The operation status of power transformers

the hub equipment in the operation of electrical power systems

is directly related to the safe operation of the entire system. For a long time

dissolved gas analysis (DGA) has served as an effective means for diagnosing transformer faults mainly due to its immunity to complex electromagnetic fields inside the transformers and external noise. For solving the limitations of existing DGA- based fault diagnosis method and improve accuracy of fault diagnosis further

a kind of fault diagnosis method for the oil immersed power transformer based on the capsule network (CapsNet) is proposed. The proposed method

by combining the unique capability of CapsNet to handle vectors and the characteristics of dissolved gases in transformer oil

can be able to accurately model the intricate nonlinear relationship between dissolved gases and types of faults

realizing the automatic extraction of key features of the realized training of the model through dynamic routing technique and back propagation algorithm. The test result based on real DGA data shows that the proposed method achieves 90.48%、82.46% and 87.93% on the fault diagnosis accuracy

macro average and recall geometric mean

and it outperforms both SVM and DBN

demonstrating its advantages in real-world transformer fault diagnosis applications.