摘要
针对齿轮箱运作过程中故障诊断效果不理想的问题,提出了一种卷积神经网络(CNN)与粒子群优化的支持向量机(PSO-SVM)方法。首先利用特征参量求出信号的时频特征统计量,其次利用卷积神经网络对时频特征统计量进行二次特征提取,最后利用粒子群优化的支持向量机进行分类。经实验验证,此方法准确率不仅高于其他经典网络模型,而且训练时间最短。
For operation in the process of gear box fault diagnosis problem of the effect is not ideal,this paper proposes a convolutional neural network(CNN)and particle swarm optimization support vector machine(PSO-SVM)method.First by using the time-frequency characteristics of the signal parameter statistics,secondly using convolution neural network pair frequency statistic characteristics secondary feature extraction,finally,particle swarm optimization support vector machine is used for classification.The experimental results show that the accuracy of this method is not only higher than that of other classical network models,but also the shortest training time.
作者
段泽森
郝如江
张晓锋
程旺
夏晗铎
DUAN Zesen;HAO Rujiang;ZHANG Xiaofeng;CHENG Wang;XIA Handuo(School of Mechanical Engineering, Shijiazhuang Tiedao University, Shijiazhuang 050043, China)
出处
《国防交通工程与技术》
2022年第2期21-24,4,共5页
Traffic Engineering and Technology for National Defence
基金
河北省引进留学人员资助项目(CL201721)。
关键词
卷积神经网络
支持向量机
齿轮箱
故障诊断
粒子群优化算法
convolutional neural network
support vector machine
gearbox
fault diagnosis
particle swarm optimization algorithm