Particle Swarm Optimization Algorithm-Capsule Network Model for Transformer Fault Diagnosis
In order to improve the accuracy of transformer fault diagnosis,this paper proposed a particle swarm opti-mization algorithm-capsule network model for transformer fault diagnosis method.To address the drawback of insuffi-cient correlation between the features of dissolved gas data in transformers,a basic classifier combining a one-dimen-sional convolutional neural network(1D-CNN)and a capsule network(CapsNet)was constructed(Improved Capsule network,ICapsNet).To enhance the diagnostic performance of the basic classifier,an improved particle swarm opti-mization algorithm(IPSO)was used to optimize the learning rate of ICapsNet.Experimental results showed that the proposed method achieved a diagnostic accuracy of 94.44％,which was an improvement of 4.76％compared to the traditional CapsNet.Additionally,by incorporating IPSO to optimize the learning rate of ICapsNet,the model's accu-racy increased by 3.96％.The research findings provided a new approach for transformer fault diagnosis.
Power transformerFault diagnosisCapsule networkParticle swarm optimization
万方数据
维普
