摘要
齿轮箱作为航空发动机重要的传动装置,工作环境恶劣,导致振动信号呈多种信息叠加难以区分。针对齿轮箱故障特征难以提取、故障难以识别的问题,提出一种基于一维卷积神经网络结合改进鲸鱼优化支持向量机的航空发动机齿轮箱故障诊断方法,实现航空发动机齿轮箱故障快速、精准诊断。使用一维卷积神经通过其内置的卷积和池化对振动信号进行故障特征提取,在鲸鱼优化算法中引入混沌映射、非线性因子和自适应权重对其进行改进;使用改进后的鲸鱼优化算法对支持向量机进行参数寻优,再将一维卷积神经网络提取的故障特征输入到经改进鲸鱼优化参数后的支持向量机中进行故障诊断。仿真结果表明:所提的故障诊断模型对齿轮箱故障具有良好的诊断效果,与其他方法相比效果更好、泛化能力更强。
As an important transmission device of the aero-engine,the gearbox works under harsh environmental conditions;as a result,vibration signals are difficult to distinguish from multiple information superimpositions.Since it is difficult to extract the gearbox’s fault features and identify the faults,a fault diagnosis method for the aero-engine gearbox is proposed,based on a one-dimensional convolutional neural network combined with an improved whale optimization support vector machine,in order to achieve fast and accurate fault diagnosis of the aero-engine gearbox.The one-dimensional convolutional neural network is used to extract the fault features from the vibration signals through its built-in convolution and pooling;the whale optimization algorithm is improved by introducing chaotic mapping,nonlinear factors,and adaptive weights;the improved whale optimization algorithm is used to conduct the parameter optimization on the support vector machine;then,the fault features extracted by the one-dimensional convolutional neural network are inputted into the support vector machine with the improved whale optimization parameters,so as to carry out the fault diagnosis.The simulation results show that compared with other methods,the fault diagnosis model has a higher standard of diagnosis on gearbox faults,with better effect and stronger generalization ability.
作者
贾丽臻
雷欣然
李耀华
JIA Lizhen;LEI Xinran;LI Yaohua(Transportation Science and Engineering Institute,Civil Aviation Flight University of China,Tianjin 300300;Aviation Engineering Institute,Civil Aviation Flight University of China,Tianjin 300300)
出处
《机械设计》
北大核心
2025年第7期98-106,共9页
Journal of Machine Design
基金
中央高校基本科研业务费(3122022052)。
关键词
齿轮箱
故障诊断
一维卷积神经网络
改进鲸鱼优化算法
支持向量机
gearbox
fault diagnosis
one-dimensional convolutional neural network
improved whale optimization algorithm
support vector machine
作者简介
贾丽臻(1989—),女,讲师,博士,研究方向:产品创新设计方法及理论。E-mail:jializhen_1314@163.com;通信作者:雷欣然(1999—),男,硕士研究生,研究方向:故障诊断。E-mail:lei479364581@163.com。
