建筑电气系统故障诊断方法研究  

Research on Fault Diagnosis of Building Electrical System

作　　者：王亚慧[1] 张龙[2] 韩宁[2] 

机构地区：[1]北京建筑工程学院电气与信息工程学院,北京100044 [2]北京林业大学工学院,北京100083

出　　处：《计算机仿真》2014年第2期436-440,共5页Computer Simulation

基　　金：北京市自然科学基金(8111002)

摘　　要：研究建筑电气系统的故障诊断问题。现代化建筑物中,电气故障发生的频率越来越高,针对目前建筑电气系统缺少有效诊断故障方法的问题,同时考虑到在建筑物实际运行环境中典型故障样本数据获取非常有限,因此采用支持向量机(简称SVM)算法,使用建筑电气故障模拟硬件实验平台数据建立了其故障诊断模型,对系统5种故障状态进行诊断分类,仿真结果错判总数为零。最后与人工神经网络诊断方法的对比研究表明,在小样本情况下SVM诊断有效,非线性映射和泛化分类能力更好,更适合于工程实际应用。Study the fault diagnosis of Building Electrical System. In modern buildings, the electrical faults occurred with increasing frequency. To diagnose building electrical system faults effectively, only considering the typical fault samples in the building operation environment data acquisition is very limited. So this paper used the support vector machine（SVM） algorithm, and the fault diagnosis model was established by using the data of building electri- cal fault simulation hardware experimental platform. The system status diagnostic classification was carried out with five kinds of faults on, and it came out that the total misjudged number of simulation results was zero. Compared with artificial neural network diagnosis, the study results show that SVM diagnosis is effective under the condition of small samples, the abilities of nonlinear classification and generalization are better, so it is more suitable for practical application of engineering.

关 键 词：建筑电气系统 故障诊断 支持向量机 

分 类 号：TM743[电气工程—电力系统及自动化]