Torque anomaly detection of nuclear power electric valve actuator based on DAE-WDSVVD

Abstract

Abstract The abnormal detection of nuclear power electric valve actuator components can effectively improve its operation safety and reliability. With the rise of artificial intelligence technology, data-driven fault diagnosis methods have become more and more popular. However, in practical application, there are few or almost no fault data of valve actuator. For the problem of anomaly detection of actuator components of valve in the scenario of only normal data, anomaly detection method based on the fusion of deep autoencoder (DAE) and weighted deep weighted support vector data description (WDSVVD) is proposed. It uses normal data to train the depth self-encoder, and the reconstruction error of the depth self-encoder to train the support vector data description. Compared with the traditional anomaly detection method, it significantly improves the anomaly detection accuracy and can realize more sensitive and robust component anomaly detection.