Adaptive clustering method and system based on intra-region distance

Abstract

The invention discloses an aadaptive clustering method and system based on an intra-region distance. The method comprises the steps of obtaining data defining all data elements of an initial cluster; defining an initial clustering region center of the initial clustering region, wherein the region center represents a point of an average value of distances of all data elements in each region; determining a distance value in an initial clustering region in each initial clustering region according to the data elements of the initial clustering and the central point of the initial clustering region; comparing the distance value in the initial clustering region with a preset threshold value, combining or splitting the initial clustering region according to a comparison result, and determining a data element in the current region and the center of the current region; and updating the current region, and outputting the updated data elements of the clustering region after one or more iterations.

Description

technical field [0001] The present application relates to the technical field of online analysis of power systems, in particular to an adaptive clustering method and system based on intra-area distance. Background technique [0002] Clustering models and methods are capable of processing a collection of data elements to generate combinations defining assignments of each of the data elements to one or more data elements. [0003] According to the national energy policy proposed by my country in 2020 to peak carbon dioxide emissions in 2030 and achieve carbon neutrality in 2060, green energy dominated by photovoltaics, wind power, and hydropower will be further developed in the future. In order to realize the full and full consumption of new energy, large-scale and large-scale cross-regional UHVDC transmission and consumption will become a new model of the power grid. Under the new mode, the degree of coupling at the sending and receiving ends of the power grid is deepened, a...

AI Technical Summary

Problems solved by technology

For cross-regional interconnection grids, due to the different labeling requirements implemented in each region, the data labels are not uniform, and manual adjustment is required first, which is not operable in large-scale large-area data scenarios

Based on the nearest neighbor classification method, k-means clustering algorithm, and support vector machine algorithm, the number of clusters (or partitions) needs to be specified in advance, and the partitioning algorithms are all open-loop algorithms, which cannot be adjusted independently to achieve optimal clustering

[0005] Aiming at the technical problems of adaptive clustering and grouping analysis of large-scale online data of power systems in the above-mentioned prior art, no effective solution has been proposed so far

Images