Substation misoperation prevention method based on improved convolutional neural network

The existing substation anti-misoperation system mainly uses logic locking to prevent errors, which has problems such as complex operating rules and susceptibility to human interference. Therefore, this paper proposes a substation misoperation prevention method based on improved convolutional neural network. By extracting the characteristics of substation wiring diagram, adjacency matrix is constructed to form a network topology model. On the basis of traditional convolutional neural network algorithms, SE channel attention has been added, and an improved convolutional neural network has been used to classify topological structure feature vectors, achieving misoperation type analysis. This paper directly realizes intelligent misoperation prevention in substations through deep learning, avoiding human factors and improving the accuracy and efficiency of misoperation prevention. This method can provide strong support for the safety production management of the power system and is of great significance for improving the safety and stability of power supply.