Development of power transformer thermal models for oil temperature prediction

Abstract

This paper describes a new thermal model of oil-immersed, forced-air cooled power transformers and a methodology for model construction using intelligent learning applied to on-site measurements. The model delivers the value of bottom-oil and top-oil temperatures for thermal performance prediction and on-line monitoring of power transformers. The results obtained using the new thermal model are compared with the results of a traditional thermal model and the results derived from artificial neural networks.