首页|基于TSFFCNN-PSO-SVM的飞机起落架液压系统故障诊断

基于TSFFCNN-PSO-SVM的飞机起落架液压系统故障诊断

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
针对飞机起落架液压系统故障诊断精度低,深层故障特征提取困难的问题,提出了一种基于双路特征融合卷积神经网络(TSFFCNN)与粒子群优化支持向量机(PSO-SVM)结合的起落架液压系统故障诊断模型.该诊断模型以起落架多节点压力信号作为输入,采用一维卷积神经网络(1DCNN)与二维卷积神经网络(2DCNN)并行多通道网络结构自适应提取深层特征信息,并在融合层将深层特征信息融合,通过优化后的SVM分类器对融合特征进行故障分类.为验证所提诊断模型的故障分类效果,基于AMESim搭建了典型飞机起落架液压系统仿真模型,构建了几种典型故障类型数据集.基于仿真数据的诊断结果表明,所提故障诊断算法精度能达到99.37％,能够有效实现起落架液压系统故障诊断;与其他智能算法对比,基于TSFFCNN-PSO-SVM故障诊断模型具有更好的平稳性与可靠性,诊断精度更高.
Fault diagnosis of aircraft landing gear hydraulic system based on TSFFCNN-PSO-SVM
In view of the problems of low fault diagnosis accuracy of aircraft landing gear hydraulic system and difficulty in extracting deep fault features,a fault diagnosis model of landing gear hydraulic system based on the combination of two-stream feature fusion convolutional neural network(TSFFCNN)and particle swarm optimization support vector machine(PSO-SVM)was proposed.The diagnosis model took the pressure signal of multiple nodes as input,the lD convolutional neural network(1DCNN)and 2D convolutional neural network(2DCNN)parallel multi-channel network structures were adopted to adaptively extract deep feature information,and the deep feature information was fused in the fusion layer.The fusion features were classified into faults through the optimized SVM classifier.In order to verify the proposed fault diagnosis model,a typical aircraftlanding gear hydraulic system simulation model was built based on AMESim,and several typical fault type data sets were constructed.The diagnostic results based on the simulation data showed that the accuracy of the proposed fault diagnosis algorithm can reach 99.37％,which can effectively realize the fault diagnosis of the landing gear hydraulic system;compared with other intelligent algorithms,the fault diagnosis model based on TSFFCNN-PSO-SVM had better stability and reliability,higher diagnosis accuracy.

aircraft landing gear hydraulic systemfeature fusionfault dagnosistwo-stream feature fusion convolutional neural network(TSFFCNN)particle swarm optimization support vector machine(PSO-SVM)

冯东洋、姜春英、鲁墨武、叶长龙、李胜宇

展开 >

沈阳航空航天大学机电工程学院,沈阳 110136

沈阳航空航天大学计算机学院,沈阳 110136

飞机起落架液压系统 特征融合 故障诊断 双路特征融合卷积神经网络(TSFFCNN) 粒子群优化支持向量机(PSO-SVM)

辽宁省自然科学基金

2019-KF-01-11

2024

10.13224/j.cnki.jasp.20220111

航空动力学报
中国航空学会

航空动力学报

CSTPCD北大核心
影响因子：0.59
ISSN：1000-8055
年,卷(期)：2024.39(1)
  • 13