Abstract: Based on the fuzzy clustering analysis, a momentum wheel health ranking method was proposed. The ground test data and on-orbit data of 50N·m·s momentum wheel in actual model tasks were selected. After the data were processed, the instantaneous dynamic speeds of the momentum wheels in closed loop conditions were studied, then the data were converted into the fuzzy similarity matrix by the correlation coefficient method, and cluster analysis was conducted for the maximum threshold of risks. Meanwhile, the maximum threshold was used as a criterion to design the judgment criteria. The data of the four momentum wheels were used to verify the practicability of the method and compared with the kurtosis test method. The results show that this method can accurately rank the health of momentum wheels.

Key words: fuzzy；clustering, momentum wheel, health, ranking