Identification and Protection of Power System Vulnerabilities Based on Autoencoder

In the face of an increasingly complex and ever-changing network attack environment, the security of the power system plays a crucial role. To cope with the constantly changing network threats, this paper uses autoencoder to identify vulnerabilities generated during the operation of the power system and provide protection strategies. Firstly, feature extraction is performed on the operational data in the power system to construct feature vectors. Then, an autoencoder is used to encode and decode the feature vectors, detecting abnormal data. Finally, classify the abnormal data to correspond to different types of vulnerabilities, and then make security protection choices based on different types of vulnerabilities. Based on the above security protection strategies, the vulnerability response time has been shortened, thereby ensuring the stable operation of the power system.