In order to improve the effectiveness of fault diagnosis of aero-engine lubricating oil system,this paper studies the fault diagnosis technology of aero-engine and proposes an improved grey Wolf algorithm to optimize the nuclear limit learning ma-chine.Firstly,the parameter data of aeroengine lubricating oil system is preprocessed,and the kernel space is mapped by kernel independent component analysis to eliminate the correlation between the original feature vectors and extract the feature coefficient matrix.Secondly,the KELM fault model was created from the extracted feature matrix.In order to reduce the impact of random-ness of network parameters artificially adjusted on the diagnosis results,IGWO algorithm was used to optimize the network pa-rameters of KELM and create the fault diagnosis model.Finally,the IGWO-KELM fault diagnosis model is verified by experi-ments.The results show that the proposed IGWO optimized KELM fault diagnosis method can effectively improve the accuracy of aero-engine fault diagnosis.The diagnostic accuracy reaches 96%,which has a good application prospect.
NETL
NSTL
万方数据
维普
