Research on Train Energy Optimization Based on Dynamic Adaptive Hybrid Algorithms

To address the challenges of locomotive and track modeling and the poor convergence of intelligent algorithms in train energy optimization, a multi-objective optimization model is proposed. Based on the uniform bar dynamics model, an interval division method for constant slope resistance values is developed to improve the applicability and accuracy of the energy consumption model under complex track conditions. Additionally, dynamic inertia weights and learning factors are introduced into the PSO-SA algorithm to enhance the algorithm’s adaptive adjustment capabilities at different optimization stages, alleviating the conflict between global search and local convergence. The proposed method not only improves the convergence speed of the solution but also optimizes train speed profiles, reducing traction energy consumption and improving punctuality. Simulation studies carried out using the new reference line demonstrated a 19% reduction in average train energy consumption, validating the correctness and feasibility of the proposed method, which shows great potential for applications in the field of automatic energy-saving driving for trains.