Estimating the Ultimate Capacity of Piles Using Machine Learning Models Based on CPT Data

There has been an increased interest recently on using the in situ cone penetration test (CPT) for estimation of the pile capacity. In this paper, three tree-based machine learning (ML) models, namely the decision tree (DT), random forest (RF), and gradient boosted tree (GBT), are developed for estimating the ultimate capacity of piles from CPT data. A database that comprises 80 pile load tests and associated CPT data collected in Louisiana was used to develop these ML models. The measured ultimate load capacity (Qm) was determined using the Davisson's interpretation method from the load-settlement curve from each pile load test. Among the developed ML models, the GBT demonstrated the most accurate ML model compared to others. The estimation of ultimate pile capacity from the GBT model is compared with those obtained from the four best-performed direct pile-CPT methods (based on a previous study), which are the UF, probabilistic, ERTC3, and LCPC methods. The GBT and pile-CPT methods were evaluated and ranked based on multiple statistical criteria analysis. The results clearly showed that the GBT model outperforms the four pile-CPT methods in estimation of the ultimate capacity of piles.