基于门控循环单元的涡扇发动机剩余寿命预测

Remaining Useful Life Prediction of Turbofan Engine Based on Gated Cycle Units

准确预测飞机发动机的剩余使用寿命,既便于采用及时合理的维修措施,降低维护的经济成本,又提高了飞机的安全性。提出了一种基于门控循环单元的剩余使用寿命评估模型。从大量训练集中采集数据的时序特征,筛选后将合适的特征数据进行门控循环单元网络模型训练,生成预测模型。该预测模型对测试集进行剩余使用寿命评估效果良好,验证了方法的可用性。该方法有效监测航空发动机的动态健康状态,为基于深度学习方法建立剩余寿命评估模型提供了新途径。

Accurately predicting the remaining service life of aircraft engine is not only convenient to adopt timely and reasonable maintenance measures,reduce the economic cost of maintenance,but also improve the safety of aircraft.In this paper,a residual service life evaluation model based on gated cycle unit is proposed.The time series characteristics of data collected from a large number of training sets are screened,and the appropriate characteristic data are trained in the gated cycle unit network model to generate the prediction model.The prediction model has a good effect on the remaining service life evaluation of the test set,which verifies the availability of the method.This method can effectively monitor the dynamic health state of aeroengine,and provides a new way to establish the residual life evaluation model based on deep learning method.