Table 1 Methods for rock stress measurement classified by operational type (the rock volume involved in each method is also given) Rock volume (m3) Hydraulic fracturing Overcoring HTPF Borehole breakouts Strain recovery methods Core-discing Acoustic methods (Kaiser effect) Jacking methods Surface relief methods Earthquake focal mechanism Fault slip analysis Relief of large rock volumes (back analysis) 0. may not alter the average results.

from experience on rock2. 3. signal amplifier. 2 presentsFig. (4) data registration equipment. classical concept for the interpretation of hydraulic fracturing pressure records was developed by Hubbert and Willis in 1957. Classical hydraulic fracturing stresses and geology. (3) flow meter manifold and manifold for control of fracturing flow and packer pressure. impression packers and high-pressure pumps to appr. Ljunggren et al. . The decades. the geological boundaries. the basic principles have remained unchanged and of the 1960s it was proposed to both techniques have been practised now for severalderive the state of stress from such hydraulic fracturing operations. measurements. during testing is shown in Fig. The main factors limiting the volumeused for fluid injection operations in for which the stress field may be judged to be representative are sealed-off borehole intervals to the vertical depth variation. Several authors. chart recorder and portable PC. The down-hole principlemeasurements: (1) guidewheel for multihose on adjustable working platform.g. for example. Example of equipment for hydraulic an example of equipment that is used for both hydraulic fracturing and HTPF rock stress fracturing and HTPF measurements. during the 1940s. (5) high pressure water pump and (6) 400 l diesel fuel tanks. These methods are hydraulic fracturing and HTPF. 1 2 3 4 2. two methods dominate the others: hydraulicin the oil industry to stimulate methods and borehole relief methods.2. (2) drum for 1000 m multihose. fractures in borehole wall rock. e. Although bothproductivity from low permeable oilmethods have undergone continuous development over thebearing formations. Hydraulic methods There exist two stress measurement methods that use hydraulics as an active method to stimulate the rock surrounding a borehole and hence to determine the stress field. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 977 Fig. 1 shows the rock volumes that.2. may be involved for some different stress The term hydraulic fracturing is measurement situations. In the beginning years. m generate high-pressure water during either the formation of new fractures or reopen-ing of pre-existing fractures.1.C. the 2002 paper by Rummel et al. 2. and theinduce and propagate tensile presence of major faults. Both 6 5 methods use the same type of equipment. including straddle 10 packers. see. Bjarnason [5] and Ljunggren [6]. hydraulic pump and tank. Fig. have presented the hydraulic fracturing method. [4]. It If we study the commercial application of stresswas first applied.

1. .Fig. Representative volumes involved in the stress measurement tests.

offering the least resistance. The orientation of the fracture is obtained from the fracture traces on the borehole wall. these components are the maximum and minimum horizontal stresses. suitable at the early stages of projects when no underground Furthermore. it is especially limitations normally imply that the advantageous for measurements should be done in vertical holes. the orientation of initiated fractures coincides with the orientation of the maximum horizontal stress. The hydrofracture will initiate at the point. This generates tensile stresses at the borehole wall. the method is 2D: only the maximum and minimum normal stresses in the plane perpendicular to the borehole axis are established. Ljunggren et al. The sealed-off section is then slowly pressurised with a fluid. In its conventional form. Fig. Hence. the method is most suited for surface measurements in vertical or sub- . 4 shows an example of a schematic view of a hydraulic fracturing system. the theoretical access exists. 3. For aFig. which requires space. of a borehole is sealed off with a straddle packer. [4]). in a vertical or sub-vertical hole where it is assumed that one principal stress is parallel to the borehole. usually water. areas are usually close to horizontal and vertical. The fracture plane is normally parallel to the borehole axis. A schematic view of a hydraulic fracturing system (from Rummel et al. Pressurisation continues until the bore-hole wall ruptures through tensile failure and a hydrofracture is initiated. it can often The method is also not be assumed that the components measured in a verticalsignificantly affected by the drilling borehole are two of the principal stresses.978 C. normally less than 1 m in length. 4. The fracture will therefore develop in a direction perpendicular to the minimum principal stress. A section. The fracture orientation may be determined either by use of an impression packer and a compass or by use of geophysical methods such as a formation micro-scanner or a borehole televiewer. and propagate in the direction. Thus. vertical borehole. processes. Due to its efficiency. Down-hole principle during (a) hydraulic fracturing and (b) HTPF measurements. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 Fig. and is thereforeequipment. normally in the horizontal plane. and two fractures are initiated simultaneously in diametrically opposite positions on the borehole per-iphery. Since the principal stress directions in tectonically passive and topographi-cally flatmeasurements at greater depth. Hydraulic fracturing Hydraulic fracturing is an efficient method for determining thenormally includes large 2D stress field.

vertical bore-holes. of the least stress in the plane The hydraulic fracturing method allows a direct measurementperpendi-cular to the borehole axis. which is normally the least .

be planar. evaluated and those data processed using computer codes. In from core drilling. The HTPF method has been practised for effects. The method has been applied in four projects in strike and dip of the existing the Nordic countries. fractures with different strikes and homogenous.2. includes less assumptions on the Classical hydraulic fracturing requires sections in thestress field but requires a larger borehole free from fractures. The method is across the fracture plane. [7] it is shown that the general theory for calculating the mea-surements. Ljunggren [6. pre-existing uncertainties in the assumptions—a continuous. it is suggested that major horizontal stress from the hydraulic fracturing suffers from at least 10–12 isolated. either the 2D or 3D stress state can be determined. it is of importance that the each discrete fracture to be fracture tested is of a size at which the normal stress can be tested. The 3D underestimated when the major principal stress divided by the alternative of the HTPF method minor principal stress is close to. Hydraulic tests on pre-existing fractures (HTPF) Cornet and Valette [8] first presented the theoretical basis features. In some redundancy. may also affect the possibilities of success as 18–20 successful tests are obtained they act as weakness planes and thereby may control theto resolve the 3D stress field.C. Nor does the method and practical use for the HTPF method. 2. dip and strike of the hydraulic frac-turing. Given these parameters obtained until all field-testing has for a sufficiently large number of fractures with different strike been completed. it is suggested that at least in gneissic rock. test depth. The HTPF method relies only on four field A drawback.9]. is also that tested fracture. The shut-in pressure is equivalent to the normal no preliminary results can be stress acting across the fracture plane. A 3D determination hole equipment must be requires a larger number of fractures to be tested. is required. As compared to classical direction of the initiated fracture. may be somewhatinterval of interest. The method is arequire determination of the development of the hydraulic fracturing technique because it tensile strength of the rock and it uses the same equipment and is based on measurement of theis independent of pore pressure same parameters. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 979 horizontal stress. For successful horizontal stress (B710– 20% or more). and isotropic rock to-gether with the fracturedips are found and tested in the reopening.2. hydraulic fracturing. Depending on assumptions made time consuming than regarding the stress field. a factor of 3. field pressure data. borehole wall within the depth determined from hydraulic fracturing. linearly elastic. In a few meters long so that the induced fractures do not interact the 3D alternative the vertical stress with existing ones. shut-in pressure. Ljunggren et al. such as when core discing is indicated in the coreexist in the system of equations. Borehole relief methods The family of borehole relief methods can be divided into the .3. This requires good assumed to be uniform and the geometry of the fracture must accuracy in the depth calibration. field data and dips. It is probable that the major horizontal stress. Geological features. Instead of inducing new fractures in intact rock. The accuracy is less good for the maximumpractise however. neither weakness planes such as [10]. Ljunggren and Raillard [11]. positioned at the exact location of When conducting HTPF tests. the HTPF method allows either a 3D more hydraulic fractur-ing as the downor 2D determination of the stress state. the method has the advantages of less limitations as regards geological 2. These sections should be at least number of fractures to be tested. The Theoretically the 2D solution requires at least six different maximum horizontal stress is calculated from equations including a failure criteria and parameters evaluated from the fractures to solve the problem. the HTPFfoliation planes nor core discing method is based on the re-opening of existing fractures found in should cause any problems in successful the borehole wall and thereby determining the normal stress obtaining measurements. such as foliation planespractise. or higher than. 12 unknowns stresses. Theoretically. In a study by Rutqvist et al. As long as a variation in some 15 years. sh and the accuracy is good (B75%). compared to parameters. Bjarnason and Rail-lard fractures exists in the rock mass. Hydraulic fracturing may be difficult to apply does not have to be a principal with an acceptable success rate in rock domains with very high stress.

Overcoring of measuring cells in pilot-holes Stiff/solid cells are more Overcoring based on the principle of overcoring a pilot unusual than the other two groups hole in which the measuring cell is installed can be divided and have a general problem with into further groups as follows: the difference in material properties between the rock and 0* soft inclusion cells.3.1.following sub-groups: 0* overcoring of cells in pilot holes. inclusion material [1]. 0* overcoring of borehole-bottom cells and 0* borehole slotting. 2. 0* deformation meters measuring displacements of the wall during overcoring and 0* stiff/solid cells. .

5. the gauges can be damaged the principle of the Borre Probe cell is presented. given in [1] and is in principle the same for acceptable field conditions (discussed later). gauges bonded to pilot-hole wall under pressure from the nose cone.3. These instruments measure one or several change in pilothole diameter during the process of overcoring. . The method is instruments for determining in situ well known and has much testing against. The method is considered to be reliable. Bywall during overcoring measuring at least six strain components in different directions The principle of deformation meters on the wall of a borehole. (2) drill +36 mm pilot hole and recover core for appraisal. 5.3. is that they allow the 3D state of stress to be the most reliable and accurate determined from one single measurement point. Table 2 the Sigra IST. the strain. which Some of the disadvantages of the runs on batteries and permits a continuous logging of thegauges are that: it requires an strain gauges before.stresses in rock by overcoring [1]. instead of 0* CSIR cell. 2. (4) probe releases from installation tool. which enhances the evaluation process. It could also be argued that scale is awith the gauge pistons. temperature changes and calibration against known boundaryThe theory for the USBM is described stresses. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 2. The USBM Common to all these three instruments. three non-parallel The triaxial cells with strain gauges in a pilot-hole are quiteholes are necessary to calculate the sensitive to the isotropy. Two commercial 0* CSIRO cell and instruments of deformation-type 0* Borre Probe cell. High stresses may also put limitations on the method as these may initiate core discing. if the core breaks. A description of the CSIR and CSIRO cell may be difference between the two found in Amadei and Stephansson [1]. potential limitation here as small scale variations in the rock material composition may affect the results. Another general limitation for all three instruments are that relatively long unbroken (40– 60 cm) overcores are required and hence similar lengths of the borehole free from fractures. homogenous and isotropic rocks. and a major gauge is extensively used and one of advantage.3. Ljunggren et al.length. and the gauge rock. during and after the overcoringunbroken core of at least 300 mm in process.980 C. Soft inclusion cells The principle of a soft cell is based on the theory of linearmeasuring displacements of the elasticity for continuous. The most common instruments based on the above principle are: The Borre Probe cell includes a built-in datalogger. the results from the triaxial cells oftendepends on the minerals in contact show a certain scatter. As a consequence.2. The instrument is installed in a pilot in anisotropic rocks have also been developed [12]. instruments. gages are the USBM gage and Sigra in situ stress tool (IST). (3) lower probe in installation tool down hole. for example. the complete stress tensor at the testfor measuring displacements is the location can be determined. Tables 2 and 3 summarises the characteristics and differences between the summarises the characteristics and instruments. Deformation meters hole and later overcored. In Fig. 3D pilot hole overcoring measurements: (1) advance +76 mm main borehole to measurement depth. Table 2 Characteristics of the most common soft overcoring cells Fig. Theories for stress measurements same as for the soft inclusion cells. Principle of soft. homogeneity and grain size of thein situ stress field.

via cable Yes. Tested for 1000 m . built in datalogger Borehole requirements 38 mm pilot hole. modified versions: up to 1000 m Normally: up to 30 m Practised to 620 m. probe bonded in place and (6) overcore the probe and recover to surface in core barrel. Accepts water-filled holes Normally: 10–50 m. Problems in waterfilled holes 36 mm pilot hole. Modified versions accept water 38 mm pilot hole. usually 90 mm drillhole.(5) raise installation tool. Instrument CSIR cell CSIRO cell Borre probe cell No of active gauges 12 9/12 9 Measuring depths Continuous logging No Yes. usually 150 mm drill hole. 76 mm drillhole.

Reading of the strain gauges is taken before and after overcoring overcoring. while the hemi-and at the base of the gauge a spherical or conical strain cell is attached to the hemi-sphericalstrain rosette consisting of 3 or 4 or conical bottom of the borehole. Fig. modified versions 4 3. Fig. The CSIR The Doorstopper cell. 76 mm drill Accepts water-filled holes . usually 90 m drillhole. Modified versions ac water Sigra IST Yes. Duringhole. could be cemented on the bottom of 60 mm Table 3 Characteristics of two instruments of the deformation-type gauge A modified doorstopper cell called the Deep Door-stopper Gauge System (DDGS) has been developed jointly by the Rock Mechanics Laboratory at Ecole Polytechnique in Montreal! and the Atomic Energy of Canada. is attached at theDoorstopper is 35 mm in diameter polished flat bottom of a borehole. Ljunggren et al.C.4. The device utilises an Intelligent Acquisition Module. built in datalogger 25 mm pilot hole. a remote battery-powered data Continuous logging No Borehole requirements ! Instrument USBM No of active gauges Measuring depths Normally 3. they do notstrain gauges is cemented. Designed for 1500 m 38 mm pilot hole.3.14]. Overcoring of borehole-bottom cells boreholes and overcored. Leeman [13. After the cell has been positioned properlyis pushed forward by compressed at the end of the borehole and readings of the strain gaugesair and glued at the base of a drill have been performed. of the cell. 7. 6. The cell require a pilot hole. developed a cell with strain gauges that 0* Doorstoppers and 0* spherical or conical strain cells. the instrument is overcored. The DDGS was designed to allow overcoring measurements at depths as great as 1000 m in subvertical boreholes [16]. Leeman op. the changes in strain/deformation are recorded. Hence. modified versions up to 1000 m Used to 700 m. in two or three levels Normally 10–50 m.cit. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 981 2. The cell Methods for overcoring of borehole-bottom cells discussed is often referred to as CSIR (Council in this paper include the following: for Scientific and Industry Research) Doorstopper and has been used for measurements in 60 m deep boreholes.

Fig. Installation of Doorstopper (after INTERFELS): (a) cored borehole NW=76 mm. 6. . (b) borehole bottom flattened and (c) polished.

the strain cell is bonded to the rock surface at the bottom of the borehole. Underground Research Laboratory (URL). possibilities for successful measurements in relatively weak or broken rock. or spherical methods. Obara and Sugawara [15]. Ljunggren et al. . [18. is that they do Successful measurements have been performed at thenot require long overcoring lengths. Furthermore. Other advantages. e. Coordinate system and strains to be measured on a conical bottom developed.g. As the methods do not require and triaxial strain cells were not applicable at depths deeper a pilot hole there are also better than 360 m because of the high stress situation. a Dorrstopper measurement requires less time. Thereafter. solutions for anisotropic rocks have also been Fig. Compared to triaxial cells. The principle of the DDGS installa-tionDoorstopper. that measurement at one point only enables the stresses in the plane perpendicular to the borehole to be determined. the end of the borehole must be flat which require polishing of the hole bottom. valid only for the modified Doorstopper.19] surface (after Obara and Sugawara [15]). / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 advantage for the logger that collects and stores strain data during stress An measurement tests. and 2–3 tests can be conducted per day. The latest version of the conical strain cell. For doorstopper cells. where both hydraulic fracturingcm). Its bottom surface is then reshaped into a hemispherical or conical shape using special drill bits. Using a hemispherical or conical strain cell for measuring rock stresses. include possibilities for continuous mon-itoring and the application in water-filled boreholes. Doorstopper methods have been developed and practised for more than 20 years worldwide.e. 7. as well as in rocks under high stresses in which core discing is common.. has been suc-cessfully tested by. 8. however. Thei. The disadvantage with the doorstopper is. Corthesy et al. only some 5 cm. equipped with 16 strain components. a borehole is first drilled. as compared to measurements were made at borehole depths as great as 518the pilot hole methods (at least 30 m (943 m depth from surface). Canada [16]. as well as the conical is shown in Fig. developed a mathematical model to account for both non-linearity and transverse isotropy in the analysis of overcoring measurements with the CSIR Doorstopper.982 C.

(6) When overdrilling is completed. (3) The IAM and Doorstopper gauge are removed from the installation equipment. Installation of the DDGS [17]: (1) After flattening and cleaning of the bottom. the core is taken up and a bi-axial pressure test done to estimate the Young’s modulus. . 8. (2) When the DDGS is at the bottom the orientation of the measurement is noted in the orientation device and the strain sensor is glued. (4) The installation assembly is retrieved with the wire line system. the strain change in the bottom is measured by the time.Fig. (5) The monitoring and overdrilling start. the instruments are lowered down the hole with the wire line cables.

5. 10.C. it has only been applied in boreholes at shallow 2. Surface relief methods depths. mainly saw cuts the slots.3. methods. the doorstopper is the cut into the wall. It is based on the theory of linear elasticto allow for pilot hole behaviour of the rock and uses the Kirsch solution formeasurements. [20]). Summary—borehole relief overcoring.measures the rock response to stress good agreement has been found relief (by cutting or drilling) by recording the distance between gauges or pins on a rock surface before and after the relief. and the stress parallel to the borehole axis must be known. As a 2. been reported in the literature (e. Of the other borehole.application areas and limitations which is mounted against the wall of a large-diameterare fairly well known. due to their The conical or between 10 and 15 measurements can be made during a day’slimitations. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 983 Measurements with the conical borehole technique havebetween stress measurements been made mostly in Japan.Some of them. and they require preparation of the borehole bottom. Borehole slotting consequence. In general. Matsui et al. a small length of the rock is required for 2.g.0 mm wide and upto perform stress measurements to 25 mm deep. are still Disadvantages with the conical or hemispherical cell are thatin the development stage. on a regular basis in many underground projects. either in thehave not become commercially form of a cone or as a sphere. The instrument is fully recoverable and provides continuous monitoring of strain as the slot is cut. such as that the borehole must be dry. aremethods. Tangential strains induced by release ofin highly stressed rock volumes or tangential stresses by the slots are measured on thewhen the fracturing is too intense borehole surface. Like the Doorstopper. The methodbeen used in Asia. Borehole breakouts Borehole breakout is a phenomenon that occurs when the rock is unable to sustain the compressive stress concentrations around a borehole. have only been used not require any overcoring. 120₃ apart.6. Another limitation is their pooravailable. The other stresses and strains around a circular opening. This results in Fig. pneumatically drivennext most used technique. Each slot is typically 1. although theoretically The most significant advantage with the method is that it does adequate.5. their main The borehole slotter consists of a contact strain sensor. Bock [21]. however. Thereafter. Hemi-spherical or conical strain cells have There exists a variety of the somostly been used in Japan and successful applications havecalled borehole relief methods.3. hemispherical cell has mainly work. 9. . The category is most suitable for measurement on tunnel surfaces For information the reader is referred to Amadei and Stephansson [1] and de Mello Franco [22]. Cross-section of the borehole slotter (INTERFELS). Ljunggren et al. three slots. 9. Today the pilot hole methods dominate and are used success in water-filled boreholes. borehole and in various rock types. For the conical cell. the stress relief is achieved at methods an overcoring distance of 70 mm and then the strains remain at constant values [15]. The method is also quick andon a few occasions. The borehole slotter has been designed to work in This category of methods boreholes with a diameter between 95 and 103 mm. Examples of the technique are the flat jack method and the curved jack method. see Fig.4. 2. This technique has been found with the borehole slotter and to be a useful method for measuring rock stress in a single measurements with other techniques. see Fig. A small. also contains a number of limitations.

Breakouts also provide a valuable link between . This has led several authors the purpose of rock stress determination. It hasof various other factors. not directly been found that the shape and depth of breakouts in vertical holes the stress concentrations. Haimson and Lee [23]). Attempts can be found in the literature where the authors use the must be used with caution as depth and width of the borehole breakouts in order to estimate the breakouts can be enlarged because magnitude of the rock stresses (e.breakage of the wall on two diametrically opposed zones.depend on the magnitude of the major and the minor horizontal in situ called ‘breakout’. Breakouts were foundto suggest that the geometry of the to occur along the direction of the least principal stress. However.breakouts could be used for Therefore. this approach orientation of in situ stresses. breakouts are generally used to determine theestimating the magnitude of the in situ stresses.g. Leeman [13] first reported the use of borehole breakouts forstresses.

These discssuggest that disc thickness is sometimes exhibit parallel faces but are often shaped like a horseindicative of the level of the saddle. Development of borehole breakouts. direction the same cracks breakouts exist. the laboratory tests rock core often appears as an assemblage of discs. dipmeter. Another advantage is that it may reveal informationthe direction of the least horizontal valid at great depth where other methods may be insufficient. the thickness of discs decreases monotonically with increasing sH.6. the In addition. Tearea of core discing and borehole Kamp [25]. Also. [26]. The use of borehole breakouts as ancarried out at the University of indicator of the stress field has been used in some projects withWisconsin over the last decade in the deep drilling such as: the KTB hole in northeastern Bavaria. to analyse this discing process in order to extract information on the To characterize breakouts. the Cajon Pass hole in the vicinity of the San breakouts [29]. so the method cannot be used to are practically horizontal. resulting in saddle-shaped discs Anisotropy of the rock may disturb the breakout locations and having a trough axis oriented in the sH thereby also jeopardise the usefulness of the information. Core The method’s major advantage is that it is quick to use and tensile fractures initiated below the requires only measurement of the diametrical changes of thecoring-bit extend toward the axis of borehole wall to obtain information on the orientation of thethe core with slight downward tilt in stress field. breakouts do not normally exist in Swedish rocks above 1000 m. When boreholes are core drilled in highly stressed regions. [27]. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 Fig. As drilling obtain information above these depths. This observation reinforces 2. a televiewer. core. hydraulic fracturing and focal mechanism data [24]. Core discing the idea that discs recovered from oriented cores could be used as indicators of the in situ sH orientation. Ljunggren et al. For example. Attemptsapplied stresses. sH. s : In the maximum horizontal h The major limitation of the method is that it only works if stress. This phenomenon has been called ‘core discing’. logging of the diametrical shape oflocal in situ stresses. for given magnitudes of sh and sv. Such tools can be a caliper like a and Haimson [29]. . but also in the base of the Stephansson et al. overcoring. breakouts). giving rise to discing. Shamir et al.984 C. these fractures open be used to determine the magnitudes of the stresses. formation micro-scanner or a high Much experimental work has been resolution TV camera.stress. Careful measurements of core disc have been made dimensions show that. and thehigh stresses bring about failure. The studies show that Andreas fault in southern California. not only at the borehole wall (resulting in borehole for deep earth gas in Precambrian rocks of Sweden. 10. Hakala [28] the borehole is required. the method cannotadvances. direction.

thickness and trough axis attitude. also pointed out that. The results indicate that the sH magnitude andmaximum horizontal stress from the measurements of characterorientation could be estimated from the average discistic dimensions. Haimson op.Hakala [28] has suggested a methodology for interpretation of together with borehole breakouts.Fig. and sH is established. provided extracted oriented-core is disced and the relation between thickness Based on numerical analyses.cit.. core discing can providein situ upper limits on in situ stress levels and help assess the . 11.

Anelastic strain recovery (Asr) for estimation of rock stresses. cit. However. which is used to change support systems and excavation schemes on an not commonly used. it tends to concentrations are higher than the rock strength. Theserecovery. after relief from the surrounding stress field. 12. The ASR method. determination Akutagawa [31] also presented material to account for non. is of stresses are removed from the course valuable and a guide for further decisions.C. the confidence of the interpretation can be increased considerably if the same Both methods are applied on information can be achieved from both normal coring anddrill cores from bore-holes. Suchrelax and expand.of the magnitude of the stresses is elastic rock behaviour. Methods based on the above principle more difficult and a constitutive are often referred to as back-analysis methods. 2. model for the rock must be used. Other advantages are that they may be costly. By measuring the strain mations/displacements of the rock mass will occur. Field measurements show that the opening and propagation of preferential microcracks usually accompanies anelastic strain During excavation of underground openings. The limitations include that the techniquebeen tested in very deep can only be used on underground openings and only during . require limited and simple measure-ment procedure and the core is orientated. This relaxation consists of stresses from core discing. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 985 the excavation process of underground openings.8.7. Relief of large rock volumes The advantage of back analysis methods is that they are quick to The ASR method requires that use. 11.recovery. Sakurai and obtained. obtained already during the drilling stage.8. maximum stress direction and less in the direction of the minimum stress. Methods that utilise the principle of strain recovery are: 0* anelastic strain recovery (ASR) and 0* differential strain curve analysis (DSCA) According to Hakala.1. which may be include large volumes. the mean disc spacing. The method requires numerical analyses and does not allow unique solutions. Ljunggren et al. 0* the tensile strength of the rock. In practice. More recently. core. 2. Thedetermined. Example of relation between disc thickness td (normalised by core expands more in the direction of the diameter) and sH for given sh and sv [29]. defor. see Fig. the shape of the fracture (morphology) and the extent of the fracture in the core. due to that information. Thus the orientation displacements measured are related to the in situ stresses byof the principal stresses can also use of numerical methods. op. Fig. When a drill core is removed one can of course also conclude that rock stress from the rock mass. 0* 0* 0* 0* 0* Poisson’s ratio of the rock. the orientadisplacements may be measured using instrumen-ted crosstion of the principal strains can be sections in the openings. overcoring at the same depth level. The following minimum an instantaneous elastic component information is needed for the interpretation: and a time-dependent anelastic component. the uniaxial compressive strength of the rock. Sakurai and Shimizu [30]. Strain recovery methods Measurement of stress field parameters using the method of strain recovery is based on the principle that a core.. core discing can only be used as an indicator 2. has mainly almost real time basis. When core discing occurs.

Three pairs of radial inductive displacement transducers and one axial transducer are used to measure the anelastic response of a core sample. Instrument for ASR measurements of a drill core.Fig. . 12.

After an and then increased. compression in the laboratory. Amadei and Stephansson [1]. situ rock stresses. there is a oriented drill core is brought up to the surface.2. It the strain due to the closure of the micro-cracks can be obtained by has been hypothesised that the the core with strain gauges. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 boreholes where methods such hydraulic fracturing have not been possible or too expensive to use. Teufel [32]: temperature variation yielding thermal (1 ) strains.9.e. about the last (9 ) accuracy of core orientation. 0* electromagnetic methods [38]) and 0* acoustic methods (Kaiser effects. The difference was attributed to difficulties in over the past 15 years as a potential method to determine in measuring small strains. the micro-cracks significant increase in the rate of have had time to develop and to align themselves in the direction of acoustic emission as the stress the original stresses. can be detected by stressing the Thus. (6) drilling mud–rock not gained much popularity in (5 ) interaction. Ljunggren et al. Despite volumes. assumptions are made: could be inferred by monitoring the * The principal directions of the in situ stress field coincide with acoustic emission on core samples cut from different directions and the principal directions of the strain due to the closure of the loaded cyclically in uniaxial micro-cracks. [34]). There . (8) core recovery time and made. The following nine parameters could significantly limit the application of the ASR method. Geophysical methods with other methods. Differential strain curve analysis (DSCA) 1950 on acoustic emission of rock This method is based on the concept that careful monitoring of revealed that when the stress on the strain behaviour of a rock specimen upon re-loading can reflect rock relaxed from a certain level its past stress history.8. Major limitations include that it includesKaiser effect was conducted by measurements on a micro-scale and involves very smallHolcomb in 1993. (4) non-homogeneous recovery deformation. the stresses deter-mined by AE were consistent with the stresses from the overcoring method and hydraulic fracturing method. exceeds its previous higher value. which has been investigated overcoring. the research carried out by Holcomb revealed The following geophysical methods for estimation of in situ that using the acoustic emission emitted during uniaxial compresstresses have been reported in the literature: sion laboratory tests to infer in situ 0* seismic and micro-seismic methods (i. mentioned method (the Kaiser Studies show that the stress orientation obtained with the ASR method did not coincide well with that determined byeffect). once one principal stress is known (e.986 C. practice. stress experienced by a rock in situ To determine the in situ stresses. The 0* The ratios between the three principal stresses are to be related to the ratios between the three principalmaximum stress that a rock specimen has been subjected to strains due to the closure of the micro-cracks. The techniques listed above have rock anisotropy. good correlation between stresses determined with Kaiser effect and 2. the other two principal stresses can be determined. Research originally conducted by Kaiser in 2. 0* radio-isotope [36]. dehydration of core samples.g. however. (3) pore fluid (2 ) pressure diffusion. Matsuki and Takeuchi [33]. Utagawa in 1997 concluded that 0* sonic and ultrasonic methods [35]. An emission (AE). method 0* atomic magnetic method [37]. Some remarks need to be (7 ) residual strains. By subjecting the core to hydrostatic loading. The method is only 2D. Studies have found that theencouraging results obtained by stress orientations determined with the DSCA method several authors showing a fairly correlated poorly with those determined by the over-coring. advantage with the method is that it may be applied to An extensive review of the estimate stresses at large depth as long as it is possible to different studies conducted on the recover a core. 1950—now). stresses could not be justified. Martin et al. assumingspecimen to the point where is a the vertical stress is equal to the weight of the overburdensubstantial increase in acoustic rock).

opening or closure of micro-cracks in The measurements are made on a total of 5 cubic pieces. measured instead of the strain. The analysis is based on [39]. with a the pieces. It is not a placed in a pressure cell and directly commercially available method and there still exists a subjected to a defined hydrostatic discussion on its justification as a method to determine loading while the transmission times for the compres-sion and shear waves stresses. The principles of the method are the same as the DSCA the change and different in velocity of except that the compression and shear wave propagations arethe waves during the loading steps. A method called RACOS is a relative new geophysical method to are measured between opposing determination the in situ stress magnitudes and their orientations block face. The pieces are pressure and estimated vertical stress from AE.was also significant correlation between the overburden length of 25 mm. The method has . The method is based on the idea The velocity change depends on the that the re-loading on a sample can reflect it’s past stress history.

that most geological structures were formed a long timeuse the method to the ago and the mechanism controlling the stress field at that time may have changed estimate directions of the significantly since then. Can be applied when high stresses exist and overcoring and hydraulic fracturing fail Existing information. however. Involves a fairly large rock volume. erosion may play a role. Requires drill rig Only 2D. it is of most interest in situations where both overcoring and hydraulic fracturing fail Estimation of stress at early stage Borehole breakouts 2D Focal mechanisms Kaiser effects ASR/DSCA/ RACOS Back calculation Analysis of geological data 2D 2D/3D 2D/3D For great depths Simple measurements Usable for great depths 2D 2D/3D Quick and simple. the method only gives 2. the structures.C.11. If a acquisition of the world. caution must be used if stress information is estimated from geologicalseismic network exists. To obtain stress profiles HTPF 2D/3D Core discing 2D Since the method is time consuming. sensitive to several factors Theoretically not unique solution Very rough estimation. Relatively quick Limitations Scattering due to small rock volume. It should be pointed out. volcanoes. High certainty due to large rock volume Low cost Restricted to information on orientation. these analyses have It is. Plate tectonic driving forces may have had a different pattern than today. depth down to 1000 m For weak or high stressed rocks Shallow to deep measurements. The theoretical limitations in the evaluation of sH: Disturbs water chemistry Time-consuming. low reliability Occurs mostly in deep holes Rough estimations Estimation of stress state at great depth Can only be used during construction of the rock cavern At early stage of project Several been applied to sedimentary rocks and metamorphic rock but the results have not beenstress. Quick Measurements in existing hole. Earthquake focal mechanisms on a continuous basis if earthquakes occur. attempts have been made to correlate stress magnitudes to data from but From the literature there can be found numerous suggestions for estimating orientationsearthquakes and/or magni-tudes of the stress field from geological structures. however. which is obtained already at the drilling stage Existing information obtained at an early stage. method allows monitoring of events 2. Seismologists have studied the relation between the fault slip occurring during an event and As earthquakes the state of normally occur at great depth. possible to Stephansson [1]. A discus-sion of the subject can be found in Amadei and limitations. lineaments. Hence. Requires drill rig Only 2D.10. dikes. Requires existing fractures in the hole with varying strikes and dips Only qualitative estimation Suitable for 987 Measurements. Ljunggren et al. Theory needs to be further developed to infer the stress magnitude Information only from great depths Relatively low reliability Complicated measurements on the micro-scale. for example faults. compared with any established stress measurement method. Geological observational methods . / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 Table 4 Stress measurement methods and key issues related to their applicability Method Overcoring Doorstopper Hydraulic fracturing 2D 2D 2D/3D 2D/3D Advantages Most developed technique in both theory and practice Works in jointed and high stressed rocks Measurements in existing hole. and a number of glacial periods have passed in some parts principal stresses. Low scattering in the results.

0* A few methods dominate the others. they have been applied world wide. the methods have been benchmarked against each other.e. 3. i. These methods are commercially available. many reference data exist. When studying the stress measurement methods available. which may not be relevant for engineering at shallow depth. between 0 and 1000 m. their accuracy given ideal conditions is well known. Summary of methods and conclusions The previous sections have given a brief review of the most common methods to determine or estimate information about the in situ rock stresses.information on the stress orientations at great depths. they . the following can be concluded.

Klee G. Report No 06819-REP-0120010019-R00. overcoring. Swedish Nuclear Fuel and Waste Management Co. Lulea( University. Lisbon. Amadei B. 77–82. Swedish Rock Engineering Research. Bjarnason B. 1986. Oskarshamn. 1998–1999. Sweden. Ljunggren C. In-situ stress determination by hydraulic tests on pre-existing fractures at Gidea( test site. 1997. Sweden. Posiva Oy. Martino J. Thompson P. p. 1990:82 D. Stripa Project 8713.65:45–114. Stephansson O. Finland. 63pp. 1998. Leite MH. Nguyen D. Int J Rock Mech Min Sci . Sugawara. Obara Y. 18] 19] 8] Ljunggren C. Corthesy! R. 2002. et al. Uncertainty in the maximum principal stress estimated from hydraulic fracturing measurements due to the presence of the of the induced fracture. the evaluation of focal mechanisms and most of the geophysical methods. Japan. Gill DE. Application of the integrated stress measurements strategy to 9 km depth in the KTB boreholes. 0* There exists a number of methods that are based on the interpretation of geological and/or stress related phenomena where no actual measurement of the stress field is performed. p. Leeman ER. Leite MH. Japan. 1986:12L. the doorstopper technique and back analysis. J S Afr Inst Min Metall 1964. Leeman ER. Corthesy R. Andersson J. An integrated approach to rock stress measurement in anisotropic non-linear elastic rock. Corthesy R. Lulea( University of Technology. but the information may be obtained at a lesser price and they may also serve as a complement to the true measurement methods. 9] 10] 11] 12] 0* A category of methods includes those that do not have the general advantages of the most common true measurement methods but which under certain circumstances may be of great value. Stockholm. The mesurement of stress in rock—Part 1. 1995. Nac.. Thompson PM. Normally these methods are less accurate. borehole breakouts. och overborrning. SveBeFo Rapport 37. Application of the Doorstopper gauge system to deep in situ rock stress determinations. 1987. 1997. Raillard G. 1990. 1983. 1986. Improvement in accuracy of the conical-ended borehole technique. SKB. A review of recent developments. The methods in this category are hydraulic fracturing. Raillard G. Norway: NTH and SINTEF Publishers. Ljunggren et al. Rock stress measurements in borehole V3. In: Proceedings of the International Symposium on the Determination of Stresses in Rock Masses. Bergspanningsm. Canada. Brudy M. Ljunggren C. Lecture Notes in Engineering. and analysis of geological structures. Ottawa: CANMET. Licentiate thesis. Working Report 99-54. Proceedings of Workshop on Rock Stresses in the North Sea. Stephansson O.988 C. Rock stress measurements by means of hydraulic fracturing in Borehole KOV01. Rutqvist J. 79pp. Lulea( University. The methods are summarised in Table 4 with focus on some of the key parameters. Berlin: Springer. IPR-02-01. 1990. Trondheim. Bjarnason B. The measurement of stress in rock.atningars. 59–64. Typical methods that fall into this category are core discing. 1971. 13] 14] 15] References 1] 2] Amadei B.. 1997.37:107–20. Doctoral thesis. The modified Doorstopper cell stress measuring technique. New developments in hydrofracturing stress measurement techniques and a comparison to the HTPF method. the DSCA method. For certain conditions. Atomic Energy of Canada Lim-ited. Sweden. Gill DE. Chang Y. Sweden. Rock stress and its measurement.89:11527–37. and have been developed for the purpose of determining either the 2D or 3D stress field parameters. Ljunggren C. p. 77pp. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 have been practised for many years.. London: Chapman & Hall. these methods have not yet found their full commercial platform and are the ASR method. these methods may also be the ones to use to obtain stress field information. Rock stress measurements by means of hydraulic methods at Hastholmen. In: Proceedings of the Conference on Stresses in Underground Structures. Sweden. Tsang C-F. 23–32. Cornet FH. Rock anisotropy and the theory of stress measurements. De Eng. Int J Rock Mech Min Sci 2000. Valette B. Lab. as compared to the true measurement method mentioned above. Research Report TULEA 1986:22. p. 40pp. In situ stress determination from hydraulic injection test data. 200– 29. In: Proceedings of the International Symposium Rock Stress. Civil. 122pp. (in Swedish). Weber U. In: Proceedings of the International Symposium Rock Stress. 154–64. och naturliga variationer vid hydraulisk sprackning. Rock stress measurements at great depth using the modified doorstopper gauge. representativitet—Matnoggrannhet. 2000. Mostly. 16] 3] 17] 4] 5] 6] 7] Rummel F. 1999. J Geophys Res 1984. p. Hydrofracturing rock stress measurements in the Baltic Shield.

Pau. Determination of the natural stress state in a Brazilian rock mass by back analysing excavation measurements: a case study. Rummel F. 1997. 26] Shamir G. 1995. 22] de Mello Franco JA. Int J Rock Mech Min Sci 2002. 24] Zoback ML. 21] Bock H. 433–43 [Chapter 16]. Rotterdam: Balkema. Estimating in situ stress conditions from borehole breakouts and coredisking—experiment results in granite. Toshinori S. editor. In: Proceedings of the Workshop on Rock Stresses in the North Sea. Nakamura N. Savilahti T. Hydrofrac stress profile to 9 km at the German KTB site. Santiago JAF. Nature 1989. Lee CF. Japan. Norway: NTH and SINTEF Publishers. 25] Te Kamp L. In: Proceedings of the International Symposium on Rock Stress. vol. Trondheim. Global pattern of tectonic stress. Tokyo.30:395–411. Oxford: Pergamon Press. Stress orientation profile to 3. 7–10 October. Rock mechanics of the deep borehole at Gravberg. Zoback MD.5 km depth near the San Andreas fault at Cajon Pass. California. 27] Stephansson O. Measuring in-situ rock stresses by borehole slotting. vol. Eighth ISRM Congress. Mansur WJ. p. In: Hudson JA. 2. Zoback MD. 3.341:291–8. J Gephys Res 1992. p. 1993. 23] Haimson BC. Bjarnason B. Japan. . 20] Matsui H. Comprehensive rock engineering.39(8):1005–32. 147–53. 1989.Geomech Abstr 1993. Telles JCF. p. In: International Symposium Rock at Great Depth.97:5059–80. 19–24. Sugihara K. Kumamoto. p. Comparison of the results of stress measurements determined by various methods at the Kamaishi mine. 95–100. Sweden. 863–70. et al. In: Proceedings of the International Workshop on Rock Stress Measurement at Great Depth. p. Armelin JL. 1995.

34th US Symposium on Rock Mechanics. Heidelberg. vol. Use of changes in electrical conductivity of rock to study the stress state in the rock mass and its aquifer properties. 1999. Morgantown. 37] Cook JC. p. 477–84. 1967. an overview. Development of in-situ stress measurement technique using ultrasonic wave attenuation method—a progress report. p. Semi-annual report on electronic measurements of rock stress. 1993. p. Stockholm. 39] Braun R. p. Three-dimensional in situ stress determination by anelastic strain recovery of a rock core. Rock Mechanics. Initial stress back analyzed from displacements due to underground excavations. Germany. Numerical study on core damage and interpretation of in situ state of stress. 29] Haimson B. In: Proceed-ings of the ISRM International Symposium on Integral Approach 989 Proceedings of the 34th US Symposium on Rock Mechanics. Jahns E. 234pp. Japan. Heidelberg Academy of Science and Humanities. 1994. Rock Mechanics Contributions and Challenges. Shimizu N. 30] Sakurai S. 19pp. Moscow: Nauka Press. Trudy VNIMI 1961. In: Proceedings of the International Symposium on Rock Stress and Rock Stress Measurements. p.C. Takeuchi K. 34] Martin CD. Rotterdam: Balkema. 72-10. Determination of in-situ stress magnitudes and orientations with RACOS. 35–43. 679–86. Lang PA. In: Proceedings of the 30th US Symposium on Rock Mechanics. US Bureau of Mines Technical Report No. 38] Petukhov IM. 42:110–8. 1990. / International Journal of Rock Mechanics & Mining Sciences 40 (2003) 975–989 28] Hakala M. 33] Matsuki K. Determination of in-situ stress from partial anelastic strain recovery measurements of oriented cores from deep boreholes. 1. In: Lecture notes of the short course in modern in situ stress measurement methods. 15–26. The World Stress Map and Beyond. Marmorshteyn LM. Madison. 1997. 1998. . Seven years of in situ stress measurements at the URL. In: 36] Riznichanko YV. 31] Sakurai S. Ljunggren et al. Lulea:( Centek Publishers. to 135–43. Santiago. 1972. 1998. Chile. Euroconference on Earth Stress and Industry. Stromeyer D. Morozov GI. Borehole breakouts and core diskning as tools for estimating in situ stress in deep holes. September 3–5. Posiva Report 99-25. Madison. 1989. Study of rock stress by geophysical methods. Read RS. 1986. Back analysis of in-situ stresses in a rock mass taking into account its non-elastic behaviour. Applied 32] Teufel LW. 1993. Akutagawa S. Peng SS. 35] Sun YL. In: Proceedings of the International Symposium Rock Stress.