Reinforcement Learning Algorithm Based Stress Wave Modal Optimization and Deicing Efficiency of Transmission Line

作者：Qi Ding, Shicong Deng, Yu Luo, Xiaolei Niu, Rongbang Wu · 年份：2024 · DOI：10.1109/eeps63402.2024.10804380 · 被引用次数：2 · 研究领域：Vibration and Dynamic Analysis、Thermal Analysis in Power Transmission、Icing and De-icing Technologies

This study introduces a reinforcement learning-based stress wave modal optimization method to improve the de-icing efficiency of transmission lines. Detailed experiments demonstrated that the optimized stress wave parameters significantly enhanced de-icing performance, reducing de-icing time by 30%, decreasing energy consumption by 25%, and increasing de-icing area by 40%. The reinforcement learning algorithm exhibited strong convergence and stability, as indicated by consistent reductions in the loss function and increases in reward values under various conditions. These findings validate the proposed method's effectiveness and reliability in optimizing stress wave applications for de-icing processes.