An Interpretable ML-based Model for Predicting p-y Curves of Monopile Foundations in Sand

2025-01-08

AI Summary

Modeling lateral pile response in sand and predicting p-y curves lack both accuracy and physical interpretability. To address this, we propose an interpretable machine learning framework integrating XGBoost with SHAP (Shapley Additive Explanations). Trained on a comprehensive database comprising field measurements and high-fidelity numerical simulations, the model achieves superior predictive accuracy for p-y curves—outperforming conventional empirical and semi-theoretical methods. Notably, this study is the first to apply SHAP analysis to pile–soil interaction, quantitatively elucidating the nonlinear contributions of key parameters—including embedment depth and relative density of sand—with attribution patterns strongly consistent with classical soil mechanics theory. The framework thus bridges the gap between data-driven prediction and mechanistic understanding: it delivers high accuracy while ensuring model transparency and physically grounded interpretability. This work establishes a new paradigm for intelligent, physics-informed geotechnical modeling.

Technology Category

• Machine Learning: Transparent, Interpretable, Explainable ML
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Predicting the lateral pile response is challenging due to the complexity of pile-soil interactions. Machine learning (ML) techniques have gained considerable attention for their effectiveness in non-linear analysis and prediction. This study develops an interpretable ML-based model for predicting p-y curves of monopile foundations. An XGBoost model was trained using a database compiled from existing research. The results demonstrate that the model achieves superior predictive accuracy. Shapley Additive Explanations (SHAP) was employed to enhance interpretability. The SHAP value distributions for each variable demonstrate strong alignment with established theoretical knowledge on factors affecting the lateral response of pile foundations.

Problem

• Predictive Modeling
• Soil-Pile Interaction
• Lateral Displacement

Innovation

• XGBoost model
• pile-soil interaction
• SHAP interpretability