Prediction of Transformer Windings Temperature at the Moment of Power Failure Based on Regression Curve Fitting Algorithm

There are three methods for temperature rise test: direct load method, mutual load method and short circuit method. The national standard GB1094.2-2013 stipulates that the short-circuit method is the standard method for the temperature rise test of oil-immersed power transformers. This paper presents a method and system for predicting winding temperature of oil transformer at the moment of power disconnection. Through a set of winding resistance values continuously measured at a fixed time interval after the power is disconnected, the winding temperature is calculated, and the cooling curve is fitted based on the continuous winding temperature. Thus, the winding temperature at the moment of power supply disconnection can be obtained, which improves the efficiency and accuracy of the calculation of winding temperature and solves the problems of inaccuracy, invisibility and low efficiency of the existing methods.