Keywords: soil consolidation, Physics Informed Neural Networks (PINNs), coefficient of consolidation, artificial neural network, soil subsidence., geotechnical modeling.

The coefficient of consolidation is a critical parameter in geotechnical engineering, influencing the design and safety of various infrastructure projects. Traditionally, coefficient of consolidation is estimated through laboratory consolidation tests, which are time-consuming, subject to operator variability, and may not accurately reflect in-situ conditions. The challenge becomes more significant when site-specific information is incomplete or unavailable. With the increasing integration of Physics-Informed Neural Networks (PINNs) in geotechnical modeling, this study proposes a novel PINN-based framework that incorporates the Mikasa’s one-dimensional consolidation equation to estimate site-specific coefficient of consolidation values under varying data availability. In this study, we developed a PINN-based model to estimate the coefficient of consolidation using only subsidence data, without requiring explicit information on embankment surcharge loading history. The results demonstrate that the proposed approach can reliably infer coefficient of consolidation capturing the essential features of the consolidation process. This work highlights the adaptability, efficiency, and physical consistency of the PINN framework, particularly in data-scarce geotechnical settings. By reducing dependence on traditional laboratory testing and prior loading records, this approach offers a scalable and interpretable alternative for consolidation analysis in both research and practical applications.

download the full-text of this paper (PDF, 8 pages, 533 Kb)

go to the previous paper
go to the next paper
return to the table of contents
return to the volume description