Adaptive thresholds algorithm for fault detection of liquid rocket engine in ground test

Presented an adaptive thresholds algorithm (ATA) for fault detection of liquid rocket engine (LRE), the parameter thresholds of which were calculated from real-time estimated mean values, standard deviations and fixed band-width factors trained by test data using the training algorithm also introduced. The performance of ATA is higher than both the System for Anomaly and Failure Detection (SAFD) algorithm and the Redline System (RS) algorithm, indicated by off-line detection results of test data measured in twenty-two ground tests of a large LRE.