A power transformers' predictive overload system based on a Takagi-Sugeno-Kang fuzzy model

The improving of the utilization factors of mineral-oil-filled power transformers is of critical importance in the competitive market of electricity. Utilities need to change dynamically the loadability rating of transformers without penalizing their serviceability. As a key issue of loadability all aspects of the thermal performance, and in particular those related to the determination of tolerable windings hot-spot temperature (HST), overload practice and its impact on remanent life expectation should be investigated. So, this paper deals with a methodology for the identification of a Takagi-Sugeno-Kang (TSK) fuzzy model able to reproduce the thermal behavior of large mineral-oil-filled power transformers for implementing a protective overload system. The TSK fuzzy model, working on the load current waveform and on the top oil temperature (TOT), gives an accurate global prediction of the HST pattern. In order to validate the usefulness of the approach suggested herein, some data cases, derived from various laboratory applications, are presented to measure the accuracy and robustness of the proposed fuzzy model.