Combined Pile-CPT Method: Using Log-Normal Distribution Properties to Optimize the Estimation of Ultimate Pile Capacity

摘要

There are different methods proposed for estimating the ultimate axial capacity of the piles using cone penetration test (CPT). These methods represent correlations between CPT data and pile capacity based on different properties such as the pile width, pile type, installation method, soil type, etc. However, there is no perfect pile-CPT method that reflects all the factors affecting these correlations, instead, each pile-CPT method considers some of them. In this paper, the results of 80 pile load test cases of precast prestressed concrete (PPC) piles driven in Louisiana, were studied using seven (7) different pile-CPT methods: LCPC, ERTC3, probabilistic, UF, De Ruiter and Beringen, UWA, and Schmertmann. The results of these methods were compared and their performance was examined for different cases where different percentages of the pile capacity is due to sand layers. The multidimensional unfolding (MDU) technique and the log-normal distribution of estimated to measured pile capacities for these pile-CPT methods were used to develop a novel combined method that enables designers to estimate, more reliably and accurately, the ultimate capacity of driven piles for different soil categories.