Online Estimation of Transformer Hot Spot Temperature by Considering the Effects of Load Profile Modeling

One of the most valuable components in power systems is power transformer whose failure may result in a significant power loss. Therefore, one of the critical issues in power transformer operation is its health monitoring. Moreover, it was shown that the aging rate of transformers is very sensitive to the hot spot temperature, and when this temperature exceeds a threshold value, the aging rate increases. Given the fact that using temperature sensors in prefabricated and built-in transformers is not practical, thermal models are used to estimate transformer hot spot temperature. Since the transformer hot spot temperature is a key factor in the condition monitoring of a transformer, and in case it will exceed a threshold value, preventive actions should be taken in the proposed algorithm, the sensitivity of this important parameter with respect to the load profile sampling time is investigated. This paper proposes a fast online algorithm for the estimation of power transformer hot spot temperature by reducing the number of calculations without sacrificing accuracy. The proposed algorithm is applied to a 250 MVA transformer using MATLAB software. The results were compared with the actual factory test results and the efficiency of the proposed algorithm was shown.