This study presents an evaluation of the performance of eight cone penetration test (CPT) methods in predicting the ultimate load carrying capacity of square precast prestressed concrete (PPC) piles driven into Louisiana soils. A search in the DOTD files was conducted to identify pile load test reports with cone penetration soundings adjacent to test piles. Sixty piles were identified, collected, and analyzed. The measured ultimate load carrying capacity for each pile was interpreted from the pile load test using Butler-Hoy method, which is the primary method used by DOTD. The following methods were used to predict the load carrying capacity of the collected piles using the CPT data: Schmertmann, Bustamante and Gianeselli (LCPC/LCP), de Ruiter and Beringen, Tumay and Fakhroo, Price and Wardle, Philipponnat, Aoki and De Alencar, and the penpile method. The ultimate load carrying capacity for each pile was also predicted using the static "-method, which is used by DOTD for pile design and analysis.

Prediction of pile capacity was performed on sixty piles, however, the statistical analyses and evaluation of the prediction methods were conducted based on the results of thirty five friction piles plunged (failed) during the pile load tests. End-bearing piles and piles that did not fail during the load tests were excluded from the statistical analyses.

An evaluation scheme was executed to evaluate the CPT methods based on their ability to predict the measured ultimate pile capacity. Four different criteria were selected to evaluate the ratio of the predicted to measured pile capacities. These criteria are: the best-fit line, the arithmetic mean and standard deviation, the cumulative probability, and the Log Normal distribution. Each criterion was used to rank the prediction methods based on its performance. The final rank of each method was obtained by averaging the ranks of the method from the four criteria. Based on this evaluation, the de Ruiter and Beringen and Bustamante and Gianeselli (LCPC/LCP) methods showed the best performance in predicting the load carrying capacity of square precast prestressed concrete (PPC) piles driven into Louisiana soils. The worst prediction method was the penpile, which is very conservative (underpredicted pile capacities).