Multisource Heterogeneous Monitoring Data Fusion‐Based Transmission Line Galloping Prediction Method

Transmission line galloping poses significant hazards and seriously threatens the security of power systems. Therefore, timely and accurate prediction of line galloping is crucial to prevent and mitigate its damage. Existing galloping prediction methods typically rely on single‐channel optical fiber vibration signals, which suffer from incomplete information representation. These methods fail to effectively capture multisource real‐time monitoring parameters of the actual operating environment, compromising prediction accuracy and reliability. To address these problems, we propose a multisource heterogeneous monitoring data fusion‐based transmission line galloping prediction method. Firstly, a revised grey target model is developed to evaluate the influence of different multisource heterogeneous monitoring data on galloping prediction. Secondly, a heuristic‐optimized long short‐term memory (LSTM) algorithm with an improved information processing mechanism is proposed, where the sparrow algorithm adaptively optimizes key LSTM parameters. Finally, simulation results prove the superiority of the proposed algorithm in terms of prediction accuracy and convergence.