Low-voltage Theoretical Line Loss Calculation Based on Improved K-means Clustering and Fitting

Abstract

With the problem of inaccurate calculation results due to missing data such as topology and operating load when calculating low-voltage theoretical line loss, this paper establishes a theoretical line loss calculation model, and uses machine learning algorithms Solve. In the solution process, the influence of abnormal data is reduced by improving the K-means algorithm, and for typical daily types, the relationship between the power consumption in transformer area and the variable line loss power is obtained by the fitting algorithm to obtain the theoretical line loss reference value. The basic data of this method is highly fault-tolerant, suitable for actual engineering calculations, can provide accurate theoretical line loss reference values, meet the requirements of line loss analysis in transformer area, and provide an effective basis for the formulation of loss reduction measures.