基于PCA BP的航空发动机大修周期预测方法  

Method of PCA BP based Aero Engine Overhaul Makespan Prediction

作　　者：符式峰 贾晓亮[1] 安磊 常笑 FU Shi feng;JIA Xiao liang;AN Lei;CHANG Xiao(Northwestern Polytechnical University,Xi′an 710000,China;Maintenance Group of Equipment Department of Air Force,Beijing 100000,China)

机构地区：[1]西北工业大学,陕西西安710000 [2]空军装备部装备保障大队,北京100000

出　　处：《航空计算技术》2022年第3期62-66,共5页Aeronautical Computing Technique

基　　金：国家自然科学基金项目资助(52075452)。

摘　　要：航空发动机大修具有影响因素多、因素之间耦合关系复杂等特点,针对数据驱动的航空发动机大修周期预测,提出基于主成分分析和反向传播神经网络(PCA BP)的航空发动机大修周期预测方法。在分析影响航空发动机大修周期主要因素的基础上,采用PCA方法得到影响航空发动机大修周期的主成分因素,并将其作为BP神经网络的输入。基于某型航空发动机大修数据对PCA BP模型进行训练和测试,并与BP神经网络模型进行比较,结果表明PCA BP模型预测精度更高,验证了方法的有效性。Aero engine overhaul has complicated influencing factors and complex coupling relationship between factors.Aiming at data driven aero engine overhaul makespan prediction,an aero engine overhaul makespan prediction method and model based on principal component analysis and back propagation neural network(PCA BP)is proposed.Firstly,the key factors affecting aero engine overhaul makespan are analyzed,and the dimension of the influencing factors for aero engine overhaul makespan is reduced by PCA.Then the obtained principal component is taken as the input vectors for BP neural network.Finally,a PCA BP model is trained and tested using a set of aero engine overhaul data and further compared with the BP neural network model.The results show that the PCA BP neural network model has higher prediction accuracy,and the effectiveness of the proposed method is verified.

关 键 词：航空发动机 大修周期 主成分分析 BP神经网络 

分 类 号：V23[航空宇航科学与技术—航空宇航推进理论与工程] O242[理学—计算数学]