“仿医疗”的水电机组故障诊断系统

Fault diagnosis system of hydropower unit based on imitating medical diagnosis

水轮发电机组的故障诊断过程和人类求医问药的医疗诊断过程有很强的相似性。本文从论述两者相似性入手,深入分析了水电设备故障诊断的难点和现状,指出了对智能故障诊断的两个错误认识:一是将智能诊断理想化、绝对化,拒绝技术工作人员的创造性参与;二是将智能诊断系统建立在能自动采集的状态监测、计算机监控系统等基础之上,忽略了诊断型试验、处方型试验、日常运行记录等一些不能自动获取数据、资料的辅助诊断手段。文章提出的"仿医疗"水电机组故障诊断系统是一全面借鉴医疗诊断过程和原理的交互型智能系统,可在状态监测、工作日志等过去资料汇总分析基础上做出故障类型划分和初步诊断,并根据需要提出辅助于故障判断的诊断型试验步骤及方案,协助确诊。该系统还包含辅助于确定故障处理或检修方案的处方型试验,借此获得最佳解决方案。文章以机械、电磁和水力等三种不稳定的区分、水轮机转轮叶片裂纹原因分析及机组动态不平衡处理等为例介绍了"仿医疗"智能诊断系统概略架构设想。

Fault diagnosis of hydropower unit is similar to medical diagnosis of humans. This paper analyzes the difficulties and state-of-art of the former from the similarity and demonstrates two misconceptions about intelligent fault diagnosis. One is that intelligent fault diagnosis is thought to be ideal and absolute, and not enough attention is paid to creative participation of technical stuff. The other is that intelligent fault diagnosis is thought to be an easy process, and some auxiliary diagnostic methods are ignored, such as diagnostic-type test and prescription-type test. The fault diagnosis of hydropower unit presented in this paper is an interactive intelligent system which imitates medical diagnosis process of humans. With this system, the types of faults could be classified and their preliminary diagnosis could be made by imitating medical diagnosis according to the data of condition monitoring, working records and so on. Auxiliary test for diagnosis procedures and programs could be designed to aid determination of faults if necessary. The system also includes prescription-type tests which could aid us in determination of the treatment, maintenance and repair schemes of faults for best solutions. To illustrate its application and medical diagnosis imitation, the paper gives three examples, i.e. classification of the instability faults caused by mechanical, electromagnetic and hydraulic factors, reasons of the cracks on turbine runner blades, and treatments of the dynamic unbalance of hydropower units.