Detection Algorithm for Malicious Control Commands in Active Distribution Networks Based on GloVe

Abstract

Malicious control commands posing a critical threat to the reliable and solid operation of active distribution networks (ADN), necessitating more effective identification and detection methods than traditional approaches. This paper proposes a detection algorithm for malicious control commands in ADN based on the global vectors for word representation (GloVe) model. The GloVe model converts control commands into high-dimensional word vectors, and the coupling degree between the command and physical logic semantics is analysed using cosine similarity, thereby detecting malicious control commands. Experimental simulations demonstrate that this method performs excellently in terms of accuracy and detection performance, effectively identifying and preventing malicious control commands, thereby ensuring reliable and solid operation of the ADN.