Study For Modeling Method Of Civil Aircraft Engine Using Performance

Aeroengine performance model has high application value in aeroengine's development,and maintenance. Performance model with good precision and high reliability is also the application foundation to on condition maintenance. Structure and technical details of ACMS and the concept of PHM have been introduced. The application of the engine performance model in PHM system has been discussed.For predicting exhaust gas temperature of aeroengine, a modeling method is performed.Original database was derived from real-time monitoring data of engine monitoring system of PW4000 engine. We took the mean impact value(MIV) as the evaluation criteria for feature selection, screening 8 features as input variables. Extreme learning machine(ELM)algorithm for single-hidden layer feedforward neural network was applied for modeling,and K-fold cross validation was used for selecting the best number of hidden neural. After the neural network has been trained, three extend tests have been taken. Results show that training time of ELM is much less than that of back propagation algorithm, and the hole modeling method has satisfactory extensive performance and affordability.The conclusion is that the modeling method has practicability, with acceptable prediction accuracy, computing speed and extensive characteristic. The method has the value for further research and application.