Performance evaluation and optimization of recycled concrete pile foundations based on deep learning

This study proposes a deep learning-based method for the performance evaluation and optimization of recycled concrete pile foundations. By integrating the Particle Swarm Optimization (PSO) algorithm and the Simulated Annealing (SA) algorithm to improve the performance of the BP neural network, a hybrid SA–PSO–BP prediction model is constructed to achieve the prediction of vertical bearing capacity and settlement of recycled concrete pile foundations. Case studies demonstrate that the SA–PSO–BP model controls the prediction errors for pile foundation bearing capacity and settlement within ±8%, outperforming other models. This model excels particularly in capturing time-dependent characteristics under long-term loads, providing a reliable intelligent decision-making tool for the design optimization and safety assessment of pile foundation engineering.