Design method reliability assessment from an extended database of axial load tests on piles driven in sand

The accurate prediction of axial capacity remains a challenging task for piles driven in sands. Rigorous database studies have become key tools for assessing the efficacy of design methods. This paper employs the 117 high-quality entries in the recently developed Zhejiang University – Imperial College London (ZJU–ICL) database to check for potential biases between nine prediction procedures, considering a range of factors. The analysis highlights the critical importance of addressing age after driving, open and closed ends, tension versus compression, and concrete compared to steel. It also shows the hierarchy of reliability parameters associated with the alternative approaches. The “full” Imperial College pile (ICP) approach and The University of Western Australia (UWA) approaches are found to have significant advantages in eliminating potential biases. It is also argued that design load and resistance or safety factors should be varied to match the design and site investigation methods applied, as well as the loading uncertainty and degree of load cycling, which often vary between applications. Noting that predictions for base capacities Qb are inherently less reliable than for shaft Qs, especially in rapidly varying ground profiles, credible lower bound parameters (cone resistance, qc) are recommended for Qb assessment. It is also recommended that the potential effects of cycling be addressed carefully in cases that involve substantial environmental loading