An indirect data-driven model-updating framework to estimate soil–pile interaction parameters using output-only data

This paper presents a data-driven model updating framework to estimate the operational parameters of a laterally-impacted pile. The goal is to facilitate the estimation of soil–pile interaction parameters such as the mobilized mass and stiffness, as well as geometrical data such as embedded pile length, using output-only information. Accurate knowledge of mass, stiffness, and pile embedded length is essential for understanding foundation behavior when developing digital-twin models of structures for the purpose of damage detection. The method first employs subspace identification to determine modal parameters and quantifies their uncertainties using output-only data. The covariance matrix adaptation evolution strategy (CMA-ES), a stochastic evolutionary algorithm, is subsequently used to update the model. The effectiveness of the approach is demonstrated through its application to numerical models in this paper, to quantify errors, and subsequently to data from a documented full-scale field test of a pile subjected to an impact load. The work underscores the potential of statistical updating in advancing the accuracy and reliability of soil–structure interaction parameter estimation for systems where only output data might exist.