Transformer area electric characteristic parameter-based low-voltage transformer area clustering method

Abstract

The invention provides a transformer area electric characteristic parameter-based low-voltage transformer area clustering method. The method includes the following steps that: step 1, standardization processing is performed on transformer area electric characteristic parameters; step 2, an initial clustering center is selected; step 3, k-means clustering is performed on a transformer area, so that clustering results can be obtained; step 4, the outline coefficients of the clustering results are calculated; and step 5, an optimal clustering result is selected. According to the transformer area electric characteristic parameter-based low-voltage transformer area clustering method of the invention, the optimal clustering result is selected according to the outline coefficients, and therefore, the accuracy of clustering can be increased; and the initial clustering center is selected through calculating an evaluation index, and therefore, compared to a traditional k-means clustering method, the clustering results of the improved clustering method are no longer sensitive to the initial clustering center. With the method of the invention adopted, the problem of poor training accuracy of intelligence algorithms such as a neural network, which is caused by the dispersion of the numerical values of the line loss rate of the transformer area can be effectively solved, and technical support can be provided for analysis on the line loss of the transformer area.

Description

technical field [0001] The invention relates to a clustering method, in particular to a low-voltage station area clustering method based on the electrical characteristic parameters of the station area. Background technique [0002] The low-voltage distribution network station area is the terminal link in the power system, and the line loss rate is one of the important assessment indicators for the station area management. However, the construction and management of low-voltage station areas are uneven, the number of station areas and end users is large, the account management is incomplete, the distribution of lines is complex and diverse, and the success rate of power collection systems varies greatly. Most of the theoretical calculation methods use the voltage loss method, equivalent resistance method, etc. For the station area, both the calculation of the theoretical line loss rate and the evaluation of the statistical line loss rate require a lot of manpower and material...

AI Technical Summary

Problems solved by technology

However, considering that there are large differences in the grid structure of low-voltage distribution network areas in my country, when the nature and proportion of power consumption are within a certain range, the power consumption and load rate vary greatly, and intelligent algorithms such as neural network theory are directly applied and Can not get ideal results, so it is necessary to classify and analyze the station area

Images