A Probabilistic Approach for Forecasting the Allowable Current of Oil-Immersed Transformers

The overall growth of electrical energy consumption and the spread of distributed generation pose severe problems to system planners and operators, mainly due to the difficulties in expanding and upgrading electrical distribution systems. These problems can be partially overcome through the smart operation of power system components. In particular, under the new smart grid paradigm, lines and transformers could be loaded beyond their nameplate (static) rating in favorable thermal conditions, without loss of rated life or breakdown. This paper deals with the loading of oil-immersed distribution transformers, and it proposes a probabilistic approach based on the monitoring of electrical and environmental conditions. The approach consists in forecasting the predictive distribution of the transformer dynamic rating, and then in forecasting the transformer allowable current. The probabilistic approach takes into account the unavoidable uncertainties involved in the thermal modeling of the transformer; indeed, it selects the allowable current through an index that takes into account both the probability of the allowable current to be higher than the predicted dynamic rating and the corresponding expected load curtailment. Numerical applications are performed on real data to evaluate the effectiveness of the proposed procedure.