U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
|This report is an archived publication and may contain dated technical, contact, and link information|
Publication Number: FHWA-RD-99-194
Date: June 2000
Development and Field Testing of Multiple Deployment Model Pile (MDMP)
7.1.1 MDMP Test NB2
The maximum (peak) pore pressure measured during MDMP test NB2 was observed following the final downward displacement associated with the alignment of the static load frame at 39.72 min after the start of installation (see Figure 47). This maximum pore pressure of 217.33 kPa (31.52 psi) was used as the initial pore pressure, denoted by ui. By subtracting the hydrostatic pore pressure of 58.12 kPa (8.43 psi), the initial excess pore pressure (ui) was 159.21 kPa (23.09 psi). For the analysis of normalized pore pressure, the time associated with the end of driving was considered to be that of the aforementioned adjustment since it was accompanied by a significant gain of pore pressure. From this assigned end of driving to the end of the test, the excess pore pressure record was normalized by the maximum initial pore pressure (). Figure 88 shows the normalized excess pore pressure versus time after driving. The consolidation ratio of the soil at the point of measurement was the difference between 1 and the normalized excess pore pressure ratio:
The normalized excess pore pressure continued below the zero line (i.e., beyond 100% consolidation) because the actual measured pore pressure at the end of the test was below the average hydrostatic water pressure at the site. These values could be corrected to ensure that the final pore pressure would not decrease below the hydrostatic level; however, it was chosen not to do so in order to reflect the actual accuracy of the field measurements. From 80% to 20% normalized excess pore pressure (20% to 80% consolidation), the slope of the best-fit line on the log-normal scale represented the rate of radial consolidation, Hut, and was 0.6047. The time at 50% dissipation, t50, was 9.854 h (35476 s). When adjusted to the PLS cell radius (19.177 mm), t50(pls) was 2.493 h (8975 s). The time adjustment (see section 5.6) allowed the comparison of the absolute time of any dissipation process to another, regardless of the pile size. As previous data analysis used the PLS cell as its standard (Paikowsky et al., 1995), the current measurements were adjusted to the same size as well.
7.1.2 MDMP Test NB3
The maximum (peak) pore pressure during model pile test NB3 was 224.02 kPa (32.49 psi). This maximum pore pressure was used as the initial pore pressure value (ui) in the normalization process. The maximum/initial excess pore pressure was, therefore, obtained by subtracting the hydrostatic pressure of 88.53 kPa (12.84 psi) from the initial excess pore pressure, resulting in a ui of 135.49 kPa (19.65 psi). Figure 89 shows the normalized excess pore pressure versus time after driving, from the end of driving to the end of the test. The normalized excess pore pressure at the end of the test was above zero as the final measured pore pressure was above the average hydrostatic pore pressure calculated for that depth. From 80% to 20% normalized excess pore pressure(equivalent to 20% to 80% consolidation), the slope of the best-fit line on the log-normal scale, Hut, was 0.6011. The time at 50% dissipation, t50, was 7.849 h (28256 s). When adjusted to the PLS cell radius (19.177 mm), t50(pls) was 1.986 h (7149 s).
7.1.3 Comparison With Other Test Results
Paikowsky et al. (1995) found, based on a database analysis, that the rate of pore pressure dissipation, Hut, was 0.466±0.089 (33 cases) for normally consolidated soil and 0.498±0.067 (12 cases) for normally consolidated Boston Blue Clay (BBC). The complete range of Hut values for normally consolidated soils was from 0.325 to 0.763, with a narrower range of 0.351 to 0.584 for the normally consolidated BBC. Table 30 summarizes the results of the excess pore pressure dissipation during the Newbury Testing as compared to the study presented by Paikowsky et al. (1995).
Table 30. Summary of Excess Pore Pressure Dissipation Parameters and Their Comparison to a Large Data Set.
The rate of pore pressure dissipation in both tests was practically identical, with an average Hut of 0.603. This value falls within the range of recorded values for all normally consolidated soils (from 0.325 to 0.763), but indicated a faster dissipation than the mean value of Hut=0.466±0.089 (33 cases). The measured rate was also about 20% higher than the mean value found for BBC at the Saugus, MA site (Hut=0.492±0.072).
The above measurements and ranges suggest that: (1) the use of an average Hut value from a large data set provided a reasonably good initial estimation of the dissipation rate, but a site-specific investigation was required for an accurate evaluation, and (2) the normally consolidated BBC contained a larger variation of soil composition and/or deposition sequence, which affected the radial consolidation process. Preliminary investigation of the soils at the Newbury site (conducted by John Chen at UMass-Lowell) suggested that the clay at the testing location contained a larger proportion of silt content than the typical BBC clay.
The time required to achieve 50% dissipation (t50) normalized to the PLS cell radius (19.177 mm) was 1.796±1.02 h (6466±3665 s) for all normally consolidated soils and 1.57±0.334 h (5655±1202 s) for normally consolidated BBC. The actual range of t50 values for 27 normally consolidated soils was found by Paikowsky et al. (1995) to be 0.655 h (2359 s) to 6.03 h (21722 s) and 1.08 h (3878 s) to 2.21 h (7958 s) for normally consolidated BBC. The average t50 for the MDMP tests at the Newbury site normalized to the PLS ell radius is 2.24 h (8062 s). This value fell within the range for all non-consolidated (NC) soils, but indicated a longer time before 50% dissipation was completed at the Newbury site, compared to the previously measured dissipation times in BBC.
The above measurements and ranges suggested that, as with the Hut parameter, using the average t50 from the large data set provided a reasonably good initial estimation of the time to 50% dissipation, but a site-specific investigation was required for an accurate evaluation. Based on the large variability of the t50 parameter, a range should be used when predicting the dissipation time.
Table 31 presents a typical time range for a 0.3048-m (1-ft) diameter pile to reach various amounts of dissipation. From these analyses, the combined effect of the two parameters, Hut and t50, could be examined. Even though the dissipation rate was faster at the Newbury site, the time to 80% dissipation was about the same as for the Saugus site and all NC soils based on the longer time required for 50% consolidation.
Table 31. Evaluated Pore Pressure Dissipation Time (Adjusted to the PLS Diameter) Based on the Newbury Test Results Compared With a Large Data Set.
7.2.1 MDMP Test NB2
The shear transfer measured along the friction sleeve during the static load tests was normalized to the maximum measured shear transfer and presented in Figure 88 along with the normalized pore pressure readings. Both peak and residual shear transfer values from each static load test were included in the normalized shear transfer relations. Utilizing the data between the load transfer ratios of 0.1 to 1.0, the rate of capacity gain (Cgt) for both peak values and residual values was 0.589. The time to 75% gain of capacity, t75, was 67.7 h for peak values and 70.7 h for residual values. Following the standard normalization used by Paikowsky et al. (1995), t75 was adjusted to a 152.4-mm (6-in) radius pile. The t75(152.4 mm) was 1083.2 h for peak values and 1131.2 h for residual values.
7.2.2 MDMP Test NB3
The shear transfer measured along the friction sleeve during the static load tests was normalized to the maximum measured shear transfer and presented in Figure 89 along with the normalized pore pressure readings. Both peak and residual shear transfer values from each static load test were included in the normalized shear transfer relations. Utilizing the data between the load transfer ratios of 0.1 to 1.0, the rate of capacity gain (Cgt) was 0.599 for peak values and 0.631 for residual values. The time to 75% gain of capacity, t75, was 43.2 h for peak values and 38.2 h for residual values. Following the standard normalization used by Paikowsky et al. (1995), the t75 was adjusted to a 152.4-mm- (6-in-) radius pile. The t75(152.4 mm) was 691.2 h for peak values and 610.9 h for residual values.
7.2.3 Comparison With Other Test Results
Based on the analysis of data set compiled at UMass-Lowell, Paikowsky et al. (1995) found that the expected Cgt from 39 cases was 0.367, with a standard deviation of 0.096. The time to 75% gain in capacity was 370.7±338.7 h when normalized to a 152.4-mm (6-in) pile radius. Table 32 summarizes the gain in capacity behavior based on the Newbury MDMP test results, along with the data presented by Paikowsky et al. (1995) for other locations.
Table 32. Summary of Gain of Capacity Parameters and Their Comparison to a Large Data Set.
The average Cgt for the Newbury site was 0.594 for peak values and 0.610 for residual values. The Cgt values were consistent for the two MDMP tests at the Newbury site and were much higher than those determined by Paikowsky et al. (1995) based on other test results. This difference may be explained by one or more of the following reasons:
(1) The multiple load tests in each of the MDMP loading sequences followed the capacity gain process very accurately, revealing a significant delay in the capacity gain process (see Figures 88 and 89), followed by a sharp increase. Most of the cases used by Paikowsky et al. (1995) did not contain such detailed data and, hence, large time intervals between load tests resulted in a significantly more moderate slope of the capacity gain process. For example, when connecting the initial normalized capacity values to the final ones, the Cgt for peak values of NB2 and NB3 were 0.44 and 0.33, respectively.
(2) Based on knowledge acquired from other studies, the MDMP was designed such that the pore pressure measurements were situated away from both ends of the pile. This location ensured "true" radial dissipation/consolidation at the point of measurement and, hence, significantly longer dissipation time compared to cases in which other effects took place, allowing possible vertical dissipation.
(3) The number of case histories for which capacity was monitored with time is limited. In many of those cases, the final load test may not necessarily have represented the end of consolidation and/or the maximum capacity (as may be the case for test NB2). Obtaining a t75 based on normalization to the maximum load (not maximum capacity) would result in a time shorter than the one actually required for the entire capacity gain process. Even the concept of using t75 when monitoring capacity gain was developed as a result of a lack of data in the initial stages of capacity gain, along with the possible failure to complete the process (Paikowsky et al., 1995).
(4) The influence of the multiple load testing on the interfacial shear strength was unclear when combined with all aspects of penetration, pore pressure build-up, radial stress changes, etc. It was, however, known that multiple shear of the same material contributed to its strength. In other words, the testing procedure of the phenomena itself affected the test results and, hence, resulted in a higher strength that was achieved over a longer period of time.
The average t75(152.4 mm) for the Newbury site was 887.2 h for peak values and 871.1 for residual values. The average value represented about a ±22% difference from the actual measurements, which seemed to be a very large variation within the same layer. Several possible explanations are: (1) The MDMP test NB2 was not completed and the final measured capacity did not necessarily represent the maximum shear transfer that actually existed. When using this value in the normalization process, naturally the obtained t75 was smaller than the actual, along with a decrease in the Cgt parameter. (2) A large variation in the soil was detected in the subsurface exploration study. Its influence on the radial consolidation process has yet to be assessed.
Table 33 presents the differences between the data collected at the Newbury site and the data presented by Paikowsky et al. (1995). The Cgt parameter found at the Newbury site was 1.6 times higher than values determined from the data set. The t75 obtained at the Newbury site was also larger than the one obtained at other locations, resulting in an overall longer capacity gain time even though the capacity gain was faster. This fact becomes apparent when examining Table 33; at 50% pore pressure dissipation, the pile designed with the Newbury site parameters would only gain 27.6% of the overall capacity, while the pile designed from data set parameters would gain about 50% of the capacity. Even at the 80% pore pressure dissipation, still only 57.2% of the capacity gain had taken place at the Newbury site.
Table 33. Evaluated Gain of Capacity (Adjusted to 152.4-mm Radius Pile) Based on the Newbury Test Results Compared With a Large Data Set.
Section 5.6 outlined the predicted measurements of the MDMP based on the data presented by Paikowsky et al. (1995). The following sections present a comparison between the measured and the predicted values and the associated observations and conclusions. Since sections 7.1 and 7.2 showed some similar comparisons, references will be made to those sections when appropriate.
7.3.2 Pore Water Pressure Increase Due to Driving
As presented earlier, the pore water pressure increased markedly from hydrostatic pressure to an elevated level due to the effects of driving. Referring to Table 30, the pore water pressure increased a total of 159.2 kPa and 135.5 kPa for MDMP tests NB2 and NB3, respectively. When normalized with the vertical effective stress without considering the additional stresses caused by the embankment located near the test site, the normalized excess pore pressure was 1.98 and 1.31 for MDMP tests NB2 and NB3, respectively. These values had been added to the data set compiled by Paikowsky et al. (1995) and are presented in Figures 90 and 91. When considering the possible effects of the embankment, the normalized excess pore pressure decreased to 1.74 and 1.14 for MDMP tests NB2 and NB3, respectively. The normalized excess pore pressure obtained from MDMP test NB2 was lower than the mean, however, within 1 standard deviation of the data presented by Paikowsky et al. (1995). The data for MDMP test NB3 was also lower than the mean, however, outside the 1 standard deviation range. Therefore, both measurements of the initial excess pore pressure (ui) seemed lower than those anticipated. This observation was confirmed by u3 measurements of CPT tests showing u3/sv' values of about 2.3 at approximate elevations of -3.05 and -9.14 m (-10 and -30 ft) (Paikowsky and Chen, 1998). These observations suggest two possibilities: (1) incomplete saturation of the MDMP pore pressure system (air in porous stones) and/or (2) insufficient frequency of data collection not able to record the initial peaks in the pore pressure measurements.
7.3.3 Excess Pore Pressure Dissipation and Capacity Gain
Figure 92 presents the range of originally predicted pore pressure dissipation, u(t)/ui, and capacity gain ratio, Rs(t)/Rsmax, based on data presented by Paikowsky et al. (1995). Figure 92 also depicts for the same ratios, the actual measured relationships for both tests NB2 and NB3. From this graphical representation, the observations presented earlier in section 7.2.3 became obvious. The pore pressure dissipation rate was faster than the predicted range as was evident by an increased negative slope of the measured pore pressure lines for both tests NB2 and NB3. The relationships in Figure 92 were based on the measurements described in Chapter 6 and the rates described earlier. If the aforementioned assumption was correct and i should have been higher as suggested by the CPT tests, then the dissipation relationship would have started earlier, hence, better matching the predicted dissipation zone.
The capacity gain rate in Figure 92 was faster than the predicted range as shown by an increased positive slope of the measured capacity gain lines for both tests. Also, at approximately 50% dissipation, only about 25% to 30% of the capacity gain had occurred, whereas the predicted ranges indicated that at approximately 50% pore dissipation, about 50% of the capacity gain would have occurred. The practical completion of the process however, was closer to the predicted values, relating the decrease in the excess pore pressure to 20% of the initial excess pressure build-up, and to the capacity increase to the level of 80% of the maximum capacity.
Figure 93 presents the effect of pile radius on the time for 50% excess pore pressure dissipation (t50) for clays with an OCR of 1 to 2. The values of t50 for MDMP tests NB2 and NB3, respectively, were 35476 s and 28256 s. Both of these values fell within the range previously presented by Paikowsky et al. (1995), indicating that t50 may be estimated from data presented in Figure 93 and further validates the normalization procedure as presented in section 5.6.3. If, however, the actual initial pore pressure was higher and earlier (based on the CPT), then t50 would have decreased and better matched the data presented in Figure 93.
7.4 Radial Consolidation
The radial coefficient of consolidation, ch, can be evaluated from the dissipation tests. The pore pressure filter element location was positioned to ensure radial dissipation and, hence, the radial (horizontal, cylindrical) consolidation equation was being used:
for which T50(h) is the time factor associated with 50% radial consolidation, r is the piles radius, and t50 is the time for 50% excess pore pressure dissipation. According to Levadoux and Baligh (1986), T50(h) was 33 and, hence, the coefficients of horizontal consolidation were as follows:
These values were about half of the ch values calculated from the CPT dissipation tests. For the three CPT tests around 10 to 10.7 m (33 to 35 ft) penetration, ch = 0.034 cm2/s. For the two CPT tests around 13 m (43 ft) penetration, the ch varied between 0.07 and 0.16 cm2/s (see Paikowsky and Chen, 1998). The difference between the coefficients of consolidation obtained from the CPT and the MDMP was most likely the result of the variation in the initial pore pressure measurements that markedly affected the determined t50.
7.5 Time-Dependent Radial Stresses
The total radial stress decreased for 35 h after the end of driving. At that time, a dramatic increase of about 110 kPa (16 psi) took place where the radial stresses became approximately constant at about 200 kPa (29 psi). A thorough analysis was required to explain the phenomenon in detail; however, a possible interpretation was offered. When the pile was installed, the penetration was accompanied by large soil displacements, remolding, and shear, resulting in a substantial build-up of pore water pressure and significant changes in radial stresses. The clay at this stage was in a complete liquid-plastic state, with total pressures approximately equal to that of the pore water pressure, and with a possible zone of water around the pile's shaft. These pressures resulted in about zero radial effective stresses and a negligible interfacial friction. The described initial stage was well depicted in Figure 94a, which outlined the pore pressure, total stresses, and resulting effective stresses on a logarithmic time scale. In the "classical" one-dimensional consolidation process, this initial stage would have been followed-up by pore pressure dissipation accompanied by an equal increase in the effective stresses. The presented data in Figure 94a and b clearly showed a different process in which both the pore pressure and the total stress decreased with time. The decrease in the total stresses could be explained by redistribution of stresses around the pile following the driving. The total stresses, as a result of external loading, remained constant in a one-dimensional consolidation condition. In the pile penetration process, the loading due to penetration was fast, creating an immediate increase in stresses in an area surrounding the pile, extending to a point within the far field in which no change in the stresses was noticed. This zone did not remain stable, however, and as a result, changes took place in which the high stresses first developed next to the pile wall and transferred outwards, redistributing the load to soil further away from the pile wall. This process of radial soil pressure redistribution continued in parallel to the pore pressure dissipation at a rate somewhat lower than that of the pore water pressure dissipation. As a result, the radial effective stresses increased at a very gradual rate (of approximately 1.5 kPa/h on a semi-log scale) for the first 35 h after driving. The gradual increase of the radial effective stresses from approximately 0 kPa at the end of driving to approximately 56 kPa at 35 h after driving was probably accompanied by the disappearance of the water film around the pile, the change of the soil state from plastic liquid to plastic solid, and the increase of the contact between the soil and the pile wall. At that time (about 35 h after draining), a sudden increase in the total stresses (and accordingly in the effective stresses) took place. While the previous process might explain the events, the actual increase might have been triggered by a load test that was carried out at the time, accompanied by the transformation of the shear zone from interfacial shear to internal shear within the soil some distance away from the shaft.
7.6 The Relationship Between Pore Pressure Dissipation, Frictional Capacity Gain, and Radial Stresses
Figure 95 presents a summary of the different parameters affecting the shear resistance along the MDMP. The absolute frictional resistance increased with time, starting with values close to zero at the end of driving to about 25 kPa (3.5 psi) approximately 5 days later. The increase of the frictional resistance was consistently accompanied by an increase in the radial effective stresses. This rough observation suggests that for frictional material, the ratio between the shear resistance and the normal stress remained approximately constant. Such relationships are seen in the lower part of Figure 95, presenting the ratio between the frictional force to the radial effective stress.
When examining the relationships presented in Figure 95, it is important to note that the shear measured along the frictional sleeve did not necessarily take place along the pile/soil interface. As a matter of fact, when the MDMP was pulled out of the ground, a layer of clay (with a diameter equal to the internal casing diameter) was attached to the pile, clearly indicating that shear was taking place away from the interface. Correcting for such observations would result in lower frictional stresses along the sleeve at a later time after installation and approximately constant frictional resistance from about 35 to 45 h after penetration. The timing was closely associated to the period for which a large increase in total and effective stresses was observed.
7.7 Observed Heave
Figure 63 indicated a sharply increased surface load cell force measurement following the driving of model pile test NB2. This increase was attributed to the pile/soil upward movement (i.e., heave) as no displacement was measured at the surface load cell location. This behavior, on a smaller magnitude, repeated itself whenever there was a stop in the pile motion as shown in the white areas of Figure 63. The initial heave load of 7.17 kN (1612 lb) was subtracted from the measurements and the net measured forces were presented in Figure 64. The data in Figure 64 suggested that when the top motion was stopped (no change in top displacement with time, e.g., from about 25.5 to 26.5 min after the start of installation), the internal MDMP load cell measurements gradually decreased, whereas the surface load cell recorded an increase in force. Details of the measurements are presented in Figure 65. This behavior may be interpreted in the following ways: (1) The increase in the load was due to heave, in which the pile moved upwards, and (2) the pile moved together with a mass of soil around it. As the load cells inside the pile did not record an increase in load, the only possibility was shear in the soil some distance away from the pile. These observations were further supported by measurements conducted during the pushing period in which the surface load cell continued to measure the heave effect, while the internal load cell measurements reflected shear that took place due to loading.
No heave effects were observed during MDMP test NB3. The differences between the tests may be explained through the variation in the soil conditions with depth. The soil in which NB2 was driven was stiffer and the driving resistance was about 10 blows per 100 mm. During installation of MDMP test NB3, the pile was penetrating under its own weight prior to driving and was held in place while the hammer was attached. The driving resistance for NB3 was about four blows per 100 mm at the end of driving.