Study On Synchronous Generator Excitation Device Hierarchical Fault Diagnosis

As an important part of synchronous generator excitation device,to maintain stable voltage of the power pattern,reasonable distribution of reactive power,and improve power system stability.Its performance determines the operation of generator's stability,at the same time it also affect the power system safe,stable and economic operation.When the excitation device appear failure,if you can't take timely measures,to eliminate the fault,will lead to the generator outputing voltage instabilitly,power quality decline,even stop,bring huge economic loss to power plants.This article through a large number of literature research,analyzing and summarizing the current research status of the excitation device fault diagnosis,putting forward the hierarchical classification of excitation device fault diagnosis,the simple fault directly diagnosed in the ground level' excitation controller,complex fault in plant-level is diagnosed using multi-source information fusion algorithm,real-time and reliability is all guaranteed.In this paper,it design the overall architecture for the excitation device hierarchical fault diagnosis system,the architecture based on hydropower station unified information platform,it is an integral part of a unified information platform,divided into local diagnosis,plant-level diagnosis and remote diagnosis.In view of the excitation device'regional online monitoring and fault diagnosis,using DSP as main control chip,use of traditional diagnosis methods,it can diagnose simple fault and display the alarming information,analyzing the cause of the problem of some typical faults,and according to the degree of diagnosis for fault classification,studing the simple fault diagnosis methods,including spectrum and working condition analysis.Plant-level diagnosis for excitation device' complex fault,this paper,by using the comprehensive diagnosis of multi-source information fusion algorithm,and In the case of excitation-loss fault.First of all,modeling and simulation for the excitation system,through to the part of the excitation-loss and all excitation-loss fault simulation,the dynamic process of excitation-loss is analyzed.By recording fault waveform,analysing of the excitation-loss variation characteristics of various electric parameters.By the simulation of excitation-loss fault characteristics,fault diagnosis based on BP neural network model is established,extracting some failure samples for BP neural network structure and training.The trained BP network diagnose some fault case information,through Matlab simulation test,show that based on the the multi-source information fusion complex fault diagnosis is of high accuracy.Based on the above research,using the QT software develop unified information platform system of hydroelectric power station,to realize the human-computer interaction interface of fault diagnosis.