摘要
运用小波包频道能量分解技术提取了不同频带反映矿用通风机不同工作状态的特征向量,以此作为支持向量机多故障分类器的故障样本,经训练的分类器作为故障智能分类器可对通风机的工作状态进行自动识别和诊断.并以不对中故障为例,进行了实用验证.研究结果表明,支持向量机在小样本情况下仍能准确、有效地对通风机的工作状态和故障类型进行分类.
Which reflected different working state of ventilator, was extracted from different frequency segment with the technology of wavelet packet frequency segment power decomposition, and taking it as input fault of support vector machine （SVM） multi -fault classifier. The trained classifier, as fault intelligent classification, had very strong identification capability, which could identify automatically the working state of ventilator. And the shaft - misalignment was conducted. The result shows that SVM can classify working condition of ventilator accurately and effectively even in the case of smaller number of samples.
出处
《煤炭学报》
EI
CAS
CSCD
北大核心
2007年第1期98-102,共5页
Journal of China Coal Society
基金
河南省科技攻关项目(0424260115)
关键词
小波包
支持向量机
通风机
故障诊断
wavelet packet
support vector machine
ventilator
fault diagnosis
