Research On Fault Diagnosis And Location Of Oil - Immersed Transformer Based On

As the hub of the power system, the transformers’safety closely relate to security and stability of the entire power system. In order to solve the main technical challenges which are encountered in applying oil dissolved gas analysis (Dissolved Gases Analysis, DGA) to diagnose the internal insulation fault of the transformer, by applying the diagnostic model and method in this paper, they can improve the accuracy and reliability of the fault diagnose and the fault point location. Tested by practical examples, it’s proved to be practical and effective. The main works are as follows:(1) According to analyze the generating process on the thermodynamic theory of the hydrocarbon gaseous compounds and other feature gases, it can identify the causes of the feature gases and the external conditions of the feature gases and discover the relationship between the external conditions and the component of the feature gases.(2) To study and master the mechanism of different type of transformer faults, it can figure out the influence that different fault brings to the insulating oil and other insulating materials, such as physical and chemical changes, whether they can form new substances as well as their composition and quantity.(3) Through exploring the intrinsic link between the generating process of the feature gases and the mechanism of the transformer faults, it can determine the relationship between different type of faults and different component of the feature gases, then achieve one-to-one relationship between fault types and the component of the feature gases.(4) To introduce the theory of support vector machine to the fault diagnosis of transformers, on the basis of studying the multi-classification of support vector machine in-depth, this paper builds a fault diagnosis model by analyzing the abnormal phenomena of the transformer and exploring the relationship with insulating oil dissolved gas component when a fault occurs. Practice has proved that it can determine the fault type of transformers effectively and improve work efficiency greatly. (5) By introducing the theory of support vector machine, this paper further analyzes the relationship between the transformer faults and oil dissolved gas component, combining with some related information, this paper introduces the multi-classification algorithm of support vector machine to the fault point location problem to build a transformer fault point location model. By application to the actual examples, it proves that the fault point location model can effectively identify the fault point of the transformers and provide a reference for the research on the fault point location of transformers.