基于CBR循环发电机组故障诊断系统研究

Research into Cycle Generating Unit Fault-diagnosis System Based on CBR Technology

针对某钢铁公司循环发电机组的故障诊断问题,应用CBR技术进行故障诊断系统的设计。提出基于VSM的相似度分类器策略完成CBR关键技术的设计。通过计算特征项权重用向量表示机组故障案例,采用余弦法计算案例间的相似度,比较输出相似度大于阈值的案例。最终在asp.net平台上开发了基于Web的故障诊断系统。案例分析结果表明,基于CBR故障诊断系统能提高新故障维修解决的准确性和高效性。

In view of the fault diagnosis problem of cycle generating unit in a steel company, a fault diagnosis system based on CBR technology is designed. VSM （vector space model） similarity classifier is adopted to complete the design of the key technology of CBR. Fault cases are represented by vectors through calculating the weight of fea- ture items; the similarity between the cases is calculated by cosine method; and the cases whose similarity is greater than the threshold value are output. Finally, the fault diagnosis system is developed based on WEB on ASP. NET platform. Case analysis results show that the proposed fault diagnosis system based on CBR can improve the accuracy and efficiency.