Pile test data and discussion

Chapter 6 Testing data and discussions 139 CHAPTER 6 PILE TEST DATA AND DISCUSSION 6.1 Introduction The results of the pile load tests (constant rate of penetration tests at different rates, statnamic tests at different peak loads, and maintained load tests will be presented and discussed in this chapter. Based on the existing models, which are used to derive the static load-settlement curve from that of a rapid load pile test, a new model will be proposed. It was found that the quake values of the pile shaft resistances of the rapid load tests were much higher than those of the static load tests. An existing theoretical model will be modified and applied for the static load pile tests to build the relationship between load and settlement and to quantify these quake values, and then it will be developed for the rapid load pile tests. The gradual decrease of the pile shaft resistance after its peak value to a residual pile shaft resistance, which is known as the softening effect, plus the changes of pore water pressures and the inertial behaviour of the soil around the pile will be reported and discussed. 6.2 Typical results of the pile load tests Several measurements for each test were obtained from the instrumented model pile and clay bed instruments. The following definitions of the testing components will be used: ♦ Total pile load - load applied to the pile top and measured by a load cell mounted at the pile top. ♦ Measured pile shaft load - load measured by the shaft load cell. ♦ Pile tip load - load measured by the pile tip load cell. Chapter 6 Testing data and discussions 140 ♦ Total pile shaft load. It was not possible to measure the total pile shaft load directly. Therefore, it was deduced by subtracting the pile tip load from the total pile load. ♦ Pile settlement - vertical displacement from its original pre-load test position measured by an LVDT mounted at the pile top. ♦ Pile velocity - deduced from the pile settlement with time or from an accelerometer which was incorporated in the pile. ♦ Pore water pressures at the pile tip and on the pile shaft - measured by the pile tip and pile shaft pore water pressure transducers. ♦ Pore water pressures in the clay beds - measured by pore water pressure transducers which were incorporated in clay beds at different locations. ♦ Soil accelerations in the clay beds at different locations - measured by accelerometers located at different locations in the clay beds. ♦ Top and side chamber pressures. - measured by Druck PDCR 810 water pressure transducers. Typical results of pile load tests are shown in Figure 6.1 to 6.4. In the following sections these results will be discussed and analysed in more detail. 6.3 Pile shaft load results and models for the pile shaft load The total pile shaft load which was deduced from the total pile load and the pile tip load was less reliable than the measured pile shaft load as the pile tip load cell worked unreliably (see Section 6.4). Therefore, the measured pile shaft load acting on the shaft load cell rather than the total pile shaft load will be presented and used in this section. With a CRP test, the rate of shearing was not constant but it increased gradually from zero to the target rate (Figure 6.2). Normally, the desired rate for a CRP test was only achieved when the pile shaft load had reached the ultimate load value. Therefore, the rates of shearing, which were obtained from measured pile settlements, will be used in deriving the pile shaft static load from the pile shaft rapid load. In reality a CRP test at a high rate is similar to the first part of a statnamic test. Chapter 6 Testing data and discussions 141 Smith (1960) worked on the dynamic resistance of a pile and proposed a linear dependence of the damping resistance upon the shearing rate. However, the nature of the non-linear relationship between the damping resistance and the shearing rate cannot be ignored. For this reason, Gibson and Coyle (1968) carried out triaxial tests at different rates of shearing for both sand and clay and proposed a non-linear damping resistance. Following this, several researchers have worked on this problem and proposed several soil models for the relationship between the damping resistance and the rate of shearing. In order to examine the capability of the non-linear damping models several typical models will be used for analysis of the test results. Following this a new model will be proposed to derive the load-settlement curve of a static pile load test from a statnamic pile load test. 6.3.1 Non-linear models The following non-linear damping resistance models will be used for examination of the test data: ♦ Gibson and Coyle’s model (1968): This non-linear damping resistance is expressed in the form: Rt = Rs + RsJTvN (6.1) where Rt is the total dynamic resistance Rs is the static resistance JT is the damping factor v is the velocity of shearing N is the parameter drawn from the test results, which was 0.18 for clays and 0.2 for sands. ♦ Randolph and Deeks’ modified model (Hyde et al. 2000): ( )ββ ααττ 6101 −−⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ Δ+= os d v v (6.2) where τd is the dynamic shear resistance τs is the static shear resistance vo is the reference velocity (taken for convenience as 1 m/s) Chapter 6 Testing data and discussions 142 Δv is the relative velocity between the pile and the adjacent soil α, β are the damping coefficients ♦ Balderas-Meca’s model (2004): ⎥⎥⎦ ⎤ ⎢⎢⎣ ⎡ ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛−⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛+= 20.020.0 01.01 oos d vv vατ τ (6.3) where τd is the dynamic shear resistance τs is the static shear resistance vo is the reference velocity (taken for convenience as 1 m/s) v is the relative velocity between the pile and the adjacent soil α is the damping coefficient. It is a function of the pile displacement and varies linearly from zero when the pile displacement is zero to 0.9 when the pile displacement is 1% of the pile diameter. After the pile settlement achieves 1% of the pile diameter the α parameter remains constant. Due to the dependence of the damping parameters upon soil properties two sets of damping parameters for each model will be used for the examination of the pile test data. The first set of damping parameters are those proposed by the model’s authors, and the second set is deduced from the best match between the test data and the models. Thus, the two sets of the damping parameters for three models will be determined as follows: ♦ Gibson and Coyle’s model: These authors only recommended N = 0.18 for clay meaning only one set of damping parameters is applied in the model, with the JT value achieved by a best match between the test data and the model. ♦ Randolph & Deeks’ modified model: α = 1 and β = 0.2 are used for the first set of damping parameters and β = 0.2 and α achieved by the best match process are the second set of damping parameters for the model. ♦ Balderas-Meca’s model: β = 0.2 and α increasing linearly from zero when the pile displacement is zero to αmax = 0.9 when the pile displacement is 1% of the pile diameter are used for the first set of the damping parameters. β = 0.2 and αmax obtained from the best match process are used for the second set of damping parameters. Thus, β = 0.2 and α increasing linearly from zero when the pile displacement is zero to αmax when the pile displacement is 1% of the pile diameter are Chapter 6 Testing data and discussions 143 the second set of the damping parameters. After the pile displacement reaches 1% of the pile diameter α is constant and equals αmax. It was desirable to produce highly repeatable clay beds. However, test results showed that the five clay beds exhibited slightly different damping characteristics. To demonstrate this two CRP tests at rates of 0.01mm/s and 100mm/s in each bed are compared in Figure 6.5 to 6.9. The ratios of the maximum measured shaft loads of the CRP test at a rate of 100mm/s to that of the CRP test at a rate of 0.01mm/s are 1.97, 1.54, 1.82, 1.74, and 1.95 for clay Bed 1 to Bed 5 respectively. Taking the average value of five ratios, 1.8, as a benchmark, the deviations of the ratios from the benchmark are: 9.30% for Bed1; 14.56% for Bed 2; 0.77% for Bed 3; 3.32% for Bed 4; and 7.81% for Bed 5. Due to this variation the matching process between the models and the test results to deduce the damping parameters will be carried out independently for each bed. The pile bearing capacities in constant rate of penetration tests at the rate of 0.01mm/s are taken as the pile static bearing capacity benchmark and are used for comparison with the derived pile static bearing capacities. However, the pile static bearing capacity benchmark for each clay bed was not constant since the consolidation was maintained during the testing programmes. In addition, significant local consolidation developed on the soil around the pile under a series of pile load tests. For this reason, the CRP test at a rate of 0.01mm/s was repeated several times in each bed and a rapid load pile test used the nearest CRP test at rate of 0.01mm/s as its static bearing capacity benchmark. A rapid load pile test and a CRP test at the rate of 0.01mm/s in each clay bed will be used to check the suitability of the existing models. The static load-settlement curves, which are derived from the dynamic load-settlement curves by using the chosen models, will be compared with the measured static load-settlement curves which are the results of CRP tests at the rate of 0.01mm/s. Chapter 6 Testing data and discussions 144 The comparison of some load-settlement curves are shown in Figures 6.10 to 6.24. Combining the results of all pile load tests in which some are shown in Figures 6.10 to 6.24 the following conclusions can be deduced: ♦ Rate effects are present and the damping resistance is non-linear with the rate of shearing as suggested by Hyde et al. (2000). ♦ The quake of the pile shaft load, i.e. the penetration at which the pile shaft load reaches the ultimate resistance, in a dynamic pile load test is larger than that of a static pile load test. The quakes in static pile load tests for pile shaft load vary between 0.5% to 0.9% of the pile diameter (70mm), whereas the quakes of dynamic tests vary over a wide range and the quicker the loading rate the larger the quake value. The largest pile shaft quake for rapid load pile tests is 5.4% of the pile diameter (B4/12/CRP-400). ♦ The pile shaft load-settlement curve of a rapid load pile test can be subdivided into three sections as shown in Figure 6.21: i) in the first section the relationship between the pile shaft load and the pile settlement was approximately linear; ii) in the second section the relationship between the pile shaft load and the corresponding pile settlement was non-linear; iii) finally, in the third the pile shaft load reached the ultimate value and remained approximately constant with further pile settlement. On the other hand the pile shaft load-settlement curve of a static pile load test either did not exhibit or exhibited a much less obvious second section. Normally, in the first section the settlement of a rapid load pile test was larger than that of a static pile load test. Figure 6.21 shows that the first section of the rapid load pile test was complete at a settlement of about 0.76mm whereas that of the static pile load test was complete at a settlement of 0.6mm. Due to this it is difficult to derive the pile shaft static load- settlement curve from that of a dynamic pile load test since the shape of the load- settlement curve of a dynamic pile load test is not similar to that of a static pile load test. ♦ The damping load not only depends upon the shearing rate but also upon the soil loading stages described above. It can be seen from Figure 6.10 to 6.24 that the damping load develops gradually from the first section to the final section. ♦ Comparing the derived load-settlement curves of the Randolph and Deeks and Gibson and Coyle models it could be said that Randolph and Deeks’ model is only another form of the Gibson and Coyle model. Chapter 6 Testing data and discussions 145 ♦ In general, the three models predict damping load well when the second section of the load-settlement curve of a dynamic pile load test develops over a small settlement (Figure 6.13 to 6.15) whereas they overpredict the damping loads of the first and the second sections of the load-settlement curves when the second section develops over a large settlement (Figure 6.19 to 6.24). ♦ Balderas-Meca’s model takes into consideration the development of the damping load between the soil stages by changing the damping parameter, α, with the pile settlement. However, quake values for the pile shaft load are over a relatively wide range so that the quality of the derived load-settlement curve is not consistent (Figure 6.12 and 6.15). ♦ The viscous damping that occurs during a rapid load pile test can be represented by a non-linear power law incorporating damping coefficients. ♦ Apart from the above models other models reported in the literature review have also been examined. However, they did not provide a good prediction for statnamic load tests. The main reason is that these models were only proposed for the ultimate shear resistance. ♦ The soil damping characteristics depend on soil properties, i.e. liquid and plastic limits. Summarizing data from previous studies, Hyde et al. (2000) showed that the damping parameter, α, can vary by orders of magnitude for different clays. 6.3.2 A new non-linear model (the Proportional Exponent Model) for pile shaft rate effects In pile load tests settlement criterion, which is normally chosen as 10% of the diameter of the pile, rather than the pile ultimate load is used for the determination of a pile’s capacity. Therefore, it is desirable to derive the load-settlement relationship for a static condition from that of a statnamic test. The most widely used Unloading Point Method (UPM), which assumes that the damping load is linear with the shearing velocity and can overestimate the ultimate static pile capacity by up to 30% for piles in clay. The non-linear power laws reviewed in Section 6.3.1 could predict the ultimate static pile bearing capacity well. However, in order to get a better estimate the pile’s bearing capacity at loads below the ultimate static pile bearing capacity, Chapter 6 Testing data and discussions 146 which is the zone of most interest in determining the pile’s serviceability, the available non-linear power laws need to be modified. The available non-linear power laws should be modified in such a way that they can simulate the gradual development of the damping load from the first to the final stages of loading. This can be achieved by gradually increasing the damping parameters with the development of the dynamic pile load, and the damping parameters are constant when the pile’s dynamic load reaches the ultimate value. Thus, the new non-linear power law should be in the same form as the available models when the pile’s dynamic bearing capacity reaches its ultimate value. It is proposed therefore to use a model with a proportional exponent of the velocity term, the general form of which is as follows: ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛+= )(1 ultimated d s d s d v v τ τβ ατ τ (6.4) where τd is the dynamic shear resistance τs is the assumed static shear resistance determined at a pile velocity of 0.01mm/s τd(ultimate) is the ultimate dynamic shear resistance vd is the pile velocity which can vary from zero to 2500mm/s during a statnamic test (Japanese Geotechnical Society, 2000). vs is the assumed static pile velocity which is 0.01mm/s in this study. α, β are the damping coefficients The damping parameters depend on the soil properties. However, from previous researches (see Section 2.9), combined with test results from this study, the damping parameter β = 0.2 can be used for clay. Thus Equation 6.4 becomes: ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛+= )( 2.0 1 ultimated d s d s d v v τ τ ατ τ (6.5) The validity of this equation will be examined by using the results of the calibration chamber pile load tests. When the pile shaft resistance reaches the ultimate value τd/τd(ultimate) = 1 and Equation 6.5 becomes: Chapter 6 Testing data and discussions 147 2.0 )( )( )( 1 ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛+= s ultimated ultimates ultimated v vατ τ (6.6) where τd(ultimate) is the ultimate dynamic shear resistance. τs(ultimate) is the ultimate static shear resistance. vd(ultimate) is the shear velocity corresponding to the ultimate dynamic shear resistance. Due to the rate effects of the five clay beds being slightly different as mentioned in Section 6.3.1 the damping parameter will be deduced independently for each bed. Equation 6.6 is used to calculate the damping parameter, α, from the pile shaft load ratio. Plots of τd/τs - 1 against vd are shown for Beds 2 to 5 in Figures 6.25 to 6.28. The best match between the test data and Equation 6.6 was achieved using least square regression gives the α values of 0.085, 0.135, 0.115, and 0.145 with the fitting regression coefficient R-squared values of 0.90, 0.88, 0.87, and 0.89 for Bed 2 to Bed 5 respectively. The average damping parameter (αaverage = 0.12) is deduced for all five clay beds as shown in Figure 6.29. It can be seen that the use of a damping parameter varying from α = 0.068 to α = 0.16 in Equation 6.6 covers for all five beds’ pile shaft dynamic loads (Figure 6.29). Figures 6.25 to 6.28 plus 6.29 show that damping effects are fairly consistent for each clay bed, but vary from bed to bed. This is thought to be attributed to the variation of the materials used as mentioned in Section 5.2. To examine the sensitivity of the damping parameter β, several values of β (0.16; 0.18; 0.22; 0.24) are plotted in Figures 6.25 to 6.28. In general β = 0.18 and β = 0.22 are lower and upper boundaries for the test data. In the following sections Equation 6.5 will be used to derive the full static load- settlement curves from the rapid load pile tests for Beds 2 to 5 using the damping values above. ♦ Clay Bed 1: As mentioned in Chapter 3 (see Sections 3.4.6 and 3.4.7) the pile was installed into the clay bed after 14 days of triaxial consolidation under a pressure of 280kPa. After pile installation the clay bed continued to be subjected to a pressure of Chapter 6 Testing data and discussions 148 280kPa. This installation during the triaxial consolidation caused disturbance between the pile shaft and the bored hole which could not be eliminated during the subsequent the second stage of consolidation. It was found that the pile bearing capacity of this bed was much lower than that of the other beds even though the clay beds underwent the same consolidation history. Only the results of one CRP test at a rate of 0.01mm/s (B1/1/CRP-0.01) will be used together with two rapid load pile tests, a CRP test at a rate of 100mm/s (B1/2/CRP-100) and a statnamic pile load test (B1/4/STN-15) since the others tests were mainly carried out to choose the best input parameters for the testing equipment for the pile load tests of the following beds. The CRP test at a rate of 0.01mm/s provides the static pile bearing capacity benchmark for the two rapid load pile load tests. The derived static load-settlement curves which are obtained from the rapid load pile tests by the proportional exp