An Automatic Approach for Fault Isolation Based on System Abnormity Probability Model

This paper focuses on the influence of state parameter abnormity to aeroengine performance. Based on multivariate statistical analysis, a probability model describing how the state variable influences aeroengine performance is built to isolate fault in performance monitoring of aeroengine. Using this model, the probability of each state variable leading to performance abnormity in aeroengine is calculated. With this probability, the variables causing abnormity can be preliminarily ascertained, and relevant components can be examined to find the fault source. The practical application in monitoring a certain type of turbofan engine suggests by calculating the probability of each monitored parameter causing abnormity, and with the aeroengine fault mechanism analysis, the fault source can be effectively determined.