【作者】 胡清华； 于霄； 于达仁； 徐志强；

【Author】 HU Qing-hua, YU Xiao, YU Da-ren, XU Zhi-qiang (Harbin Institute of Technology, Harbin, 150001, China)

【机构】 哈尔滨工业大学能源学院；

【摘要】 汽轮机轴系故障诊断是困扰电厂运行的主要问题,当前开发了一些专家系统来帮助运行和人类专家分析和诊断轴系故障。运行中发现专家系统应用于轴系故障诊断时出现了置信度低、各故障差异小等问题。本文提出采用D-S证据融合理论和模糊综合判别方法融合多次专家系统多次诊断的结果,通过多个故障诊断实例说明了该方法的有效性。更多还原

【Abstract】 Faults in rotor systems have been great problems in operating and managing power plants. Some expert systems are developed to help operators and experts to analyze and diagnose faults in rotor systems. There are some drawbacks of the existing expert systems, such as the lower belief and multiple faults without great difference in belief. We propose a method to improve the robustness and accuracy based on combining multiple results from experts systems in this paper. D-S evidence reasoning and fuzzy comprehensive evaluation are introduced to fuse the multiple results from different times. The test shows the proposed methods are effective.更多还原