| Relevant statistics show that the highest possibility of power transformer failures is caused by windings. It is of great significance to develop monitoring and diagnosis techniques for power transformer windings. Because the destruction of the mechanical integrity of transformer windings, such as loosening or deformation, inevitably lead to the change of the mechanism dynamic performance of transformer body, the study of transformer vibration characteristics becomes an effective way.In this paper, the transformer winding state recognition method is studied. This paper aims to study the method based on vibration analysis, and to achieve online condition monitoring of transformer windings. The paper defines a multi-point vibration feature vectors and matrices, and studies the transformer winding state recognition method based on similarity matrix, the study includes the followings:1. Based on the transformer vibration theories, study the fundamental vibration characteristics of transformer windings, study the transformer vibration propagation characteristics, and then extract the vibration characteristics of the transformer tank vibration which can present transformer windings vibration;2. Study the sensitivity of500kv and220kv power transformer tank to winding vibrations. Relevant experimental researches and analyses are conducted for500kV and220kV transformers, the two typical types of single and three-phase large power transformers. And methods of the measuring point selection are also proposed, which can effectively reflect the two typical types of transformers’ winding vibration;3. Study the overall vibration characteristics of the transformer and the relationship between the various measuring points and, on this basis, focusing on studing the diagnostic methods and algorithms of the transformer condition identification based on the similarity matrix;4. In order to monitor and diagnose the transformer on line using vibration methods, and ensures the safety and reliability of the trans-former status; this paper design and implement a internet-based online monitoring and fault diagnosis system. |
