Correlations between PENCEL Pressuremeter, Cone Penetrometer and Dilatometer Parameters

Transcription

1 Correlations between PENCEL Pressuremeter, Cone Penetrometer and Dilatometer Parameters November 13, 2007 Paul J. Cosentino, Ph.D., P.E., Florida Institute of Technology Department of Civil Engineering (321) Edward H. Kalajian, Ph.D., P.E., Florida Institute of Technology Farid Messaoud, Tebessa University, University Center CLT, Tebessa 12000, Algeria Sunil Sundaram, ECS Florida LLC, Orlando FL Thaddeus J. Misilo, Graduate Research Assistant, Florida Institute of Technology David Horhota, Ph.D., P.E. Florida Department of Transportation, Gainesville FL 1

2 Acknowledgement This research was completed with the Florida Department of Transportation under contract BD 658. The authors would like to acknowledge the following people for their invaluable guidance and help in the completion of this study; John Shoucair, Brian Bixler and Todd Britton of the Florida Department of Transportation State Materials Office. 2

3 Abstract PENCEL Pressuremeter, Cone Penetrometer and Dilatometer tests were performed at three sites in Florida. Two of the sites were predominately sands and the third was predominately clay. The study goal was to demonstrate that pushed-in PENCEL tests would produce reliable engineering parameters. The PENCEL was pushed to the desired test depth using the Cone Penetrometer equipment. Both a smooth cone tip and a cone tip equipped with a small friction reducer were used and evaluated during pressuremeter testing. A standardized testing procedure was followed for all tests. Initial or lift-off pressures, elastic moduli and limit pressures were determined from the PENCEL pressuremeter testing, while lift-off and elastic moduli were determined from dilatometer testing. Cone testing produced both friction and tip resistances. Both manual and digital pressures and volumes were recorded during pressuremeter testing. Correlations were developed within the engineering parameters obtained from PENCEL pressuremeter data and between the PENCEL, Cone and Dilatometer engineering parameters. All correlations matched published values. From the PENCEL data, excellent correlations were produced between the initial elastic modulus and the limit pressure as well as the initial elastic modulus and the elastic reload modulus. Excellent correlations were also produced between the initial elastic modulus and limit pressures. Correlations based on digital elastic and reload moduli reduced with a software package called APMT were better than those based on a combination of digital and manually recorded data. From the comparisons between pressuremeter, cone and dilatometer data, promising correlations were developed between the pressuremeter initial elastic moduli and Cone Penetrometer tip resistances. Promising correlations were also developed between pressuremeter limit pressure and Cone Penetrometer tip resistances. Consistent ratios existed between PENCEL Pressuremeter and Dilatometer lift-off pressures as well as Pressuremeter and Dilatometer initial elastic moduli. Evaluation of smooth and friction reducer cone tips indicate that the increased soil disturbance, associated with the friction reducer being pushed through the soil, decreases the engineering parameters and the friction reducer is not recommended. 3

4 1 INTRODUCTION 1.1 Brief History of the Pressuremeter The pressuremeter (PMT), a balloon type device that is inflated once inserted into the soil to produce in situ stress-strain responses, was originally developed in the 1950s by Ménard (1956) and modified by Briaud and Shields (1979) in the 1970s. Various PMT models are currently available, although they are typically based on two widths, the standard 3-inch diameter probes lowered into boreholes and the specialty 1.35-inch diameter PENCEL probes pushed when attached to cone rods (Briaud, 1992). In addition to classical geotechnical applications, Briaud and Cosentino (1989) developed procedures for using the PENCEL pressuremeter (PPMT) in pavement design. The PPMT is shown in Figure 1 with the probe connected to the unit through tubing and the pressure and volume gauges for recording data by hand (Roctest, 2005). Anderson and Townsend (1999) saw advantages in connecting the PPMT probe to Cone Penetrometer rods and either pushing the cone with the PPMT attached or pushing the PPMT separately to conduct PPMT tests. Finally, this device was further advanced by 1) developing a standardized testing procedure as recommended by Cosentino et al (2006) and 2) incorporating digital technology with data acquisition software producing significant time savings and improved accuracy as a fully reduced stress-strain curve is produced during testing (Cosentino et al, 2006), (Cosentino et al, 2007). The soil stratigraphy limits the depth to which PPMT tests can be performed similar to Cone Penetrometer testing. Often thrust pressures monitored by the equipment operators are limited to 10 kn to avoid damage. PPMT equipment has been successfully used throughout Florida in sands and clays (Anderson and Townsend 1999) (Cosentino et al, 2006). 4

5 Tubing Pressure Gage Volume Counter Probe FIGURE 1 PENCEL Pressuremeter Probe, Tubing and Control Unit on Tripod Stand (after Roctest 2005). 1.2 Typical Operation Following the system saturation, several calibrations are performed, one that accounts for the inherent membrane resistance, termed the membrane calibration and a second, called the system expansion calibration, that accounts for expansion of the tubing and thinning of the membrane during pressurization. A hydraulic correction is also applied to the pressures because the test is conducted at a known distance below the pressure gauge. The PPMT probe is then hydraulically pushed with the equipment in the CPT rig to the desired test depth and the 10- to 15-minute standardized test recommended by Cosentino et al (2006) is performed. During this standard test, a strain-controlled process is used, requiring operators to turn the crank handle to inject equal 5 cm 3 volumes of water into the probe, wait 30 seconds and then record the corresponding pressures. The probe volume is incrementally increased from its original volume (V o ) an additional 90 cm 3, or until the limit of the pressure gauge is 5

6 reached. The operators also determine the extent of the linear stress strain response range before conducting one unload reload cycle on the soil. This determination requires several complex steps; therefore, Cosentino et al (2006) incorporated digital equipment and data acquisition software, called APMT for Automated Pressuremeter that simplified the process, yielding more precise data while easing operator requirements. 1.3 Data Interpretation Once the data is collected it is typically plotted on a curve as shown in Figure 2. This figure contains both the volume (i.e., system expansion) calibration line and the membrane calibration curve, which are subtracted from the raw data to produce a reduced curve V o l u m e C a l i b r a t i o n R a w D a t a Pressure (kpa) R e d u c e d D a t a M e m b r a n e c a l i b r a t i o n Volume (cm 3 ) FIGURE 2 Raw and Reduced Field data with calibrations applied (Cosentino et al 2006). In order to determine the critical engineering parameters, various portions of the reduced curve are analyzed. Figure 3 shows four critical portions of the curve that are used for 1. estimating the at-rest horizontal pressure or initial pressure (p 0 ), 2. estimating an initial elastic modulus (E 0 ), from the elastic phase, 3. estimating an elastic reload modulus (E r ), from the reload phase and 6

7 4. estimating the limit pressure (p L ). Limit Pressure Pressure (kpa) p o Elastic Phase S i At-Rest Soil Pressure FIGURE 3 Engineering parameters typically obtained from reduced data. S r Volume (cm 3 ) Elastic Reload Phase Plastic Phase Elastic moduli are determined from the following equation (Baguelin et al, 1978): p L Unloading E = 2( 1+ ν) P V V m where, E = Young s modulus P = change in stress V = change in volume related to P V m = average volume over P ν = Poisson's Ratio There are concerns about the quality of the engineering parameters obtained from pushed-in PPMT tests due to soil disturbance. To help preserve the membranes, some 7

8 operators push the probe using a small friction reducer on the cone tip, while others push it without this device. 2 OBJECTIVE The objective of this paper is to show reliable correlations exist from the pushed PPMT data and to evaluate whether there is a need for a friction reducer on the cone tip of the PENCEL probe. 3 APPROACH To meet these objectives three major tasks were completed: 1. Field Testing 2. Data Reduction and Analysis 3. Development of Correlations 3.1 Field Testing A comprehensive field-testing program was performed in Florida, enabling both sands and clays to be evaluated. In addition to PPMT tests, Cone Penetrometer (CPT) and dilatometer (DMT) tests were conducted. Tests in normally consolidated sands were performed on the Florida Institute of Technology (FIT) campus and within the Archer Landfill, near Gainesville, FL. Tests in clays were conducted in Cape Canaveral, FL. All testing was conducted using the Florida Department of Transportation (FDOT) State Materials Office CPT rig and personnel. Over 200 PPMT and DMT tests were conducted at the three sites. To determine the effects the friction reducer, shown schematically in Figure 4, has on the soil properties, about half of the PPMT tests were conducted with and half without the friction reducer. The 33.9 mm (1.335 inch) diameter reducing ring was about 3 % larger than the 33.0 mm (1.280 inch) diameter smooth cone point. 8

9 Friction Reducer Cone Tip Smooth Cone Tip inch 33.9 mm inch 33.0 mm FIGURE 4 Schematic of PPMT cone tip with and without the friction reducer (Cosentino et al 2006). A review of the land use history revealed that the FIT site soil stratigraphy is composed of three normally consolidated sand layers. The upper 3.05 m (10 feet) mediumdense layer is a fill composed of sands interbedded with silt and clay lenses. The density averaged 17 kn/m 3 (108.3 pcf) with a friction angle varying from 32 to 35 degrees. The second layer, considered to be the original ground surface, is also 3.05 m (10 feet) and consists of very loose to loose silty sand. The density averaged 16 kn/m 3 (101.2 pcf) with a friction angle varying from 27 to 32 degrees. The third layer, starting at the 6.1 m (20 foot) level consists of dense, cemented sands. The density averaged 21 kn/m 3 (133.8 pcf) with a friction angle varying from 42 to 44 degrees. The sands at the Archer Landfill site were above the water table and were divided into three layers. As was the case with the FIT site, the stress history within the landfill is not well documented and based on DMT test data correlations could vary between normally to slightly overconsolidated sands. The upper 2.1 m (7 feet) consisted of loose sandy silt with Standard Penetration Testing N-values between 5 and 12 and angles of internal friction of 34 degrees (Townsend et al, 2001), while from 2.1 to 4.2 m (7 to 14 feet), medium dense sand with N-values from 12 to 20 and angles of internal friction of 33 degree (Townsend et al, 9

10 2001) exists. The third layer, from 4.2 to 9.1 m (14 to 30 feet) was medium dense sand to silty sand with N-values between 22 and 28 and angles of internal friction of 36 degree (Townsend et al, 2001). The Cape Canaveral site consists of interbedded sands and clays. There were two clay layers that were the focus of the research. An upper clay layer approximately 2m (6 feet) thick was normally consolidated and had an average density of 14.4 kn/m 3 (92 pcf) and a lower normally consolidated layer from the 10 to 15 m (30 to 50 feet) depth with an average density of 15.3 kn/m 3 (97 pcf). The procedure used during PPMT testing was the recommended FDOT standard (Cosentino et al, 2006). During the strain-controlled test, operators monitored stress versus volume data to determine the extent of the elastic range (Figure 3). Once this range was complete, unloading to one-half the existing pressure then reloading to the original pressure was performed followed by the remainder of the strain-controlled test (Figure 3). A detailed description of the PPMT testing procedure is given in Cosentino et al (2006). The American Society for Testing and Materials (ASTM) procedure D 6635 was followed for all DMT testing, while CPT tests were conducted in accordance with ASTM D For evaluating DMT data, Marchetti (1980) presented equations requiring several preliminary calculations to determine a Young s modulus of elasticity (E). For this research the following steps are to be followed. After obtaining the two basic test parameters; the liftoff pressure (A) or the pressure on the DMT membrane once it is pushed to the desired depth and the maximum pressure at 1.1 mm of movement (B), a corrected contact stress is found using the equation: p o = 1.05 (A- Z M + A) (B-Z M - B) where Z M is the gauge pressure when vented to the atmosphere, while A and B are calibration pressures subtracted from the lift-off and maximum readings. A corrected expansion stress is then found using the equation: p 1 = B-Z M - B The DMT modulus, not Young s Modulus, is then found from the equation: E D = 34.7(p 1 -p o ) 10

11 This DMT modulus can be converted to a Young s Elastic Modulus by first determining a constrained modulus from: M DMT = R M E D where R M is an empirical value that is a function of either the material index (I D ) defined as (p 1 p o )/(p o - u o ) or the horizontal stress index (I D ) defined as (p 0 u o )/(σ' vo ). Note that u o is the pore water pressure and σ' vo is the vertical effective stress. The constrained modulus is used in the following equation, based on Poisson s ratio (ν) to determine the elastic modulus: E = M DMT [(1+ν)(1-2ν)/(1-ν)] 3.2 Data Reduction and Analysis In conjunction with incorporating digital pressure and volume equipment into the PENCEL control unit, the stand-alone data acquisition (DAQ) program, called APMT, was developed for this research (Cosentino et al, 2007). A detailed description of APMT; including installation, system criteria and usage can be found in Cosentino et al (2006). This software uses the digital calibration data to continuously reduce the digital field data producing a stress-strain plot on the operators computer screen throughout testing. This plot allows operators to follow standardized testing procedures. APMT also has built-in modules that yield the critical stress-deformation information i.e., p0, E 0, E r and p L. Results obtained using the APMT package were compared to hand and spreadsheet calculations. Once the output was verified this package was used to determine the four key engineering parameters obtained during PPMT testing (i.e., p 0, E 0, E r, p L ). Because the digital instrumentation and associated software was developed during this work, some of the data was recorded manually and some was recorded digitally. All testing conducted at the Archer site was performed using APMT software integrated with updated digital technology (Cosentino et al, 2007). 3.3 Development of Correlations Friction Reducer versus smooth cone driving tips To evaluate the effects of using a friction reducer, 80 tests were conducted at the Cape Canaveral Site, 40 were performed with a friction reducer and 40 without. The tests were performed in 16 soundings (i.e., a series of tests performed in one location while 11

12 advancing the PENCEL probe to the desired depths), with eight being conducted with, and eight without the friction reducer. Tests performed without the friction reducer were termed smooth cone tests. A comparison was developed between the smooth cone data and the friction cone data using the initial and reload moduli, plus the initial or lift-off pressures and the limit pressures. Ratios of these four parameters at five depths are shown in Table 1. This data was erratic in the two upper depths due to inconsistencies in the soil types as the PENCEL probe was moved between soundings. However, once the soft clay was encountered the ratios between the smooth and friction reducer probes were nearly 1.00, indicating that in soft clay there is very little difference between the results conducted with and without the friction reducer, maybe because the soft clay moved around the friction reducer. Of the four parameters evaluated, the initial modulus was most affected by the use of a friction reducer. To further evaluate the need of a friction reducer cone tip, the ratios between the PPMT initial moduli and the CPT point resistances (q c ) were evaluated along with ratios of the PPMT limit pressures and q c. The E/q c ratios are commonly used for settlements of sands (Schmertmann et al, 1978). These comparisons along with published values are shown in Table 2. Again, for the first two depths the comparisons are not consistent, however, for the last three depths the values indicate that there is very little difference between the results from tests conducted with and without the friction reducer. The correlations in Table 2 also indicate that reliable engineering parameters can be obtained from PPMT testing. Data in both tables show that parameters obtained with the smooth cone are slightly higher than that from the cone with the friction reducer. This difference indicates that the additional soil disturbance associated with the friction reducer reduces the engineering parameters. 12

13 TABLE 1 Ratio between PPMT Engineering Parameters for Smooth and Friction Reducer Cone Tip Testing at Cape Canaveral Site (Cosentino et al, 2006). Depth Engineering Parameters Soil E o (smooth) E r (smooth) p L (smooth) (m) Type E o (friction) E r (friction) p L (friction) 2.5 Soft Sandy Clay Loose Silty Fine Sand Soft Clay Soft Clay Soft Clay p o (smooth) p o (friction) TABLE 2 Comparison of PPMT and CPT Engineering Parameters between the Smooth and Friction Reducer Cone Tips at the Cape Canaveral Site (Cosentino et al, 2006). Depth E o /p L E o /q c q c /p L PPMT CPT PPMT CPT PPMT CPT (m) Smooth Friction Ref A Smooth Friction Ref s B, C Smooth Friction Ref B to to to to Reference A Menard and Rousseau (1962) Reference B Schmertmann (1978) Reference C Bergado and Al Khaleque (1986) Correlations from PPMT Engineering Parameters The initial elastic modulus was compared to the limit pressures using the engineering parameters from 96 PPMT tests in the silty sands at the FIT and Archer sites. These soils ranged from very loose to dense silty sands. Figure 5 shows an excellent correlation exists when modeled nonlinearly. A nonlinear relationship would be expected because the limit pressure cannot increase infinitely as stiffness increases. Briaud (1992) presented linear correlations based on over 400 records, between the limit pressure and initial elastic modulus of p L = E o for sands and p L = E o for clays. He specifically states that the wide scatter in the data used to develop them, makes these correlations essentially useless for design; however, they give the engineers a relative feel for the engineering parameters. When a linear regression through the origin was used to describe this data the equation becomes p L = E o. In conclusion, this nonlinear relationship shows that PPMT data are 13

14 realistic and can be used by engineers. Data from sixty of the 96 tests was collected manually, while data from the remaining 36 tests was recorded digitally and reduced with the APMT software. Of those digitally recorded, 30 were from driven PPMT tests and the other ten were conducted in prebored holes, in which the prebored holes were constructed by driving an open ended pipe the same diameter as the PENCEL probe p L = 0.43E o R 2 = Limit Pressure (kpa) ,000 10,000 15,000 20,000 25,000 30,000 35,000 Initial Modulus (kpa) FIGURE 5 Correlation between limit pressure and initial elastic modulus in silty sands. Using data from 36 tests performed in silty sands from both the FIT and Archer sites, a nonlinear correlation was developed between the initial elastic modulus and the reload modulus (Figure 6). Twenty of the 36 tests were performed and recorded using the digital system. When only digital information was evaluated the regression equation became E r = 0.15E o and had a corresponding regression coefficient (i.e. R 2 ) of This improved correlation suggests that the digital instrumentation combined with the APMT software improves the PPMT data. Briaud (1992) presents a linear correlation between these two parameters where E r = 8 E o in sands. If the data in Figure 6 is represented linearly a 14

15 regression of E r = 16 E o results with an R 2 value of 0.76 and if only digital data is used the equation becomes E r = 21 E o results with an R 2 of These linear results indicate that digital testing should be used to improve pressuremeter data. 600, ,000 E r = 0.37E o R 2 = 0.86 Rebound Modulus (kpa) 400, , , , ,000 10,000 15,000 20,000 25,000 Initial Modulus (kpa) FIGURE 6 Correlation between initial modulus and reload modulus for silty sands. Similar correlations between the lift-off or initial pressure and limit pressure were developed; however the data were not well represented statistically as shown in the equation and corresponding correlation coefficient below. The poorly documented stress history in the sands at both FIT and Archer may have a significant impact on the lift-off pressures. This fact alone could produce a poor correlation. However, common empirical correlations exist between the at-rest earth pressure and the friction angle and it is believed that a similar correlation may exist between the lift-off and limit pressure. This relatively weak correlation could also indicate that the initial pressures may not be properly determined either graphically or procedurally during the beginning of the PPMT tests. Currently, equal 5 cm 3 volume increments are injected and the resulting pressure volume curve is evaluated, smaller volume increments may be necessary to better define the lift-off pressure. 15

16 p L = 24.2 (P o ) R 2 = 0.62 Linear correlations between the lift-off pressure and initial elastic moduli were developed from the testing in sands. The overall regression coefficient was 0.30, therefore the resulting equation is not presented. Again, limited knowledge of the stress history at these sites could be a major factor in the poor statistical fit. The data recorded by hand was compared to the digitally recorded data. The regression coefficient from the digital information was nearly 0.58 producing the equation below; while the data recorded manually was This 60 % increase in the regression coefficient indicates that digital stress-strain data is a significant improvement over manually recorded data. E o =68 (P o ) Correlations to Engineering Parameters from Other Instruments Using the instrumented control unit along with the APMT software the initial elastic modulus and limit pressures from PPMT tests conducted at the Archer site were compared to CPT data from this site to develop correlations. These correlations were then compared to those developed by Schmertmann (1978) and these comparisons were summarized in Table 3. The consistent results indicate that the instrumentation and associated software produce very reliable engineering parameters. TABLE 3 Comparison of PPMT and CPT Engineering Parameters at the Archer Site. Test Depth (m) Schmertmann s (1978) Correlation q c / p L Automated Test Correlation q c / p L Schmertmann s (1978) Correlation E o /q c Automated Test Correlation E o / qc

17 PPMT data was also compared to DMT data from the Cape Canaveral site. DMT liftoff pressures were compared to PPMT lift-off or initial pressures plus the DMT and PPMT initial moduli were compared. Correlations between these parameters were not conclusive; however, ratios between the DMT and PPMT parameters were developed to provide engineers with a probable range. The ratio of the DMT/PPMT lift-off pressures varied from 1.2 to 2.7, while the DMT/PPMT elastic moduli ratios varied from 0.9 to 1.4. These ranges were based on data from 80 PPMT tests and 20 DMT tests at 5 depths, 40 of the PPMT tests were conducted with the smooth cone and 40 with the friction reducer cone. 4 CONCLUSIONS AND RECOMMENATIONS Reliable correlations exist from the large number of pushed PPMT test data, indicating that this device can produce useful parameters for engineers to use in design and analysis. The data from this research indicates there is no need for a friction reducer on the cone tip of the PENCEL probe. The increased soil disturbance that may be associated with the friction reducer decreases the engineering parameters (i.e. p o, E o, E r, p L ). PPMT data produces more engineering parameters (i.e., p 0, E 0, E r, p L ) than either DMT or CPT data. A precise nonlinear correlation was developed between the PPMT initial elastic modulus and the limit pressures in sands. A reliable nonlinear correlation was developed between the PPMT initial elastic and the reload moduli in sands. This correlation improved when digital information along with the APMT software was used, indicating that the combination of digital instruments and this software will provide engineers with more reliable data than the analogue equipment alone. When digital pressures and volumes were recorded and reduced with the APMT software, a useful linear correlation was developed between the PPMT lift-off pressure and the initial elastic modulus, further substantiating the finding that instrumentation and proper software reduction improves the PPMT. 17

18 Several correlations between PPMT data and CPT data were confirmed and shown to be very consistent. Probable ratios between PPMT and DMT parameters were presented and should be improved with further research. The pushed-in PPMT test is much faster than conventional pressuremeter testing and is recommended for use in determining the soils stress-strain response and the associated engineering parameters. 5 REFERENCES Anderson, J.B., and Townsend, F.C., Validation of P y Curves from Pressuremeter Tests at Pascagoula Mississippi. Proc. 11th Panamerican Conference on Soil Mechanics and Geotechnical Engineering. Baguelin, F., Jézéquel, J.F., and Shields, D.H., The Pressuremeter and Foundation Engineering. 1st ed., trans., Tech Publications, Causthal, Germany. Beergardo, D.T., and Abul Khaleque, M., Correlation of LLT Pressuremeter, Vane, and Dutch Cone Tests in Bangkok Marine Clay, Thailand. 2nd International Symposium on the Pressuremeter and its Marine Applications, ASTM, Briaud, J.L., The Pressuremeter. A.A Balkema, Brookfield, Vermont. Briaud, J.L., and Shields, D.H., A Special Pressuremeter and Pressuremeter Test for Pavement Evaluation and Design. Geotechnical Testing Journal, ASTM 2:3. Cosentino, P.J., and Briaud, J.L., FWD Back Calculation Moduli Compared with Pavement Pressuremeter Moduli and Cyclic Triaxial Moduli. Philadelphia, Pennsylvania: ASTM Cosentino, P., Kalajian, E., Stansifer, R., Anderson, J. B., Kattamuri, K., Sundaram, S., Messaoud, F., Misilo, T., and Cottingham, M., Standardizing the Pressuremeter Test for Determining p-y Curves for Laterally Loaded Piles, FDOT Research Report. Contract BD 658. Cosentino, P., Kalajian, E., Sundaram, S., and Misilo, T., Instrumenting the PENCEL Pressuremeter Control Unit to Simplify Data Collection, Reduction and Analysis TRR In Press. Marchetti, S., In Situ Tests by Flat Dilatometer. ASCE Journal GED 106(GT3): Menard, L., 1956, An Apparatus for Measuring the Strength of Soils in Place. Master s Thesis, University of Illinois. Menard and Rousseau L evaluation des Tassements Tendence Nouvelle, Sol- Soils 1: Roctest, Inc., PENCEL Pressuremeter Instruction Manual. Plattsburgh, N.Y.: Roctest, Inc. 18

19 Schmertmann, J.H., Guidelines for the Cone Penetration Test Performance and Design. Washington, D.C.: U.S. Department of Transportation, Federal Highway Administration Report FHWA-TS Schmertmann, J.H., Hartmann, J. P. and Brown P.R., Improved strain influence factor diagrams. Proceedings of the American Society of Civil Engineers, 104(GT8), Townsend, F.C., Anderson, J. B, and Rahelison, L., 2001, Evaluation of FEM engineering Parameters from Insitu Tests, Final Report BC , FDOT, p