Fault Diagnosis Method of Spacecraft Control Systems Based on PCA-ResNet

Abstract

Due to the complex space environment, spacecraft telemetry signals are accompanied by a large amount of noise, and the accuracy of fault diagnosis is low by directly using the original telemetry signals. This paper proposes a fault diagnosis method for spacecraft control systems based on principal component analysis (PCA) and residual network (ResNet). Firstly, grayscale images are generated by denoising the telemetry signal of the spacecraft control system through PCA; Secondly, the images are input into the residual network to extract deep-level features; Finally, the Softmax classifier is used for classification to realize the fault diagnosis of the spacecraft control system. The research results show that the diagnostic accuracy of the method proposed in this paper reaches 95.33%, which is higher than other diagnostic models, and the method can be used for the actual fault classification of spacecraft control systems.

Improved time-variable transformer-based fault diagnosis method for satellite attitude control system

“…ElDali et al 17 applied a long short-term memory (LSTM) neural network to the Kepler spacecraft reaction-flywheel dataset and completed fault detection and failure prediction of the reaction flywheel. Wei et al 18 proposed a fault diagnosis method based on PCA and a residual network that can classify the faults of satellite ACSs. Yu et al 19 proposed a fault detection method based on a spatial-temporal generative adversarial network, which solved the problem of satellite telemetry data anomaly detection.…”