A Generative Deep Learning and Explainable Machine Learning Framework for Heat Transfer Prediction and Analysis in Porous Structures with Oscillatory Flows

Abstract

Predicting and interpreting thermal performance under oscillating flow in porous structures remains a critical challenge due to the complex coupling between fluid dynamics and geometric features. This study introduces a data-driven framework that integrates generative deep learning, numerical simulation based on the lattice Boltzmann method (LBM), and interpretable machine learning to predict and explain the thermal behavior in such systems. A wide range of porous structures with diverse topologies was synthesized using a Wasserstein generative adversarial network with gradient penalty (wGAN-GP), significantly expanding the design space. High-fidelity thermal data were then generated through LBM simulations across various Reynolds (Re) and Strouhal numbers (St). Among ten machine learning models evaluated via nested cross-validation (Nested_CV) and Bayesian optimization, Extreme Gradient Boosting (XGBoost) achieved the best predictive performance for the average Nusselt number (R^2=0.9981). Furthermore, model interpretation using SHapley Additive exPlanations (SHAP) identified the Reynolds number, Strouhal number, porosity, specific surface area, and pore size dispersion as the most influential predictors, while also revealing synergistic interactions among them. For example, SHAP-derived interactive thresholds, including Re > 75 and porosity > 0.6256, provide practical guidance for enhancing convective heat transfer. This data-driven framework novelly integrates a hybrid approach to predict thermal performance in porous media under oscillatory flow, with the implementation of explainable machine learning, delivering both quantitative predictive accuracy and physical interpretability, offering guidelines for identifying and designing favorable oscillatory flow and structural conditions that enhance thermal performance in complex porous media.