Forecasting dissolved gases content in power transformer oil based on particle swarm optimization-based RBF neural network

摘要

Accurate forecasting of dissolved gases content in power transformer oil is very significant to ensure safe work of entire power system. In order to realize accurate forecasting of these dissolved gases, particle swarm optimization-based RBF neural network (PSO-RBFNN) is proposed in the paper. Particle swarm optimization (PSO) has strong global search capability. Thus, PSO is adopted to determine training parameters of RBF neural network. The PSO-RBFNN forecasting performance is validated by engineering cases. The experiment results indicate that PSO-RBFNN has higher forecasting accuracy than GM, RBFNN in forecasting dissolved gases in transformer oil.