Insulation Diagnosis of XLPE Cables by Expert System and Its Application

本研究在1993年和1994年两年间开发并实际应用了6.6kV交联聚乙烯电缆绝缘退化诊断专家系统。在该专家系统中，计算机应用程序主程序用c语言编写，主推理程序采用OPS83的后向推理规则构造。此外，采用模糊推理方法对绝缘特性值进行推理。在本研究的开始，利用从停电试验中获得的一些绝缘特性值，建立了一个原型专家系统。专家系统取得了令人瞩目的成绩，系统的诊断结果与专家的诊断结果基本一致。但是，由于不允许长时间停止供电，因此不能将停电试验作为一种有利的方法来坚持。此外，仅分析绝缘特征值对监测驾驶室的劣化趋势也没有好处。为了纠正这些缺陷，专家系统增加了从目测测试中获得的绝缘退化数据知识库。通过听取专家的意见，构建知识库。对专家系统的计算机程序进行了修改，使其适用于交联聚乙烯电缆的带电状态。在这种情况下，通过分析火线试验获得的绝缘电阻值和护套绝缘电阻值进行诊断。这些改进提高了专家系统的诊断精度。由于制造形式、安装环境、安装年限等条件不同，绝缘退化的诊断标准也有所改变，在专家系统中反映了这些条件。在系统评价中发现，系统的结果与专家的结果有85%的一致性。此外，专家系统还得到了改进，绝缘特性值由带电绝缘诊断器通过GP-IB电缆自动输入。采用了藤仓公司的带电线绝缘诊断系统，系统的改进使我们能够随时监测绝缘劣化的趋势并进行诊断。本研究实现了性能优良的专家系统，并将该系统实际应用于某工厂生产的6条交联聚乙烯电缆。少

This study has developed and practical applied an expert system for diagnosing insulation degradation of 6.6kV XLPE cables, during 2 years for 1993 and 1994. In this expert system, the main program of computer applications is written in C-language, the main inference routine is constructed by a backward inference rule of OPS83. In addition, the inference with insulation characteristic value is done by a fuzzy inference method. At the beginning of this study, a prototype expert system has been performed by using some insulation characteristic value cbtained from a power outage test. The expert system has made remarkable achievements that the diagnosis results of system almost agreed with those of the experts. However, since it is not allowed to stop the power supply for a long time, the power outage test can not be insisted as an advantageous method. And also, the insulation diagnosis analyzing only insulation characteristic value is no advantage for monitoring degradation trends of cab … More le insulation. In order to correct these defects, the expert system has added a knowledge base of insulation degradation data obtained from visual inspection tests. The knowledge base has been constructed by hearing of the experts. The computer program of expert system has been modified for applying to a live-wire state of XLPE cables. In this case, the diagnosis is carried out by analyzing insulation resistance value and sheath insulation resistance value obtained from the live-wire test. The modifications have improved the diagnosis accuracy of the expert system. Since the insulation degradation differs in some conditions such as manufacturing form, installation environment and installation years, the criteria of diagnosis has reformed by reflecting the conditions in the expert system. In the system evaluation, it has been found that both results of the system and the experts coincide with each other in 85%.Furthermore, the expert system has been improved as the insulation characteristic value is automatically inputted by a live-wire insulation diagnoser via a GP-IB cable. The live-wire insulation diagnoser has been used through the kind offer from Fujikura LTD.The improvement of system enable us to always monitor the trends and diagnose the insulation degradation. This study achieves the expert system which has excellent performance, and has practical applied the system to 6 lines of the XLPE cables operating in a factory. Less