Based on the completeness of the normal and fault condition data of liquid rocket engine and the improvement of data quality

a data-driven comprehensive fuzzy clustering algorithm was proposed for fault diagnosis. The fuzzy c-means (FCM) algorithm was used to cluster the known normal sample data to obtain the optimal clustering center

and the obtained cluster center was used as a-priori sample data to select the optimal classification result of the closure method to obtain the fault detection result. Only a small amount of normal prior sample data were required to quickly and accurately detect the fault; then the FCM algorithm was used to classify the fault

and the corresponding fault type can be clustered according to the existing fault database

and the range of fault amplitudes can be given. The simulation results showed that the detection rate of the algorithm was up to 968% and the fault isolation rate was 94%. The actual test data of a liquid rocket engine show that the fault diagnosis algorithm can detect and isolate faults accurately and timely.