Research On On-line Monitoring And Fault Diagnosis Of Large Generator Carbon Brush Based On Multi-source Information Fusion

In the process of electric power production,the carbon brush introduces the excitation current into the rotor and makes the rotor of the generator establish a constant magnetic field,which is the key part of generating electric energy.As a wear part,carbon brush not only requires regular replacement.Moreover,it is often caused by many reasons such as brush ignition,combustion and other failures.When the failure of the carbon brush is serious,the collector ring and the slip ring chamber are even on fire.The generator set will also be forced to stop operation,which will stop the power production process and affect the safety and stability of the power industry.The excitation current of Large Hydrogenerator is very large,once the carbon brush current is distributed unevenly,the above problems are more likely to occur.In view of the carbon brush fault,the on-line monitoring and fault diagnosis of the large-scale hydro generator carbon brush is particularly important.The on-line monitoring of the carbon brush status can not only prevent the serious failure of the carbon brush,resulting in the generator shutdown.Moreover,the monitoring of the historical status information of the carbon brush and the status information of the related equipment of the carbon brush in the monitoring system can provide the basis for the analysis of the expert system.The fault diagnosis of information fusion based on expert knowledge can reduce the analysis time of monitoring data and avoid the delay caused by human analysis data.According to the fault inferred by expert knowledge,it can locate the fault carbon brush and know the cause of the brush fault.Before the carbon brush fault occurs,eliminate the fault to ensure the safe and stable operation of the generator.The online monitoring and fault diagnosis system of carbon brush based on multi-source information fusion described in this paper can not only collect the status of carbon brush and related equipment information,but also analyze and give the results of carbon brush fault.On line monitoring of carbon brush can grasp the operation of carbon brush in real time,avoid the discontinuity of manual measurement of temperature and current of carbon brush,and solve the problems existing in the process of manual inspection of power system.As a comprehensive process of multi-source information fusion,information fusion has its complexity.In recent years,the emergence of artificial intelligence and new methods of information theory has become an important force to promote the development of information fusion technology.In this paper,the brush failureis not only the understanding of the developers,but also the experience of experts in the field.As an artificial intelligence method in multi-source information fusion,expert system has its convenience to realize the failure of experience summary.The fault diagnosis expert system based on information fusion can analyze the real-time and historical data of the carbon brush,locate the fault carbon brush and understand the change of the working condition of the fault carbon brush.The expert system diagnosis results can find out the types and causes of carbon brush faults in time,eliminate the potential faults of carbon brush,and facilitate the operation and maintenance personnel to deal with the brush faults in time.In this paper,the expert system based on database is easy to input expert knowledge,extract fault symptoms and reasoning process of expert knowledge.Avoid the delay and complexity of Reading database data by expert system written by other software.The expert system in this paper adopts forward reasoning and conflict elimination strategies.Using the CF model of imprecise reasoning,the calculation method of fault reliability is given.Through the reliability of fault results,we can judge the probability of fault occurrence and provide basis for fault handling.Finally,the accuracy and feasibility of the system is proved by simulating the operation of the brush fault in the experimental platform.