Analysis of Physical Condition of Aero-Engines Based on Flight Data

Through the analysis of typical aero-engine fault, the function parameters under different throttle openings of the aero-engine are presented, and the fault characteristic learning as the training set are studied using the RBF neural network. An aero-engine function evaluation model is proposed in this work by comparing the test set with the expected value of the training set, and the decay degree of engine function are determined. The proposed method is validated to be an effective method of diagnosing and identifying the aero-engine fault accurately and timely by testing large amount of recorded flight data from each of various types of aero-engines. It makes it possible for early and successful diagnosis and prediction of the health condition of aero-engines.