Study on engine's rub-impact failure detection model based on flight data

During the status of real-time flight monitoring, it is related to the safety of flight that the accuracy of engine's rub-impact failure detection under the strong vibration environments. Under the strong vibration environment, when engine goes to stationary state, the engine's rub-impact failure is easy to happen, and they are often happened simultaneously. However, there is no correlation between the vibration signal failures. Traditional failure mining methods in the engine's rub-impact failure detection process, it only can detect this failure the vibration signal, so that failure detection is not accurate. In order to solve this problem, the paper proposed an engine's rub-impact failure detection method based on improved association rule mining. The associated clustering process are made with the data of engine's rub-impact failure, to obtain the classification matrix of sample space and update the data of engine's rub-impact failure. According to updated results, the association probability values of the data of engine's rub-impact failure were calculated to obtain probabilistic decision. The association rules mining was made with data of aircraft's rub-impact failure, so as to realize the optimized detection for the data of engine's rub-impact failure. The experiment results show that using the improved association rules algorithm to detect the engine's rotor-to station rub-impact failure under strong vibration environment, can greatly shorten the detection time, reduce the missing rate, and improve the detection accuracy.