摘要
旋转机械设备工作机理复杂,故障样本缺乏,难以应用传统的方法对其进行有效的异常状态检测.结合小波包分析技术及人工免疫系统理论,构造了小波包特征免疫检测器,给出了小波包特征免疫检测器的产生算法、异常状态检测方法.小波包特征免疫检测器是在对正常样本学习的基础上产生的,不需要设备运行的故障数据,适合对故障数据缺乏的设备进行有效的异常状态检测.活塞压缩机气阀的异常状态检测结果表明,小波包特征免疫检测器检测效果良好、准确率高.
Traditional method is difficult to be efficiently applied to the abnormal state detection of complex mechanical equipment because of the complex mechanism and insufficient fault samples.Combined with wavelet packet and artificial immune system,the immune detectors of wavelet packet character are constructed.The algorithm of detector production and approach of abnormal state detection are proposed.The immune detectors of wavelet packet character are produced by studying normal samples without the equipment fault data.So the approach proposed is applicable to the detection of the abnormal states of the equipment that lack fault data.As an application example,the abnormal state detection of piston compressor is investigated.Results show that the immune detectors of wavelet packet character can efficiently detect the abnormal states of gas vales of the piston compressor.
出处
《大庆石油学院学报》
CAS
北大核心
2005年第6期101-103,共3页
Journal of Daqing Petroleum Institute
基金
国家自然科学基金资助项目(50475183)
黑龙江省教育厅科学技术研究项目(10541010)
关键词
小渡包
免疫
检测器
反面选择算法
状态检测
wavelet packet
immune
detector
negative selection algorithm
state detection