Transformer Health Assessment Method Based on Loss Parameter Identification by Multivariate Linear Regression

In this paper, a state estimation method based on loss calculation is proposed, which reflects the change of transformer body parameters according to the loss and conducts health assessment. Based on the superposition theorem, the calculation formula of total loss of transformer considering harmonic loss is derived. Based on the actual experimental data, the mathematical model of the corresponding parameters is obtained by curve fitting method. The change mechanism is clarified and the accuracy of the calculation is verified. A method of parameter identification using multiple linear regression is proposed. Under the state of transformer operation, the parameters in the formula are calculated according to the actual operation data. The change of this parameter reflects the change of the health status of the transformer.