摘要
数据驱动的外啮合齿轮泵(以下简称齿轮泵)故障诊断中,存在实际作业中易受噪声干扰、故障特征冗余以及故障特征选择与分类器参数寻优繁琐问题,为此提出一种基于沙丘猫优化变分模态分解和蜣螂优化算法同步优化特征选择的齿轮泵磨损故障诊断方法。首先,搭建齿轮泵故障试验台获取原始故障数据,采用沙丘猫优化变分模态分解方法对齿轮泵4种磨损故障的振动信号进行降噪重构;然后,提取故障磨损4种重构信号的时域、频域和时频域统计特征共26种,并组成特征层;最后,基于蜣螂优化算法同步优化特征选择对故障特征集进行特征选择,同时优化支持向量机分类器参数,实现齿轮泵的磨损故障类型识别。结果显示,该齿轮泵故障诊断方法准确率高达99.6%,耗时仅49.8 s,具有较高的诊断精度和运算效率。
Aiming at the fault diagnosis of data-driven external gear pump(hereinafter referred to as gear pump),there are the problems of easy noise interference in actual operation,redundant fault features and cumbersome fault feature selection and classifier parameter optimization,so we propose a wear fault diagnosis method for gear pump based on the simultaneous optimal feature selection with sand cat swarm optimization-variational mode decomposition and dung beetle optimization algorithm.Firstly,the original fault data are obtained by building a gear pump fault experimental bench,and the vibration signals of four types of gear pump wear faults are reconstructed by noise reduction using the method of sand cat swarm optimization-variational mode decomposition.Then,a total of 26 statistical features are extracted from the four reconstructed signals in the time domain,the frequency domain,and the time-frequency domain,and the feature layers are composed.Finally,the feature selection of the fault feature set based on the dung beetle optimization algorithm and the optimization of parameters of classifier for support vector machine are carried out simultaneously,achieving wear fault type recognition of gear pumps.The results show that the accuracy rate of the fault diagnosis method for gear pump is as high as 99.6%,and the consumed time is only 49.8 s,which demonstrates the high diagnostic accuracy and computational efficiency of the proposed method.
作者
问亚鹏
张佳奇
郭锐
杨锦昌
何丝丝
张浩
WEN Yapeng;ZHANG Jiaqi;GUO Rui;YANG Jinchang;HE Sisi;ZHANG Hao(AECC Sichuan Gas Turbine Establishment,Mianyang,Sichuan 621000;School of Mechanical Engineering,Yanshan University,Qinhuangdao,Hebei 066004;State Key Laboratory of Crane Technology,Qinhuangdao,Hebei 066004;Hebei Key Laboratory of Special Carrier Equipment,Qinhuangdao,Hebei 066004;Key Laboratory of Advanced Forging&Stamping Technology and Science,Ministry of Education of China,Qinhuangdao,Hebei 066004)
出处
《液压与气动》
北大核心
2025年第8期65-78,共14页
Chinese Hydraulics & Pneumatics
基金
国家自然科学基金(52075469,12173054)。
关键词
齿轮泵
故障诊断
同步优化特征选择
蜣螂优化算法
沙丘猫优化变分模态分解
gear pump
fault diagnosis
simultaneous optimal feature selection
dung beetle optimization algorithm
sand cat swarm optimization-variational mode decomposition
