Pressuremeter testing in Ruritania

Pressuremeter testing in Ruritania A compilation of the results of ten tests in a variety of materials, selected to show what can be derived from careful pressuremeter testing Reference: CIR 2001/11 Part 1 Text report and result summaries CAMBRIDGE INSITU Ltd. Little Eversden Cambridge ENGLAND CB23 1HE Tel:- (01223) 262361 Fax:- (01223) 263947 Email:- caminsitu@aol.com Web site:- www.cambridge-insitu.com

Weak Rock Self-boring Pressuremeter and High Pressure Dilatometer Tests Cambridge Insitu Ltd March 2011 CONTENTS PART 1 WRITTEN REPORT AND RESULTS SUMMARY 1 INTRODUCTION 1.1 Instrument 1.2 Analysis 1.3 Report 1.4 Notation 1.5 Units 1.6 Personnel 1.7 Headers and Footers 2 A LIST OF TESTS CARRIED OUT 3 COMMENTS ON THE TESTS AND THE ANALYSIS 3.1 Test comments 3.2 Shear modulus 3.3 Shear strength 3.4 Insitu stress 3.5 Friction angle 3.6 Individual tests 4 SUMMARY OF RESULTS 4.1 Table 1 General results stresses 4.2 Table 2 Water pressures 4.3 Table 3 General results strength and stiffness 4.4 Table 4 Shear modulus parameters 4.5 Table 5 Frictional parameters 4.6 Plots of various parameters against depth 4.7 Plots of the variation of modulus with strain PART 2 FULL RESULTS AND PLOTS PART 3 APPENDICES The appendices contain background information on the instruments, the test procedures and the analyses. There is also a list of references not complete, but hopefully nearly so!

Weak Rock Self-boring Pressuremeter and High Pressure Dilatometer Tests Cambridge Insitu Ltd March 2011 Typical Geotechnical Investigations 1. Introduction Cambridge Insitu Ltd (CI) are usually subcontracted by the Main Contractor to carry out pressuremeter testing as part of a larger program of site investigation. This site investigation, of a mythical borehole in Ruritania, is intended to give prospective users and other interested parties a clearer idea of what information the pressuremeter test can provide. Many testing programs are aimed at providing particular insitu engineering parameters for the mechanical properties of the ground, often with emphasis on stiffness. The way the tests are run can be tailored to specific requirements of the site. The field work took place in April and May 2006. Two types of pressuremeter were used, a Self Boring Pressuremeter (SBP) and a High Pressure Dilatometer (HPD), which carries out a pre-bored pressuremeter test in a 101mm or H size pocket. Ten tests are shown in one borehole at depths between 5 and 80 metres below ground level. All of the tests were completed successfully. Details of the testing programme are given in part two of this volume. Descriptions of the material can usually be found in information provided by the main contractor, which includes details of the borehole location. This report is concerned solely with the presentation of Pressuremeter test results. Preliminary results are usually presented soon after completion of the testing. All such results are superseded by the values given in this report, though the changes are unlikely to be anything other than minor. 1.1 Instrument The Cambridge Self Boring Pressuremeter was invented in the early 1970 s by Hughes and Wroth to carry out minimally disturbed tests in soft clays and loose sands. The test is a two part process. The instrument is a miniature tunnelling machine, the central part of which is covered with an elastic membrane. The first part of an SBP test is concerned with inserting the instrument into the ground to the depth where the test is intended. The second part consists of a controlled inflation of the elastic membrane so that the instrument carries out a lateral loading of the test cavity. The quality of the test is dependent on the quality of the self boring. If hard layers, or more typically, gravel bands are encountered the resulting damage to the cutting parts means the test is compromised. However in a uniform material free of hazards, the SBP test is the best that can be achieved with a Pressuremeter. On this contract a 3 arm version was used for all except the final SBP test, where the 6 arm version was used. The HPD used was a 95mm diameter Cambridge High Pressure Dilatometer. This instrument, based on an earlier design by Dr J.M.O Hughes, was developed to carry out pressuremeter tests in soft to weak rock. In use the instrument is lowered into a 101mm pocket made by rotary coring or rock rollering. Although developed to test ground of the strength of weak rock, the pressure and displacement resolution of the instrument is such that it can operate at two extremes of ground conditions. The first is strong rock, where it is likely the 20MPa pressure capability of the instrument will determine the end of the test and only elastic deformations will be suffered by the intro.doc Print date: 28-03-11 Volume 1 Section 1 Page 1 of 3 in this section

Weak Rock Self-boring Pressuremeter and High Pressure Dilatometer Tests Cambridge Insitu Ltd March 2011 ground. The second condition is typically firm to very stiff clay, when the soil will experience substantial plastic deformation at modest pressures, and the strain range of the probe will determine the limit of the test. The pre-bored installation method is best suited to material where the unloading of the cavity wall prior to placing the probe will not result in a collapse of the material or cause reverse plastic failure. Note that for two of the tests, in soft clay, the HPD has been used because of the presence of gravel. The testing system for both types of Pressuremeter is the same. Once in position pressure is applied down an umbilical cord and a membrane covering the central part of the probe inflates and thus loads the borehole wall. The pressurising medium may be gas or oil, depending on the ground conditions and the maximum pressure desired. The expansion of the membrane is monitored by high resolution displacement sensors and the pressure applied is measured by transducers in the probe. The output of the probe is a stream of digital data, which when converted to engineering units gives a pressure/displacement curve of the horizontally oriented loading test. The Cambridge family of Pressuremeters are complex instruments by normal site standards; using strain gauged transducers throughout and microprocessor controlled data acquisition systems inside the probes themselves. 1.2 Analysis The pressuremeter loading curve can be solved directly using mathematical expressions for the expansion of a cylindrical cavity. The solution is conventionally quoted in terms of strength parameters for the material, specifically shear modulus, shear strength or friction angle as appropriate, and the insitu lateral stress. This fundamental approach is not the only way to interpret pressuremeter data, but is the common practice in the UK. The success of this method is dependent on the validity of the assumptions that have to be made: Assumptions about the soil response include that the material is fully saturated, homogeneous, isotropic and behaving as a continuum that fails in shear only. An important assumption about the instrument is that the length to diameter ratio of the expanding section is large enough for end effects to be negligible, so allowing the test to be modelled as a plane strain expansion. A further major assumption concerning the test procedure is that the loading path be either undrained or fully drained. If undrained the loading takes place at constant volume and only shear strains need be considered. This loading path would be appropriate for a test in clay. If the expansion is drained, then the material is treated as behaving with the characteristics of sand and shear and volumetric strains must be accounted for. The Self Boring Pressuremeter tests have been analysed here as either drained or undrained expansions, depending on the material type. All the later High Pressure Dilatometer tests are more amenable to drained (ie frictional) analysis, but the first two present a problem. The material is clay, and so basically undrained, but the gravel imparts a frictional behaviour. Both types of analysis have been included. intro.doc Print date: 28-03-11 Volume 1 Section 1 Page 2 of 3 in this section

Weak Rock Self-boring Pressuremeter and High Pressure Dilatometer Tests Cambridge Insitu Ltd March 2011 1.3 Report Although it is necessary to make judgements when analysing the data, this remains a factual report. The parameters derived represent what seems a reasonable choice having applied a particular analysis. However other choices are certainly possible, and the intention is that this report provides a full description of the tests and analytical methods employed so that the choices made here can be checked by the user and modified if so wished. Significant details of individual tests have been mentioned, and this includes any problems with the equipment or the procedure that may have influenced the results. Note that the original test data is available on disk as files of readings in engineering units in a format easily accessed by several common spreadsheet programs. The reported results are extracted from a number of supporting files. For the most part these are in a private format that can only be accessed by Cambridge Insitu software. However there are some EXCEL spreadsheets that may be useful in understanding how some of the results are obtained. Data is also available in AGS format. 1.4 Notation The data collection system employed on site utilises a limited keyboard that restricts the options for describing a test. In particular it stores tests in the form B** T** where ** must be a number. The B is intended to refer to the borehole and the T refers to the individual test. For this contract the pressuremeter borehole is known as B1. Our internal reference is therefore B1T2 for the second test in the borehole. Calibration tests to evaluate membrane stiffness and compression are reported in a similar manner, but using a test number that cannot be confused with an actual test. A typical calibration test reference is S2104T10. 1.5 Units Pressure is quoted throughout in Pascals. The smallest unit of pressure quoted is 1 kpa. Displacements are quoted in millimetres. Once an estimate of the insitu lateral stress has been made, allowing the original cavity diameter to be inferred, then the deflections are converted to percent cavity strain. Lengths are quoted in metres. Depths are referred to metres below original ground level. 1.6 Personnel This report was compiled by Philip Hawkins of Cambridge Insitu, using data from a variety of sources. 1.7 Headers and Footers The header used on every page of this text report refers to the contract. The footer is for CI internal use only and refers to the document name and version number. Note that comments in italics are extra explanations or comments that would not normally appear in a report. They often refer to unusual events, or are intended to explain certain peculiarities. intro.doc Print date: 28-03-11 Volume 1 Section 1 Page 3 of 3 in this section

Weak Rock Self-boring Pressuremeter and High Pressure Dilatometer Tests Cambridge Insitu Ltd March 2011 2. A list of tests carried out Note that this is not a real borehole the tests have been chosen to represent a possible geological sequence, and to demonstrate some of the advantages (and difficulties) of pressuremeter testing. Test Date Test depth (mbgl) Material Probe Maximum pressure (MPa) Remarks B1T1 01/04/06 5.0 Brown clay SBP3 0.36 Above water table. May be sandy? B1T2 05/04/06 6.1 Clay/gravel HPD 0.53 Large pocket. B1T3 8.0 0.76 Tight pocket. B1T4 07/04/06 10.0 Coarse sand SBP3 1.14 Full unloading in middle. B1T5 08/04/06 20.0 Grey clay 1.65 Holding test. B1T6 10/04/06 25.0 2.86 A lot of creep. Loops on loading and unloading. B1T7 12/04/06 37.5 Silty sand SBP6 9.35 Ran out of gas. B1T8 07/05/06 50.8 Chalk HPD 14.70 Fails slowly at first, then more rapidly. B1T9 10/05/06 67.6 Limestone 14.37 Cracks after failure. B1T10 11/05/06 80.4 Phyllite 20.03 Does not fail. Tests are listed in chronological order.

Weak Rock Self-boring Pressuremeter and High Pressure Dilatometer Tests Cambridge Insitu Ltd March 2011 3. Site Specific Commentary This part of the report collects together comments and observations that have arisen out of the fieldwork and subsequent analysis. 3.1 Test comments These tests, although presented as if in a single borehole, have been chosen from a variety of sites to provide representative examples. All show something slightly unusual, or something that is not part of a normal test. The tests in this borehole vary from soft clay to hard rock. In all there are five tests in clay, two in sand and three in various types of rock. The tests in clay and sand were performed using a Self-boring Pressuremeter, those in rock using a High Pressure Dilatometer. The exception to this are the two tests in gravely clay. The gravel precludes self-boring and the only way to get a test is to use the HPD, even though the pressures involved are very low. The HPD tests were in a pocket formed using an H size core barrel. The diameter of the pocket is nominally 101mm, but other sizes may sometimes be used. Note that the term dilatometer is the general name in rock mechanics for the type of instrument known as a pressuremeter in soil mechanics. It should not be confused with Marchetti s flat dilatometer which is a completely different device. All tests have multiple reload loops, with one or two on the unloading for two of them. All of the tests have a smooth unloading path, which, where the material has started to fail, can be used to estimate the shear strength. 3.2 Shear Modulus The way the modulus changes from one loop to the next depends on the soil type. In clays all loops, except possibly the first if it is too close to the origin, give similar results. In sand, or other frictional materials, the modulus increases with successive loops. This is because the modulus depends on the mean effective stress. In saturated clays the excess pore pressures caused by the expansion follow the total pressure almost exactly thus the effective stress remains constant. In sands, where drainage occurs, no excess pore pressures are generated and the effective stress follows the total pressure. Note too that many materials lie between these two extremes. Rocks, like chalk and limestone, show basically frictional behaviour and although relatively impermeable the modulus increases with the applied stress. For SBPM tests in clay the initial modulus is usually comparable to that from the reload loops, but is not always clearly defined. In sands it tends to be a lot less. sitecom.doc Print date: 28-03-11 Volume 1 Section 3 Page 1 of 5 in this section

Weak Rock Self-boring Pressuremeter and High Pressure Dilatometer Tests Cambridge Insitu Ltd March 2011 For HPD tests in rock the initial modulus is often of similar magnitude to that from the reload loops and can often be used to derive a realistic value for the Insitu stress, by comparison with them. This has not been pursued here. All tests except the last two show non-linearity. 3.3 Shear Strength This concept is strictly only applicable to clays, and clay-like materials. Frictional materials show a steadily increasing shear stress throughout the expansion, although some do show a flattening off at large strain. On the unloading, however, all materials show a fairly constant shear stress for much of the curve. It is this value that has been quoted. Note that the Gibson & Anderson method used for the expansion and the Jefferies method used for the contraction are strictly comparable. Each uses the slope of a straight line drawn on a semi-log plot to give the shear strength. The Palmer method gives a continuous plot of Shear Stress against Strain, and may be used on both loading and unloading. Unfortunately it tends to give rather a noisy plot, and is seriously disturbed by reload loops. These plots may be provided in exceptional circumstances. 3.4 Insitu Stress The standard technique for determining Insitu Stress is based on the method of Marsland & Randolph, and involves an iterative procedure performed by the operator. It is, therefore, subjective. It requires that the material has begun to fail before the maximum pressure is reached, which applies to all of these tests except the last. It also includes a contribution from the Cohesion, which may be large in rock. A new curve matching technique for frictional materials may be used here. One of the problems with this is that the same match can be achieved using a number of different combination of parameters. Use of the Manassero plot gives a value for the friction angle, and it is possible to separate out the Insitu Stress from the Cohesion. Note that the sum of these two should be similar to that from the Marsland & Randolph procedure. 3.5 Friction angle The method of Hughes, Wroth and Windle requires a value for the friction angle at constant volume φ cv. The values used can be based on past experience or on the results of other types of test. For small differences, the friction and dilation change in concert with φ cv. Using a combination of the Manassero result with the Carter et al curve matching procedure may allow an independent assessment of the peak friction angle and the dilation. This is still experimental at the moment. sitecom.doc Print date: 28-03-11 Volume 1 Section 3 Page 2 of 5 in this section

Weak Rock Self-boring Pressuremeter and High Pressure Dilatometer Tests Cambridge Insitu Ltd March 2011 Individual tests B1T1 A test in soft clay that is apparently quite typical but shows a couple of oddities. The response of the pore pressure cells is initially very good, but then the excess pressures start to drop away. The reload loops give moduli that steadily increase, and although the curve matching is good the modulus again increases. The material may be sandy, but it may just be because it is above the water table and is not fully saturated. Note that the curve matching shows a large zero offset. This may be genuine, but is more likely to be due to the intermediate nature of the material. B1T2 The first of a pair of tests using the HPD. The material is soft clay, but the presence of gravel precludes the use of the SBPM. This test has a large pocket, with the probe expanding a long way before encountering the borehole wall. A reload loop in the initial, very soft mush has been ignored. Curve matching has been tried, but the results look dubious, again with a very large zero offset. B1T3 Similar material to the previous test, but the presence of less gravel allows it to close back in the probe had to be pushed into the pocket. The curve matching gives similar results to the previous test. The curve matching procedure gives similar results for these two tests, even though they look very different. The Carter et al procedure for curve matching in frictional materials has also been tried on these tests. There is not such good agreement, but it is still better than from those analyses that rely on just the loading path. B1T4 This test is in coarse sand, still above the water table. All arms are disturbed, which is not unexpected in a material like this. The unusual thing about this test is the complete unloading in the middle, after the third reload loop. This was specifically asked for in the specification. Note however that it does not affect the analysis the loading curve after this unloading carries on from the original loading, as can be seen on the Hughes plot. The moduli from the reload loops increase markedly each time. Note that the unloading apparently occurs at a pressure slightly below zero. This is due to the compromise involved in measuring the membrane correction the usual technique is to draw the calibration line halfway between the loading and unloading paths, which inevitably makes the unloading correction slightly too large. It is only because this test is above the water table, and the unloading should be at zero pressure, that this becomes noticeable. B1T5 The first of a pair of tests in stiff grey clay. The drilling does not seem to have gone very well, because the initial pore pressures are very high, as is the liftoff. The test is fine and includes a long hold at constant strain. The decay of pressures during this hold allows the determination of the consolidation, and from that the permeability. A new method, involving matching computer generated decays to the actual ones, has been used here. The opportunity was taken to do permeability tests sitecom.doc Print date: 28-03-11 Volume 1 Section 3 Page 3 of 5 in this section

Weak Rock Self-boring Pressuremeter and High Pressure Dilatometer Tests Cambridge Insitu Ltd March 2011 while the probe was in position, pumping water at a controlled rate down the drill string and out of the cutting shoe. The results are very similar to that from the holding test. This test is a problem to analyse, because of the high initial pore pressures. Other methods are available for analysing the holding test, some of which are more tolerant of this. Some examples are given in Appendix E. B1T6 The clay here is much stiffer, but the drilling appears to have gone better as the excess pore pressure generated is much lower. The pore pressure response is good, but starts dropping away during the hold for the first reload loop. The holds in this test are longer, and there is a lot of creep particularly before the second loop. One consequence of this is that Arm 2 reaches the limit of its travel. The other two loops are on the unloading (again specifically requested) and it can be clearly seen that the creep is much less. The last loop gives a slightly lower modulus than the others, possibly being too far down the unloading curve? Loops on the unloading have now become more common, for the reasons seen here. They give results similar to those on the loading and do not interfere with the main curve in the same way. The Gibson and Anderson plot steps across at each hold, but each section has a similar slope. The curve matching is good, as long as all of the loading curve after the first hold is ignored. These are both the result of the drainage. The creep has been analysed. After two or three minutes it settles down, and a plot of creep against the log of the time from the start of the hold becomes straight. The slope of this line allows the magnitude of the creep to be estimated over long periods. Note however that it is very stress dependent. B1T7 A second test in sand, this time dense silty sand that pushes the capabilities of the probe to the limit note that the supply of gas ran out at just less than the maximum pressure rating. The probe this time is the six arm version, and slight difference can be seen in the arm pairs. This may be due to anisotropy, but it is difficult to determine the orientation at this depth. It is possible to measure the orientation of the HPD because the electronics assembly, which contains a compass, is below the main body of the probe. The SBPM electronics are alongside the stainless steel central tube, and hence shielded. Some examples of the SBPM contain an inclinometer, and one of our HPDs has been so modified. This can help show if the borehole has deviated from the vertical, or determine the orientation of the probe in a horizontal borehole. The Hughes plot is again straight, showing no sign of flattening off, as often happens at large strain. It is now common to use the HPD in dense sand, one advantage being that it is possible to push the test past the peak friction stage. Testing with the HPD is usually quicker than with the SBPM a big practical advantage. The main disadvantage is the lack of pore pressure cells, but as can be seen in the test shown here there is very little for them to measure. sitecom.doc Print date: 28-03-11 Volume 1 Section 3 Page 4 of 5 in this section

Weak Rock Self-boring Pressuremeter and High Pressure Dilatometer Tests Cambridge Insitu Ltd March 2011 B1T8 The first test in this borehole using the HPD for its pressure capacity. The material is chalk, which it is possible to self-bore, but there are flints as well which are not! This material would normally be expected to behave in a frictional manner, but here (after initially doing so) it then begins to expand rapidly and the behaviour reverts to clay-like. Notice how the modulus hardly increases from Loop 2 to Loop 3. B1T9 This limestone is a truly frictional material. The Hughes plot is straight (ie the friction angle is constant) but the line steps across at each break in the expansion curve. Each of these is caused by the rock cracking usually one arm is affected more than the others, showing the general direction of the cracking. The modulus here is about the largest that can normally be measured accurately with the HPD. The movement during the reload loops is similar to that inside a calibration cylinder, and so the accuracy of the correction is vitally important. B1T10 This is very hard rock. There is no sign of failure, even at the maximum rated pressure of the probe. The reload loops all lie on the main curve, which becomes straight above about 12MPa. The curvature below this is mainly due to the strips of the Chinese Lantern taking up the exact shape of the borehole wall. To achieve any degree of accuracy at this level of modulus, great care must be taken in determining the correction. The calibration cylinder used should be as similar as possible to the test conditions plain and of similar diameter, and a number of calibrations should be done to check that nothing has changed. For this particular test, individual calibrations were used for all six arms, as one in particular was more pressure sensitive than the others. The following plot shows how an error in the correction affects the accuracy of the measured modulus :- sitecom.doc Print date: 28-03-11 Volume 1 Section 3 Page 5 of 5 in this section

Weak Rock Self-boring Pressuremeter Tests Cambridge Insitu Ltd Typical tests in a borehole March 2011 4. Summary of Results 4.1 General results stresses Table 1a SBPM tests Test Depth Zero offset Total stresses (mtr) M&R Curve Arm 1 Arm 2 Arm 3 Arm 4 Arm 5 Arm 6 L/O M&R Curve match L/O L/O L/O L/O L/O L/O ave. match (bgl) (mm) (mm) (kpa) (kpa) (kpa) (kpa) (kpa) (kpa) (kpa) (kpa) (kpa) B1T1 5.0 0.05 1.13 135 48 39 - - - 74 75 106 B1T4 10.0 0.84-12 - - - - - - 147 - B1T5 20.0-0.01 0.01 960 975 1010 - - - 982 1000 780 B1T6 25.0 0.00 0.02 400 1575 960 - - - 978 1000 890 B1T7 37.5 1.32 - - - - - - - - 1690 - Table 1b HPD tests Test Depth Borehole (mtr) Notes on Table 1 :- Total stresses diameter M&R Curve match (bgl) (mm) (MPa) (MPa) B1T2 6.1 110.4 0.34 0.22 B1T3 8.0 96.8 0.34 0.24 B1T8 50.8 99.1 3.7 - B1T9 67.6 101.7 5.0 - B1T10 80.4 99.1 - - Note that other methods are sometimes available. A comparison of the Initial Modulus with that from the reload loops can be useful if there has been no degradation of the borehole wall due to stress relief. The curve matching technique, developed for use with sands, can help to separate the Insitu Stress from the Cohesion in rocks. The first two tests would normally use the SBPM but the gravel in the clay rules out self-boring. 1. L/O (Lift-off) The pressure at which the arm in question starts moving. Not always clearly defined. The average is usually similar to the value derived by the method of Marsland & Randolph, unless badly disturbed. 2. M & R These results for cavity reference pressure are derived from the Marsland & Randolph (1977) method. They are from the loading curve and hence affected by disturbance. Note that for rocks this value includes the cohesion further analysis is required to separate this from the Insitu stress. See the note alongside Table 1b. No value is quoted for B1T8 as the rock did not fail, and the analysis is not possible. 3. Curve matching Using the shear strength derived from the unloading, and the nonlinearity from the reload loops, the best possible match to the field data yields these results. Only gives realistic values if the expansion has been taken far enough to remove all the effects of disturbance, and only works for clays. Work is continuing on a similar approach for sands. 4. Zero offset Given a value for P o, a corresponding displacement is derived. This when added to the probe diameter gives the diameter of the cavity in its insitu condition. The process of curve matching involves adjusting the zero offset until the shear strength from the loading matches that from the unloading. The amount of this adjustment is another indication of the amount of drilling disturbance. Note that with the over-size shoe edge, there should theoretically be 0.5mm clearance round the probe. 5. Borehole diameter The size of the test pocket, being the diameter of the probe plus twice the zero offset. This diameter may give an indication of the quality of the pocket. RESULTS.doc Print date: 22-03-11 Volume 1 Section 4 Page 1 of 22 in this section

Weak Rock Self-boring Pressuremeter Tests Cambridge Insitu Ltd Typical tests in a borehole March 2011 Table 2 Water pressures Test Depth PPC A PPC B PPC average Unloading Uo Full b/h (mtr) (kpa) (kpa) (kpa) (kpa) (kpa) (kpa) SBPM tests B1T1 5.0 57 54 56-0 59 B1T4 10.0 0 0 0 0 0 88 B1T5 20.0 775 805 790-88 206 B1T6 25.0 102 558 330-137 255 B1T7 37.5 350 136 243 315 260 378 HPD tests B1T2 6.1 - - - - 0 70 B1T3 8.0 - - - - 0 88 B1T8 50.8 - - - 350 390 508 B1T9 67.6 - - - 650 555 673 B1T10 80.4 - - - 700 681 799 Notes on Table 2 :- 1. PPC A & B The water pressure measured by the pore pressure cells at the start of the test. This may be the local pore pressure in the ground, or may be the pressure due to the head of water above the probe. With the SBPM in rock machine configuration it is more likely to be the latter. The HPD does not have pore pressure cells. 2. Unloading The pressure at which the membrane collapses back in at the end of the test. This is only usually applicable to tests in sand or rock. 3. Uo The natural pore water pressure in the ground, assuming a hydrostatic profile from 11m below ground level. 4. Full borehole The water pressure exerted at the probe by a borehole full of water, allowing for a 1m stick-up of the casing above the ground surface. The derivation of Ko requires the calculation of Effective Stresses, both Vertical and Horizontal. The former may be obtained by subtracting the assumed ambient pore water pressure from the calculated vertical stress. The latter, however, relies on knowing which water pressure to subtract. The curve matching produces a value of Insitu Stress unaffected by drilling disturbance, and thus the ambient pore pressure should be appropriate. Both Liftoff and the Marsland and Randolph method are affected by disturbance, in particular the (usually) raised pore pressure. There is some evidence to show that subtracting the pressure measured at the start of the test gives results consistent with other methods as long as this pressure has been measured correctly. Unfortunately this is not usually the case see Note 1 above. RESULTS.doc Print date: 22-03-11 Volume 1 Section 4 Page 2 of 22 in this section

Weak Rock Self-boring Pressuremeter Tests Cambridge Insitu Ltd Typical tests in a borehole March 2011 Table 3 General results strength and stiffness Test Depth P max C u C u (Jeff.) P f P L P L (G&A) (curve) G i G ur1 G ur2 G ur3 G ur4 G MIN (G ur5 ) (mtr) (G&A) (bgl) (MPa) (kpa) (kpa) (kpa) (kpa) (kpa) (MPa) (MPa) (MPa) (MPa) (MPa) (MPa) SBPM tests (and HPD tests B1T2 & B1T3) B1T1 5.0 0.36 80 49 102 505 462 17.0 8.4 9.6 10.2-14.2 B1T2 6.1 0.53 156 77 500 925 767 2.3 11.1 17.1 16.4-25.7 B1T3 8.0 0.76 323 105 470 1086 989 1.0 16.3 23.4 25.5-27.7 B1T4 10.0 1.14-195 380 - - 9.6 26.6 44.9 51.4 74.8 - B1T5 20.0 1.65 235 225 1090 2115 2090 19.9 29.8 26.7 - - 17.3 B1T6 25.0 2.86 630 591 1446 4498 4380 70.1 79.4 64.7 75.2 66.8 52.2 B1T7 37.5 9.35-1850 4200 - - 130 313 354 394 396 (454) Test Depth P max C u C u P f P L P L G i G ur1 G ur2 G ur3 G ur4 G ur5 (mtr) (G&A) (Jeff.) (G&A) (curve) (bgl) (MPa) (MPa) (MPa) (MPa) (MPa) (MPa) (GPa) (GPa) (GPa) (GPa) (GPa) (GPa) HPD tests B1T8 50.8 14.70 4.27 3.68 5.50 26.2-0.33 1.02 1.07 1.03 - - B1T9 67.6 14.37 4.58 2.65 8.01 36.4-1.92 0.23 0.78 1.55 3.01 4.05 B1T10 80.4 20.03 - - >20 - - 10.5 5.22 7.24 9.31 - - Note the change in units for the HPD tests. Notes on Table 3 :- 1. P max This is the maximum pressure reached during the test. 2. C u This is the undrained shear strength, derived from the slope of a semi-log plot of total pressure versus current shear strain. Values are derived from the loading and from the unloading from the unloading only for sand and other frictional materials. Note that for tests where the membrane burst there is no unloading not applicable here. 3. P f This is the observed yield stress of the material, being the total stress where the loading curve departs from the initial straight line. In a linear elastic/perfectly plastic material this will be p o +c u. 4. P L This is the limit pressure of the material, from the intercept on a semi-log plot of total pressure versus current shear strain. Again two values are given, derived from the Gibson & Anderson plot and the curve matching. Note that the only real difference between them is that they use different values for the zero offset. Where there is no curve matching there is only one value, and for sands none at all. It is possible to derive a Limit Pressure for sands using some of the newer methods of analysis. 5. G i The shear modulus derived from the slope of the initial linear part of the expansion. Not always clear, as there may be no straight part of the curve. 6. G urn The shear modulus derived from the slope of the chord bisecting cycles of unloading and reloading. This procedure is valid provided the material response is linear elastic. If it is not, then a slightly more complex procedure is required and these results are given in Table 4 7. Curve matching Using the shear strength derived from the unloading, and the non-linearity from the reload loops, the best possible match to the field data yields these results. The method only gives realistic values if the expansion has been taken far enough to remove all the effects of disturbance, and only works for materials for which a mathematical model exists, ie clays. 8. G MIN The modulus at failure. Note that for B1T5 and B1T7 there are five loops, and the value for the final one is in this column. RESULTS.doc Print date: 22-03-11 Volume 1 Section 4 Page 3 of 22 in this section

Weak Rock Self-boring Pressuremeter Tests Cambridge Insitu Ltd Typical tests in a borehole March 2011 Table 4 Shear modulus parameters Linear elastic interpretation Non-linear elastic interpretation Test Loop Value P Start P β α G S G S G S G MIN (γ=10-4 ) (γ=10-3 ) (γ=10-2 ) (MPa) (MPa) (MPa) (MPa) (MPa) (MPa) (MPa) (MPa) B1T1 1 8.4 0.20 0.10 0.597 0.87 36 14 6 2 9.6 0.27 0.12 0.605 1.04 40 16 6 3 10.2 0.32 0.14 0.611 1.24 45 18 7 Curve matching :- 1.57 56 23-14.2@0.3% B1T2 1 11.1 0.28 0.09 0.853 4.32 17 12-2 17.1 0.41 0.12 0.833 6.61 31 21-3 16.4 0.20-0.12 0.785 5.90 43 26 - Curve matching :- 7.37 53 32-25.7@0.3% B1T3 1 16.3 0.51 0.17 0.649 1.91 48 22-2 23.4 0.67 0.21 0.670 3.08 64 30-3 25.5 0.37-0.18 0.649 3.94 100 44 - Curve matching :- 3.92 99 44-27.7@0.4% B1T4 1 26.6 0.15 0.10 0.858 10.4 39 28 20 2 28.4 0.28 0.13 0.852 17.2 67 48 34 3 21.4 0.48 0.22 0.859 22.3 82 59 43 4 21.4 0.96 0.29 0.845 28.8 120 84 59 B1T5 1 29.8 1.21 0.26 0.674 4.43 89 42 20 2 26.8 1.33 0.30 0.672 4.26 87 41 19 Curve matching :- 4.17 85 40 19 17.3@1.3% B1T6 1 79.4 2.34 0.68 0.709 14.9 218 112 57 2 64.7 2.68 0.72 0.701 13.8 217 109 55 3 75.2 1.69-0.59 0.668 11.2 238 111 52 4 66.8 1.04-0.66 0.658 9.52 222 101 46 Curve matching :- 0.701 13.7 215 108 54 52.2@1.1% B1T7 1 313 1.50 0.60 0.916 157 340 280 (231) 2 354 2.51 0.98 0.839 117 515 355 (245) 3 394 3.51 1.02 0.809 105 612 394 (254) 4 396 4.73 1.25 0.787 100 710 435 (266) 5 454 6.97 1.47 0.754 89.5 863 490 (278) B1T8 1 1021 5.91 1.96 0.843 323 1370 960-2 1065 10.4 3.58 0.842 367 1570 1090-3 1031 13.4 4.74 0.838 366 1630 1120 - B1T9 1 229 1.38 0.51 1 - - - - 2 784 2.16 0.59 1 - - - - 3 1554 3.10 0.87 1 - - - - 4 3014 6.15 1.54 1 - - - - 5 4047 7.77-2.75 1 - - - - B1T10 1 5219 7.66 3.42 1 - - - - 2 7239 11.7 5.39 1 - - - - 3 9306 15.9 8.01 1 - - - - RESULTS.doc Print date: 22-03-11 Volume 1 Section 4 Page 4 of 22 in this section

Weak Rock Self-boring Pressuremeter Tests Cambridge Insitu Ltd Typical tests in a borehole March 2011 Notes on Table 4 :- 1. The tables give a comprehensive breakdown of the shear modulus data for this contract. The linear elastic columns give the slope of the chord bisecting the loop. Also quoted is the pressure at the start of the loop, and the pressure change. 2. The non-linear elastic interpretation uses a power law to approximate the degradation of stiffness with strain. The multiplier and exponent of a power law are obtained from the intercept and slope of reloading data plotted in log-log space. 3. β The slope of a log-log plot of reloading data, where 1 is linear elasticity. 4. α The shear stress constant. 5. γ The plane shear strain, being approximated by 2ε c 6. γ=10-4 This column and the next two quote shear modulus at three magnitudes of shear strain. The equation is G s = αγ β-1, where G s is secant shear modulus. Note for the sake of completeness that tangential shear modulus G t is given by αβγ β-1 7. Curve matching The moduli from the curve matching procedure. The final column is G MIN the modulus at failure, with the strain to failure. Only applicable to tests in clay. Note that for the stiffer materials the value for modulus at 1% shear strain is either not included, or is in brackets. This is because the material will have failed before it reaches this strain as can be seen on the plots of modulus against strain derived from the reload loops. Note also that in the stiff rock there is no non-linearity. An increase of stiffness with successive loops is a sure sign of frictional (drained) behaviour; note however that the loops in B1T10 follow the main curve almost exactly, and the apparent increase in modulus is due to the strips of the Chinese Lantern forming to the shape of the borehole wall. Sometimes a plot of modulus against the loop starting pressure can be useful. Where there is no degradation of the borehole wall due to stress relief, it is possible to obtain an estimate of the Insitu Stress. The problem for non-linear modulus is choosing a strain level. RESULTS.doc Print date: 22-03-11 Volume 1 Section 4 Page 5 of 22 in this section

Weak Rock Self-boring Pressuremeter Tests Cambridge Insitu Ltd Typical tests in a borehole March 2011 Table 5 Test Frictional parameters Depth (bgl) Material Friction angle (constant volume) Friction angle (Peak) Angle of dilation (m) φ cv φ Ψ B1T2 6.1 Clay/gravel 24 24 0 B1T3 8.0 35.5 13.1 B1T4 10.0 Coarse sand 30 44.0 17.4 B1T7 37.5 Silty sand 32 47.3 19.6 B1T8 50.8 Chalk 25 37.2 14.2 B1T9 67.6 Limestone 43 45 2.8 Notes on Table 5: 1. The method of Hughes et al has been used to derive angles of friction and dilation. This requires a value for φ cv - the angle of friction at constant volume. The values used have been obtained from other, sometimes very simple, test methods or from past experience. Note that other analysis methods are available, and may sometimes be quoted. They are still somewhat unproven, however, and are not normally used. These analyses are Manassero (19**) which uses the same basic assumptions as the Hughes method, and a curve matching technique based on the equations derived by Carter et al (19**). Manassero uses a point by point approach to derive the radial strain and from this the volumetric strain and shear strain. It is usually used to produce (very noisy) plots of various parameters against shear strain, but can also give a plot of Shear Stress against Normal Stress from which values for friction angle and Insitu Stress may be obtained. This plot is particularly useful when the unloading is included, but it is this extension of the original concept that is unproven. The problem with the curve matching technique is that there are too many variables, and a good match may be produced by more than one combination of parameters. The problem, therefore, is to obtain independent values by other methods, but this is not always successful. The values of friction and dilation in B1T3 reduce considerably if a larger zero offset is used. Both curve matching methods would suggest that this is the case. 2. Other values of φ cv may be tried directly in the Hughes method, using the equations:- sin φ = S/[1 + (S -1) sin φ cv ] sin Ψ = S + (S -1) sin φ cv where φ is the peak angle of friction φ cv is the constant volume angle of friction Ψ is the angle of dilation. S is the slope of the Hughes plot. What follows now are plots of various parameters against depth, and modulus variation with strain. RESULTS.doc Print date: 22-03-11 Volume 1 Section 4 Page 6 of 22 in this section

Weak Rock Self-boring Pressuremeter Tests Cambridge Insitu Ltd Typical tests in a borehole March 2011 4.6 PLOTS OF VARIOUS PARAMETERS AGAINST DEPTH 4.6.1 Pore water pressures and lift-off As measured by the pore pressure cells before each test, and when each arm starts to move. Also the pressure at which the membrane collapses back in after the test, where applicable. Also shown are a possible ground water profile, and the pressure from a borehole full to the top. The ground water profile is based on an ambient level of 11m below ground level, and is assumed to be hydrostatic. 4.6.2 Total Insitu Stresses The average lift-off, and values from Marsland & Randolph, and curve matching. Also shown is the M & R value corrected for the excess pore pressure at the start of the test, assuming it has been measured correctly. Vertical stress assumes a relative density of 2 throughout. 4.6.3 Effective Stresses Based on the previous graph, assuming the ground water profile shown in 4.6.1. The corrected Marsland & Randolph values have not been included. 4.6.4 Effective Stress Ratios Using the values from the previous graph. A possible trend line may sometimes be added. 4.6.5 Shear Strength From the loading and the unloading, by the methods of Gibson & Anderson and Jefferies respectively. 4.6.6 Failure pressure, Limit Pressure and Test maximum pressure From the loading and curve matching. Also includes the maximum pressure reached during the test. It is not immediately obvious here, but the Limit Pressure from curve matching is usually about 1.3 to 1.4 times the maximum pressure. 4.6.7 Angles of Friction and Dilation The values shown are those derived by the method of Hughes et al. They vary widely in such varied material. 4.6.8 Initial Modulus and G MIN G MIN is the modulus at failure, derived using values from the curve matching. 4.6.9 Initial Modulus and G MIN (expanded scales) The same plot as 4.6.8, but on axes of 0 40m and 0 140MPa. SUBINDEX.doc Print date: 22-03-11 Volume 1 Section 4 Page 7 of 22 in this section

Weak Rock Self-boring Pressuremeter Tests Cambridge Insitu Ltd Typical tests in a borehole March 2011 4.6 PLOTS OF VARIOUS PARAMETERS AGAINST DEPTH (continued) 4.6.10 Moduli at different shear strains Tests B1T1 to B1T7 Secant Shear Moduli at strains of 1%, 0.1% and 0.01% (Includes the results from curve matching, where applicable). The same tests as shown on 4.6.9 by restricting the scales. 4.6.11 Modulus ratios G MIN divided by Po, Cu and σv, again making assumptions about the ground water profile for the effective stresses. Only applicable to those tests where G MIN may be determined. SUBINDEX.doc Print date: 22-03-11 Volume 1 Section 4 Page 8 of 22 in this section

Typical pressuremeter tests Ruritania April/May 2006 4.6.1 Pore water pressures and liftoff 0 10 20 30 H/static Depth (m) 40 50 60 PPC A PPC B Unloading Full borehole Arm 1 liftoff Arm 2 liftoff Arm 3 liftoff 70 80 0 200 400 600 800 1000 1200 1400 1600 1800 Pressure (kpa) CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 9 of 22 in this section 28/03/2011

Typical pressuremeter tests Ruritania April/May 2006 4.6.2 Total Insitu stresses 0 5 10 15 20 25 30 Liftoff M & R curve M & R corr. hydrostatic Vertical stress 35 Depth (m) 40 45 50 55 60 65 70 75 80 85 0 1000 2000 3000 4000 5000 6000 Pressure (kpa) CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 10 of 22 in this section 28/03/2011

Typical pressuremeter tests Ruritania April/May 2006 4.6.3 Effective Insitu stresses 0 5 10 15 M & R 20 Curve 25 30 Vertical stress 35 Depth (m) 40 45 50 55 60 65 70 75 80 85 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Pressure (kpa) CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 11 of 22 in this section 28/03/2011

Typical pressuremeter tests Ruritania April/May 2006 4.6.4 Effective Stress Ratios 0 5 10 15 M & R 20 curve 25 30 35 Depth (m) 40 45 50 55 60 65 70 75 80 85 0 1 2 3 4 5 6 7 Ratio (ko) CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 12 of 22 in this section 28/03/2011

Typical pressuremeter tests Ruritania April/May 2006 4.6.5 Shear strength (from loading and unloading) 0 10 G & A Jefferies 20 30 Depth (m) 40 50 60 70 80 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Cu (kpa) CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 13 of 22 in this section 28/03/2011

Typical pressuremeter tests Ruritania April/May 2006 0 4.6.6 Failure pressure, Limit pressure and Test maximum pressure 10 Pf PL(G & A) PL(curve) Pmax 20 30 Depth (m) 40 50 60 70 80 0 5000 10000 15000 20000 25000 30000 35000 40000 kpa CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 14 of 22 in this section 28/03/2011

Typical pressuremeter tests Ruritania April/May 2006 4.6.7 Angles of Friction and Dilation 0 5 Friction Dilation 10 15 20 25 30 35 Depth (m) 40 45 50 55 60 65 70 75 80 85 0 5 10 15 20 25 30 35 40 45 50 Angle (deg) CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 15 of 22 in this section 28/03/2011

Typical pressuremeter tests Ruritania April/May 2006 4.6.8 Initial Modulus and G MIN 0 5 Gi - Initial modulus GMIN 10 15 20 25 30 35 Depth (m) 40 45 50 55 60 65 70 75 80 85 0 2000 4000 6000 8000 10000 12000 Modulus (MPa) CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 16 of 22 in this section 28/03/2011

Typical pressuremeter tests Ruritania April/May 2006 4.6.9 Initial Modulus and G MIN (expanded scales) 0 Gi - Initial modulus GMIN 5 10 15 Depth (m) 20 25 30 35 40 0 20 40 60 80 100 120 140 Modulus (MPa) CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 17 of 22 in this section 28/03/2011

Typical pressuremeter tests Ruritania April/May 2006 4.6.10 Secant Moduli (at three shear strains) Tests B1T1 to B1T7 0 0.01% strain 5 0.1% strain 1% strain 10 15 Depth (m) 20 25 30 35 40 0 100 200 300 400 500 600 700 800 900 1000 Modulus (MPa) CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 18 of 22 in this section 28/03/2011

Typical pressuremeter tests Ruritania April/May 2006 4.6.11 Modulus ratios 0 Gmin/Po' Gmin/Cu Gmin/σv' 5 10 15 Depth (m) 20 25 30 35 0 50 100 150 200 250 300 350 400 CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 19 of 22 in this section 28/03/2011

Weak Rock Self-boring Pressuremeter and High Pressure Dilatometer Tests Cambridge Insitu Ltd March 2011 4.7 PLOTS OF THE VARIATION OF MODULUS WITH STRAIN 4.7.1 Secant shear modulus vs Shear strain B1T1 to B1T7 These curves come straight from the analysis of the reload loops. They are the theoretical representation of the values obtained for the plots that show the actual data points as well see the individual tests. The range of shear strain chosen is that over which the curve can be confidently predicted. It may be seen from the individual test plots that the actual data points deviate from the theoretical curve at about 1% strain (less in stiffer materials). 4.7.2 Normalised Youngs modulus vs Axial strain B1T1 to B1T7 Both the modulus and strain have been transformed, to give a plot that may be compared to that from a triaxial test. A value must be assumed for Poisson s ratio, 0.3 has been used here. Note that many other plots are possible, and are sometimes asked for. Other normalising factors may be used, cavity strain may be used, and the strain scale may be linear. MODINDEX.doc Print date: 29-03-11 Volume 1 Section 4 Page 20 of 22 in this section

Typical pressuremeter tests Ruritania April/May 2006 1000 4.7.1 Secant shear modulus vs Shear strain - B1T1 to B1T7 900 Note: The last loop on the loading in each test is plotted Secant shear modulus, Gs (MPa) 800 700 600 500 400 300 200 B1T1 B1T2 B1T3 B1T4 B1T5 B1T6 B1T7 100 0 0.0001 0.001 0.01 log shear strain CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 21 of 22 in this section 29/03/2011

Typical pressuremeter tests Ruritania April/May 2006 2000 4.7.2 Normalised Youngs modulus vs Axial strain - B1T1 to B1T7 1800 Note: The last loop on the loading in each test is plotted Normalised Youngs modulus, E/Cu 1600 1400 1200 1000 800 600 400 B1T1 B1T2 B1T3 B1T4 B1T5 B1T6 B1T7 200 0 0.0001 0.001 0.01 log axial strain CAMBRIDGE INSITU Little Eversden Cambridge CB3 7HE England Volume 1 Section 4 Page 22 of 22 in this section 29/03/2011