发电机定子温度在线监测与自动预警系统设计

On-line temperature monitoring and automatic warning system for generator stator

为实现对发电机运行过程中定子部分过热故障的及早发现和准确预警目标,对发电机定子部分的温度在线监测方法进行了研究并设计了相应的自动预警系统。采用基于运行参数状态空间辨识的指纹算法得到各监测点的标准参照值,运用阈值判据获取状态的定量信息;综合各监测点的趋势特征信息,提出了定量和定性相结合的混合策略,并运用多征兆集成推理,与知识库中专家规则匹配后得出诊断结论。最后,阐述了系统数据库规划、运行界面以及信息管理等环节的设计实现。该系统有效提高了机组定子温度监测与预警的准确率。

The on-line stator temperature monitoring is researched and an automatic warning system is designed to early and accurately detect the overheating of stator during generator operation. Fingerprint algorithm based on the state space identification of operating parameters is adopted to calculate standard reference temperature of each detector, and the quantitative information is obtained by means of threshold-criterion. The tendency features of all detectors are synthesized and the hybrid qualitative and quantitative strategy is presented,combined with multi-symptom integrating reasoning, to obtain the diagnostic conclusion by matching corresponding rules in expert knowledge database. Design and implementation of the database, monitoring display and information management are expounded. The system improves the accuracy of stator temperature monitoring and warning.