Transformer winding deformation fault diagnosis method based on improved variational modal decomposition and optimized correlation vector machine

The design of the transformer can withstand the mechanical shock during transportation and the large electromechanical power generated by short circuit and switching surge current. However, due to these forces, the windings may be slack, misaligned or deformed. The windings are the core component of the transformer, which directly affects the diagnosis of the transformer and even the entire power system. This paper proposes a feature extraction method for transformer winding vibration signals based on improved variational modal decomposition (IVMD). The improved variational modal decomposition is used to decompose the vibration signals under different states of transformer windings into a series of finite bandwidth eigenmode functions (IMF); on the basis of feature extraction, this paper also introduces the correlation vector machine (RVM), through the pattern recognition algorithm to establish the relationship between the weight divergence and the transformer winding operating conditions, and through the harmony search algorithm to optimize the option to the RVM kernel function. The algorithm proposed in this paper can effectively extract the characteristics of the mechanical vibration signal of the transformer winding and improve the classification accuracy of RVM.