本文建立了一个基于知识的汽轮发电机组在线工况监测与故障诊断系统KBTMD。该系统将传统的监测与诊断技术同智能诊断技术进行了有机的结合,采用多种诊断信息获取方式和灵活的推理与控制策略,只有良好的知识库开发与管理功能和友好的人机接口。在知识表达上采用了“框架+规则+过程”的复合表达技术,在知识组织上采用了层次组织模糊。这一系统还为开发其它旋转机械的智能诊断系统提供了一个良好的开发工具。 In this paper, a knowledge-based on-line condition monitoring and fault diagnosis system KBTMD for steam turbine generator set up. The system combines the traditional monitoring and diagnosis technology with intelligent diagnosis technology, adopts a variety of diagnostic information acquisition methods and flexible reasoning and control strategies, and has only good knowledge base development and management functions and friendly man-machine interface. In the expression of knowledge, the composite expression technique of “framework + rules + process” is adopted, and the hierarchical organization fuzzy is adopted in the knowledge organization. This system also provides a good development tool for developing intelligent diagnostic systems for other rotating machines.