摘要
针对利用精细复合多尺度散布熵(RCMDE)特征在行星齿轮裂纹故障识别时存在噪声鲁棒性差、尺度选择依赖人工等不足,通过改进RCMDE特征,提出一种基于自适应精细复合多尺度散布熵(ARCMDE)特征的行星齿轮裂纹故障识别方法。改进从以下3个方面展开:首先在计算RCMDE特征前利用变分模态分解(VMD)算法对信号进行分解,以获取预设数量的固有模态分量(IMF),再计算IMF与原信号的互信息,选取互信息大于阈值的IMF重构信号,实现降噪预处理;然后提出特征重合度及其计算公式,评价多个状态样本的均值标准差之间的重叠交叉情况,利用特征重合度选择较好的若干尺度构建特征向量;最后,结合粒子群优化(PSO)与支持向量机(SVM)实现故障模式识别。行星变速箱实验数据验证结果表明,与多尺度散布熵(MDE)和RCMDE特征相比,采用改进的ARCMDE特征构建的特征向量输入PSO-SVM的分类准确率提高了20%以上,验证了所提裂纹识别方法的有效性和优势。
To overcome the shortcomings of using RCMDE feature in planetary gear crack fault identification,such as poor noise robustness and manual scale selection,a new method for planetary gear crack fault identification based on ARCMDE feature is proposed by improving the RCMDE feature.The improvement is carried out as follows:First,the signal is decomposed by using the variational mode decomposition(VMD)algorithm before calculating RCMDE features,so as to obtain the preset number of intrinsic mode components(IMF);Then,the mutual information between IMF and original signal is calculated,and the reconstructed IMF signal whose mutual information is greater than the threshold value is selected to realize noise reduction preprocessing;Further,the feature coincidence degree and its calculation formula are proposed to evaluate the overlap and crossover between the mean and standard deviation of multiple state samples,and the feature vector is constructed by using the feature coincidence degree.Finally,the fault pattern recognition is realized by combining particle swarm optimization(PSO)and support vector machine(SVM).Experimental results of planetary gearbox and comparisons with multi-scale dispersion entropy(MDE)and RCMDE show that the classification accuracy of the feature vector constructed by the improved ARCMDE is improved by more than 20%,which verifies the effectiveness and advantages of the proposed method.
作者
吴守军
陈健
冯辅周
周超极
吴春志
卫恒
WU Shoujun;CHEN Jian;FENG Fuzhou;ZHOU Chaoji;WU Chunzhi;WEI Heng(Department of Vehicle Engineering, Army Academy of Armored Forces, Beijing 100072, China;Unit 63963 of PLA, Beijing 100072, China;School of Physical Education, Shaanxi Normal University, Xi’an 710119, China;Institute of System Engineering, Academy of Military Sciences, Beijing 100141, China;Unit 32021 of PLA, Beijing 100094, China)
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2021年第6期61-68,共8页
Journal of Xi'an Jiaotong University
基金
国家自然科学基金资助项目(51875575,51875576)。
关键词
行星齿轮
故障诊断
信息熵
支持向量机
多尺度熵
planetary gear
fault diagnosis
information entropy
support vector machine
multiscale entropy
