Validation of bearing capacity methods for working platforms with test and numerical data

Several bearing capacity calculation methods have been used to design working platforms over soft soils, yet their predictive performance remains largely unvalidated. This lack of validation leaves engineers with doubts and uncertainties about the accuracy and use of these methods in their applications. This study assessed the predictive performance of widely used approaches for working platforms, including those in recent industry guides and literature. A dataset of 85 cases, comprising small-scale, centrifuge, field, and numerical studies, was compiled to evaluate the accuracy of these approaches. Key parameters, including granular layer thickness, subgrade soil undrained shear strength, and footing type and dimensions were systematically investigated. While the BR470 and Lawton-Han methods generally demonstrated closer agreement with observed bearing capacities, traditional approaches such as Load Distribution and Punching Shear showed varying degrees of bias, either over- or underestimating the ultimate bearing capacity. No single method consistently captured the influence of granular layer thickness across all test conditions. Under very soft subgrade conditions, all approaches showed diminished estimation accuracy, as evidenced by increased scatter and deviation from a unity bearing capacity ratio, even after modifications, while most methods produced conservative results for stiffer subgrades. A practical relationship was developed to estimate the critical granular layer thickness in layered systems, reflecting a nonlinear relationship with the bearing capacity ratio between the lower and upper layers. These findings are expected to contribute to the informed use and further development of design methods for more accurate and reliable bearing capacity predictions in unreinforced working platforms over soft subgrade and provide justifications for updating the current design guides and practices.