Transformer Fault Diagnosis And Status Assessment Based On Neural Network And Support Vector Machin

Due to the increase in the size of the equipment and the increased demand for reliability of electrical energy by the users,the proportion of electrical energy costs in the repair costs of electrical facilities will also increase,and the long-standing overhaul of the electrical system equipment is mostly in the form of hourly fixed repairs.The significance of maintenance and management of electrical equipment is becoming increasingly apparent.At present,transformer operation inspection as the basic content of power transformer operation and maintenance is increasingly becoming the basis for correct maintenance strategies and the development of reasonable maintenance programmes.In this project,in-depth research has been carried out on two aspects of the insulation oil paper electrothermal ageing test of power transformers and the condition assessment and fault diagnosis in power transformer fault diagnosis technology,analyzing the insulation oil paper ageing law and proposing an intelligent assessment and fault diagnosis method based on support vector classification machine theory,conducting the following studies.(1)Simulating the working conditions of insulating oil paper in transformers and comprehensively analysing the aging law of insulating oil paper from different perspectives.(2)The traditional artificial neural network is improved to study the fault diagnosis method of large transformers based on the dissolved gas in oil analysis method,which improves the accuracy and reliability of fault diagnosis.(3)Based on the preventive measurement results,the internal inspection data and external test results of transformers were taken into consideration,and the condition data that can more fully and accurately reflect the operation of transformers were selected from the condition data required for transformer performance evaluation,starting from the perspectives of chemical substances in oil,electrical tests,insulating oil,or the influence of other causes,and after the qualitative and quantitative condition quantities of transformers were A comprehensive study has been carried out to identify the main elements of transformer condition evaluation and to develop a comprehensive approach to evaluating the condition of transformers and their insulation bodies.The method enables the condition of the equipment to be assessed on the basis of its actual current operating state,using advanced condition testing methods,reliability assessment methods and life cycle prediction methods.The location of problems,the degree of damage and trends are assessed,early signs of problems are identified and,based on the findings of the assessment and diagnosis,repairs are carried out before the performance of the equipment is reduced to a specified level or problems are imminent.The timely identification of problems in the system during maintenance plays an important role in ensuring the economic operation of the equipment,overcoming the problems of "over-maintenance" and "under-maintenance" in conventional maintenance,and improving equipment reliability.