Abstract:The traditional transformer fault diagnosis method has incomplete coding, and it is easy to misjudge the fault. With the development of transformer on-line monitoring technology and the increase of product demand, which transformer fault diagnosis technology is developing in the direction of intelligence. In order to improve the accuracy of transformer fault diagnosis, combined with the gas analysis method in oil, a transformer fault diagnosis method based on fruit fly optimization algorithm (FOA) and probabilistic neural network (PNN) model is proposed. As a new heuristic and evolutionary algorithm, the fruit fly optimization algorithm has the advantages of easy understanding and fast convergence to the global optimal solution. The structure of the PNN is simple and the training is concise. PNN has strong ability of nonlinear classification, which the sample space is mapped to the fault pattern space to form a diagnostic network with strong fault tolerance and mechanism self-adaptive ability. The approach of FOA is used to optimize the model parameters to reduce the impact of human factors on the neural network design. The simulation experiments show that FOA-PNN can be effectively applied to transformer fault diagnosis and provides a new way for transformer fault diagnosis. The fault diagnosis method of FOA-PNN has a good research value and development prospects.