using DLTs at EOD and several restrikes after that and one SNLT or SLT at the end was performed to quantify the increase of resistance (or setup) with time. The total pile resistance (R ) was estimated using a t PDA and CAPWAP from the DLTs and the resistances measured by the SLT and SNLT. The setup ratio for side (R /R ), tip (R /R ) and total (R /R ) resistances s s0 tip tip0 t t0 were also calculated for each TP. The results of test program showed that each of the TPs exhibited significant amounts of setup. Several methods (e.g., Skov and Denver, 1988; Bullock et al, 2005; Ng et al, 2013) were developed to predict pile setup with time after EOD. Among the developed models, the relationship developed by Skov and Denver (1988) is considered to be the most popular method due to its simplicity and accuracy in estimating pile setup. They proposed the following equation: where R = pile resistance at time, t; R = pile resistance at initial time, t 0; t = time elapsed since end of initial pile driving; t = reference t 0 In this study, t was selected as one day. The total (R t) and side (Rs) resistances were initial time after which the resistance increases with logarithmic increase in time; and A = rate of pile resistance (R /R ) increase per log cycle of time (t/t0 ). 0 t t0 best fitted to a linear logarithmic time scale, which suggests a linear logarithmic time relationship similar to the models proposed by Skov and Denver (1988) and others does exist for this load testing program. The tip resistance (R ) did not tip change significantly during all the restrikes for all the TPs with the exception of TP-2. As mentioned earlier, damage was noticed at the bottom 33 ft (10 m) of TP-2, which contributed to the lower R during the restrike for TP-2. tip t0 However, for the other TPs, any change in the R was mainly due to R setup, which t s is consistent with the findings in the literature (e.g., McVay et al, 1999; Ng et al, 2013), where the findings from the load tests showed that R mainly contributes to the setup process and R tip stays constant s over time during the load tests. The load test results also showed that for the four TPs, R was always overestimated during the SNLT compared to R tip computed by tip the restrikes, which lead us to conclude that the derived R from SNLTs are consistently greater those from DLT. tip Though both TP-3 and TP-4 were driven to two different locations, it is interesting to observe that both TPs had similar setup ratios (R /R = 4.4 to 4.7) and t t0 rate of setup (A = 0.23 to 0.24). TP-2 had different amount and rate of setup compared to the other two OE piles (i.e., TP-3 and TP-4). We suspect the difference is due to the damage at bottom 33 ft (10 m) of TP-2. Since the setup of the PSC pile (TP-1) is mainly due to outside friction and the setup rate for R of the PSC piles and OE s cylinder piles (TP-3 and TP-4) are close (i.e., A =0.28 and A =0.23 to 0.24), PSC OE one can postulate that no significant amount of setup was observed due to inside friction for the OE cylinder piles. Conclusions Based on this study, the following conclusions can be drawn: 1. The adhesion factor (α) calculated from the Statnamic load tests ranged from 0.47 to 0.75, with an average value of 0.61. The β-coefficient ranged from 0.20 to 0.26, with an average value of 0.22 for all soil layers in this study. These values are comparable to results from previous studies published in literature. 2. The unit side resistances of most soil layers, which were determined using differentiation of loads between strain gauge levels, mostly exhibited strain softening behavior, which ensured that the side resistance was fully mobilized. 3. The distribution of unit side resistance and design parameters (α and β) showed that the unit side resistance and α and β increased almost linearly with depth. 4. All the TPs (i.e., both open and close- end piles) exhibited significant amount of setup, which was mainly due to the increase in side resistance. The tip resistance was almost constant for the four TPs during the restrikes, except for TP-2. The lower tip resistance for TP-2 can be attributed to the damage caused during driving. 5. The load test results showed that for all the TPs, the inferred Rtip from SNLTs were consistently greater than those de- termined from DLTs in this particular site. 6. The logarithmic setup parameter (A) for the four TPs were backcalculated for both R and R . The results revealed that t s for the LDOE cylinder piles, A for R s was about 0.23 and for the PSC pile was about 0.28. The difference was mainly due to size and shape of the piles used in the study. Murad Y Abu-Farsakh, Ph.D., P.E., F.ASCE, is a re- search professor in the civil and environmental engi- neering department at Louisiana State University. Md Nafiul Haque, Ph.D., EIT, is an assistant project engineer at Ardaman and Associates in the Baton Rouge, Louisiana, office. Chris Nickel, P.E., is a pavement and geotechnical engineer administrator at Louisiana Department of Transportation and Development. Design Parameters and Pile Setup” was submitted for Session 320: Design, Testing, and Analysis of Pile Foundations of the 97 Transportation Research Board Annual Meeting (2018). An extended version of this article, “A Load Testing Program on Large-Diameter Open-Ended Instrumented Test Piles to Evaluate the th DEEP FOUNDATIONS • SEPT/OCT 2019 • 85
