Transformer Fault Diagnosis Based On Characteristic Gas Analysis In Oil System Research And Application

In power transmission and distribution system,the transformer as the main equipment in the system,has always been concerned about,the credibility and stability of power transformers in the work process directly affect the normal operation of the entire transmission and transformation grid.With the strengthening of China's power grid construction and the level of automation of the power industry,real-time monitoring of the operating state of the transformer,on-line diagnosis of abnormal conditions is particularly important.This work has become the focus of the daily work of power system maintenance.In this paper,based on the study of electrical system and its automation disciplines,combined with the author's work experience in the power industry,selecting the on-line fault diagnosis of power transformers as the main object of study,to study its fault diagnosis method.First of all,the research background,significance and research status of the on-line monitoring of transformer and its fault diagnosis are analyzed and discussed.Then,this paper analyzes the related theories of on-line diagnosis and monitoring of oil-immersed power transformers,such as the basic structure of transformer,the common types of faults,and the existing fault line of transformer.On-line diagnosis and monitoring of oil-immersed power transformers are discussed in this paper.Monitoring and technical means and diagnostic methods,obtained as a result of these commonly used diagnostic method is based on the electrical characteristics of the transformer to its unilateral diagnosis,and based on experience to judge the fault,the internal failure of the power transformer can not be detailed and accurate Then,based on the commonly used method of dissolved gas monitoring in oil,neural network to fault diagnosis is used in this paper,and a comprehensive diagnosis model is established according to the occurrence of specific faults.The system is composed of data acquisition and processing,comprehensive diagnosis,on-line monitoring,query report and system management.The comprehensive diagnosis module is the key part of the design of the subject,and the software of this model is designed.The results show that the fault diagnosis model is superior to the traditional one based on the comparison between the conventional model based on dissolved gas in oil and the neural network based on the characteristic of the fault model.Finally,this paper,combined with application examples,discusses in detail the software application method of the whole system,and gives the research conclusion and the need of improvement.The system has been implemented in Ziyang Electric Power Company,not only for the company in the power transformer maintenance work to provide advanced intelligent methods,making the transformer operation and maintenance management efficiency has been improved,and to achieve the working state of the transformer and electrical performance Real-time monitoring,improve the security of the entire power grid operation.