基于模型的发动机喘振故障关键参量仿真与预测

Model Based Simulation and Prediction of Some Key Parameters of Surge Fault in Engine

航空发动机故障中的深喘严重破坏发动机运行稳定乃至飞机安全。针对该问题,分析了表征喘振故障关键参量的异常变化特征,基于航空发动机运行过程气动热力机理所构建的发动机整体仿真模型模拟的表征发生喘振故障的关键参量变化情况,利用长短记忆神经网络模型进行数据变化趋势预测,通过与其他类型的不同网络算法预报结果的比较,验证了LSTM模型用于喘振故障关键参量预测的有效性。

Surge fault surge will seriously damage the stability of the engine operation and even the safety of the aircraft,resulting in serious consequences.The abnormal variation characteristics of key parameters characterizing surge fault were analyzed.Based on the aerothermodynamic mechanism of aeroengine in operation,the whole simulation model of aeroengine was established,and the change of key parameters of surge fault was simulated.The LSTM neural network model was used to predict the trend of data changes.By comparing with the prediction results of other different network algorithms,the validity of LSTM model in predicting the critical parameters of surge fault was verified.