Research On Fault Feature Extraction And Severity Identification Of Power Plant Thermal System

Fault diagnosis for the power plant thermal system is of great significance to improve the safety,stability and economy of the whole unit.The structure of the thermal system is complex and the parameters are highly coupled.Once the fault occurs,it will not only affect the operating efficiency of the unit,but also cause the unit to shut down when it is serious.To improve the safety and reliability of the thermal system,prevent or eliminate the fault is an urgent problem to be solved.Qualitative trend analysis(QTA)is an important technique for fault detection and diagnosis,the characteristic trend of the process variable is an important form of the running state,which provides the basis for the early judgement of the fault.Therefore,it is of practical significance to carry out the fault diagnosis of the thermal system based on QTA.The fault diagnosis method based on qualitative trend analysis mainly includes the trend extraction and recognition algorithm and the fuzzy matching strategy.Linear least-squares fitting technique is used to extract the dynamic trend of the data,and the cumulus sum technique is used to judge whether the fitting trend can be accepted.In order to improve the adaptability of the algorithm,a method of setting the grading threshold is proposed.In the stage of fault diagnosis,a typical fault sample rule knowledge base is established,the similarity of feature parameters is calculated by fuzzy trend matching strategy,and the quantitative information of feature parameters is combined to achieve the purpose of fault severity identification.Taking the feedwater heater system of a 600 MW supercritical unit as the object investigated,fault diagnosis study is carried out with its full-scope simulator.Through the simulation research on the typical faults of the feedwter heater system,the grading thresholds of different characteristic parameters are selected.The trends of typical leakage fault samples are extracted and the knowledge base of fault diagnosis rules is established.The fault disgnosis program based on fuzzy matching strategy is developed with MATLAB software and detailed fault diagnosis simulation experiments are carried out.The results show that the method proposed can be applied to diagnose the faults of different severity in the thermal system and can achieve good diagnostic results.