Adaptive dynamic estimation on the operational capacity margin of distribution transformers

Load picked up by a distribution transformer is traditionally limited by the rated capacity provided by the manufacturer, which is not adaptive to the thermal environmental condition that the transformer is working in. In order to achieve more reliable and economical operation, the operational capacity margin is suggested to be estimated in a dynamic manner in this paper, with a data-driven algorithm proposed to perform the estimation, adapting to each transformer specifically. The algorithm is based on the combination of support vector regression (SVR) and optimization for the parameters of SVR for each individual transformer. Through a case study, the algorithm was verified to be effective.