Fault Prediction of Control Moment Gyro with Lempel-Ziv Index Based on Wavelet Entropy

In spacecrafts, the control moment gyro is the core actuator for attitude adjustment and control. Its failure causes the malfunction of the spacecraft attitude control system, which affects the safety and reliability of the spacecraft in orbit. To realize the abnormal feature mining and early fault prediction of the on-orbit working telemetry data of the control moment gyro, this study proposes a failure prediction method of the control moment gyro based on wavelet entropy and Lempel-Ziv index. First, wavelet multi-resolution analysis was performed according to the discrete time of the on-orbit data and signals on each data acquisition day. Wavelet entropy was used to extract the fault characteristics in the data. However, wavelet entropy was easily affected by the length of data. To overcome this limitation, the Lempel-Ziv index was used to improve the proposed method. The results show that this method can not only predict but also locate the failure of the control moment gyro.