Thermal circuit model parameters identification of oil-immersed transformer based on PSO algorithm

The thermal circuit model of oil-immersed transformer needs to improve its accuracy in predicting the winding hottest-spot temperature, especially in the case of overload and different cooling modes. In the case of overload condition, the estimate values obtained from most existing models are smaller than the actual measurement values, which increase the potential of transformer overheating fault because of the estimate shortage. These limitations are mainly due to the thermal circuit parameters which are deviating when overloaded or having different cooling modes. In order to overcome these limitations, a parameter identification and correction approach based on PSO algorithm is proposed. The approach works on daily load pattern, permits better accuracy and improves the estimation security margin of thermal circuit in the presence of overload condition and different cooling modes.