Identification of thermal parameters for transformer FEM model by differential evolution optimization algorithm

Abstract

This paper presents a thermal model to simulate a thermal behavior of transformers. Heat dissipated is always problem in transformers. Coupled physical and mathematical models would assist in the development of a system that was both accurate and simple. Material properties, the geometry, heat transfer coefficients for each surface are introduced as the input values. For the accurate results of the temperature distributions, the exact values of heat transfer coefficients are required. This paper deals with identification of thermal parameters for FEM-based transformer model. The identification technique is based on stochastic optimization algorithm called Differential evolution (DE). The goal of DE is to identify adequate thermal parameters, in order to achieve best possible agreement between measured and model-calculated values of temperature. The inclusion of identified thermal parameters into the model results in good model's accuracy.