Adaptive Segmentation and Feature Acquisition of Test Sequence for Momentum Wheel

In aerospace engineering, the telemetry sequence is a vital reference for evaluating the performance of each component. The segmentation algorithm is an effective method for processing telemetry sequence, which lays a solid foundation for obtaining feature indicators. For the Momentum Wheel, rotational speed stability and acceleration characteristic are two crucial feature indicators. This paper presents a novel method of Adaptive Segmentation and Feature Extraction (ASFE) for processing the rotational speed sequence of the Momentum Wheel, which can realize the requirements of online testing and automatic acquisition of these two indicators. Firstly, based on the two different working modes of constant speed and adjusting speed, the rotational speed sequence of Momentum Wheel is identified as smooth and transitional. Secondly, the improved Adaptive Piecewise Constant Approximation (APCA) algorithm is proposed to identify the smooth segmentation, and the feature extraction sub-algorithm with online processing capability is also proposed to obtain the indicator of rotational speed stability. Thirdly, the fitting method based on slope threshold is used for the data of the transition segmentation, and the indicator of acceleration characteristic is extracted. The experiment results with the X Momentum Wheel shows that the accuracy of the data measurement processing is not less than 0.96 of the theoretical true value. This indicates that ASFE can significantly improve the test efficiency and meet the requirements of the Momentum Wheel test.