主元分析法和模糊积分的航空发动机气路状态监测

摘要

The aeroengine is a large-scale system,because of its complex structure,poor working conditions and other factors,the effective health condition monitoring has become one of the key techniques in the aviation field that is difficult to resolve. In order to monitor the health condition of aeroengine effectively,taking the aeroengine gas path system as an example,this paper puts forward a condition monitoring method of aeroengine based on principal component analysis and fuzzy integral. First of all,using the method of principal component analysis to extract the main element and feature information and reduce the dimension of the sample data,it achieves optimal compression samples. Secondly,based on the feature vector of the engine condition sample data,it uses Bp and Elman neural network to monitor the condition of aeroengine. Finally,using two kinds of neural network results for decision level fusion based on fuzzy integral. By a certain type of aircraft engine real validation,it showed that this method took advantage of principal component analysis and fuzzy integral,improved the accuracy of condition monitoring and met the aircraft engine condition monitoring of real-time requirements,which has a good value in engineering applications.%航空发动机是一个大系统，由于结构复杂、工作条件恶劣等因素影响，对其进行有效地健康状态监测成为航空领域长期难以解决的关键技术之一。为有效监测航空发动机健康状态，以航空发动机气路系统为例，提出一种基于主元分析和模糊积分的航空发动机状态监测方法。首先，利用主元分析法提取发动机状态样本集的主元，对样本数据进行降维，实现样本的最优压缩。其次，利用BP神经网络和Elman神经网络对发动机状态信息的特征向量进行初步状态监测。最后，利用模糊积分对采用两种神经网络的初步监测结果进行决策层融合，从而有效地实现对航空发动机气路系统的状态监测。通过某型真实航空发动机验证表明，所提出基于主元分析和模糊积分的状态监测方法，能有效提高监测的准确度，满足航空发动机状态监测的实时性要求，具有良好的工程应用价值。