论文部分内容阅读
为了兼顾数据建模的准确性和诊断的实时性,提出一种K近邻诊断证据融合新方法.利用故障特征的历史样本构建随机模糊变量(RFV)形式的故障样板模式,由KNN算法获取测试样本的K个近邻历史样本,并定义它们的RFV待检模式;经样板和待检模式的匹配获取K个诊断证据,再将各特征的K个诊断证据融合,并作出故障决策;使用RFV实现对故障数据的精准建模,利用K个历史样本丰富诊断信息,并增加诊断的时效性.诊断效果在电机转子试验台上得到了验证.
In order to take into account the accuracy of data modeling and the real-time diagnosis, a new fusion method of K-nearest neighbor diagnostic evidence is proposed.Fault samples patterns in the form of random fuzzy variables (RFV) are constructed by historical samples of fault features, and the test samples are obtained by KNN algorithm , And define their RFV test patterns. K diagnostic evidences are obtained through the matching between the template and the test pattern, and then the K diagnostic evidences of each feature are fused to make fault decision; Accurate modeling of fault data, enriching diagnostic information with K historical samples, and increasing timeliness of diagnosis Diagnostic results were validated on a motor rotor test bench.