The Fuzzy Wavelet Classifier Machine with Penalizing Hybrid Noises from Complex Diagnosis System
-
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.Abstract: Aiming at Gaussian noise, greater breadth noise, andoddity point noise from feature series of the complex faultdiagnosis system, a segment loss function is designed and waveletkernel function is constructed on the basis of wavelet basefunction. Then, a fuzzy wavelet ν-support vector classifiermachine whose outputs are fuzzy numbers is proposed, named as fuzzyrobust wavelet ν-support vector classifier machine(FRWν-SVC). FRWν-SVC, which has stronger robustness andsimpler dual optimization problem than standard fuzzy waveletsupport vector classifier machine (FWν-SVC), can inhibit sometypes of noise and oddity point noise of fault feature serieseffectively. Finally, FRWν-SVC is applied to the fault diagnosisof a car assembly line, the results showing the proposed model based on FRWν-SVC is ffective and feasible.
-
计量
- 文章访问数: 1890
- HTML全文浏览量: 58
- PDF下载量: 1207
- 被引次数: 0
下载:
