Satellite Fault Diagnosis Technology Based on Intelligent Signal Processing
Shuo Jiang, Siyue Jiang, Xiaorui Zhang, Jianyu Liu, Hetong Gao, Chengzhao Shan, Zhuoming Li · 2025
This paper investigates the intelligent fault diagnosis technique for the power converter in the main circuit of the Battery Charge and Discharge Regulation (BCDR) module within satellite power supply and distribution systems. To address potential limitations of traditional shallow networks in fault diagnosis, a Deep Belief Network is employed for fault detection. The study also explores the optimization effects of various algorithms on the parameters of the DBN model. The DBN fault diagnosis model proposed in this paper, which is based on the Adam optimization algorithm, offers a simple and feasible solution for satellite fault diagnosis. It can effectively enhance the stability of satellite power control loops and reduce the operational risks of satellites in orbit.