航空发动机试飞关键参数趋势监控的实现及应用

Realization and Application of Key Parameters Trend Monitoring in Aeroengine Flight Test

为了实现航空发动机滑油压力、滑油温度、振动值在试飞中的趋势监控,采用神经网络方法对某型发动机大量试飞数据进行训练和验证,获得了这几个参数全过程较为准确的计算模型。计算模型应用于该型号另1台发动机参数趋势监控中,在应用前,利用有限架次试飞数据修正了这几个参数的计算模型,采用动态链接库形式实现计算模型与原有实时监控系统的协同工作,进行了模型计算结果和试飞结果趋势实时对比监控。结果表明:模型计算结果和试飞结果变化趋势吻合良好,说明了神经网络计算模型的准确性以及在关键参数趋势监控中的工程实用性。

In order to conduct trend monitoring of aeroengine oil pressure, oil temperature and vibration value in flight test, accurate calculation models of these parameters in the whole engine working process were established by using the neural network method for training and validation of a large number of engine flight test data. Calculation models are corrected based on limited flight test data of another engine before applied in its trend monitoring. For this engine, calculation models were cooperated with real-time monitoring system to realize parameters trend monitoring based on dynamic link library. Real-time monitoring between calculation results and flight-test results was implemented on the engine. Results indicate that trends of calculation results and flight-test results are in good agreement. It shows the accuracy of the neural network models and the engineering practicability in the key parameter trend monitoring.