Connection Phase Estimation of Pole Mounted Distribution Transformers by Majority Voting of High-quality Solutions using Only Selected Stages

This paper proposes connection phase estimation of pole mounted distribution transformers by majority voting of high-quality solutions using sampled stages. The essential challenge of the connection phase estimation of pole mounted distribution transformers problem is that the correct connection phase combination and the combination with the best objective function value do not match because of using inconsistent data. In order to solve the challenge, the proposed method utilizes the concept of majority voting of ensemble learning andtheproximate optimality principle. The proposed method is verified to estimate the connection phase accurately and specify pole mounted distribution transformers with insufficient estimation accuracy using a distribution system model.