惩罚复杂诊断系统混合噪音的模糊小波分类机  

The Fuzzy Wavelet Classifier Machine with Penalizing Hybrid Noises from Complex Diagnosis System

作　　者：吴奇[1,2] 刘静[3] 熊福力[1] 刘晓军[2] 

机构地区：[1]复杂工程系统测量与控制教育部重点实验室(东南大学自动化学院),南京210096 [2]东南大学机械工程学院,南京210096 [3]上海海事大学信息工程学院,上海200135

出　　处：《自动化学报》2009年第6期773-779,共7页Acta Automatica Sinica

基　　金：上海市教育发展基金会晨光计划(2008CG55)资助~~

摘　　要：针对复杂故障诊断系统特征数据中具有高斯、幅值较大、奇异点等混合噪音,设计一种分段式损失函数,构造基于小波基函数的小波核函数,由此得到一种输出为模糊数的模糊小波ν-支持向量分类机,即模糊鲁棒小波ν-支持向量分类机(FRWν-SVC).它可以有效地压制故障特征时序的多种噪音和奇异点,具有很强的鲁棒性,而且它比标准模糊小波ν-支持向量分类机(FWν-SVC)具有更简洁的对偶优化问题.最后进行了汽车装配线故障诊断的实例分析,结果表明基于FRWν-SVC的故障诊断模型是有效可行的.Aiming at Gaussian noise, greater breadth noise, and oddity point noise from feature series of the complex fault diagnosis system, a segment loss function is designed and wavelet kernel function is constructed on the basis of wavelet base function. Then, a fuzzy wavelet v-support vector classifier machine whose outputs are fuzzy numbers is proposed, named as fuzzy robust wavelet v-support vector classifier machine （FRWv-SVC）. FRWv-SVC, which has stronger robustness and simpler dual optimization problem than standard fuzzy wavelet support vector classifier machine （FWv-SVC）, can inhibit some types of noise and oddity point noise of fault feature series effectively. Finally, FRWv-SVC is applied to the fault diagnosis of a car assembly line, the results showing the proposed model based on FRWv-SVC is effective and feasible.

关 键 词：支持向量分类机 小波核函数 分段式损失函数 故障诊断 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]