Pile foundations are essential in geotechnical engineering, especially when strong soil layers are too deep for shallow foundations. Besides supporting vertical loads, they must also resist lateral forces from wind and earthquakes. Accurately estimating pile bearing capacity is crucial for structural safety and cost efficiency. Conventional methods, such as those using SPT and CPT results, are mostly deterministic and often ignore the uncertainty in soil properties—an issue especially relevant in tropical regions like Indonesia with highly variable subsurface conditions. This study applies a Bayesian probabilistic approach to analyze pile bearing capacity using CPT data from Tumbak Bayuh, Bali. Typical soil parameters from literature are used as priors and updated with site-specific data. The results indicate that Bayesian inference offers more realistic and robust estimates than traditional methods by incorporating uncertainty and allowing continuous refinement as new data are added. The study also introduces a Bayesian model to assess how sample size affects parameter estimation, comparing it with frequentist approaches such as least squares and maximum likelihood. Findings highlight the strength of Bayesian updating in capturing uncertainty, offering a systematic and adaptive framework for pile analysis—especially beneficial in data-scarce or variable geotechnical environments.
I Wayan Ariyana Basoka (Mon,) studied this question.