Clustering-based typical daily load curve selecting method and device

Abstract

The invention provides clustering typical daily load curve selecting method and device. The method comprises the following steps: reading a curve within a time span, determining the number k of typical daily load curves, and selecting k curves as a set center; classifying the curves into a set (S3) nearest to the set center; and calculating a new set center, determining whether the new set center is the same as the previous set center or not, determining whether the difference with the previous target function is within a preset range or not if the new set center is not the same as the previous set center, returning to the S3 if the difference is not within the preset range, and defining the curves in each set nearest to the set center as the typical daily load curves if the new set center is the same as the previous set center or the difference of the previous target function is within the preset range. The method ensures that all the daily load curves are grasped on the basis of a clustering thought; the generated samples inside the set are similar, while the samples in different sets are different, so that the discovery of a global distribution mode is facilitated, single index calculation or averaging processing is avoided, the influence of random and subjective factors can be reduced, and the sensitivity of directly extracting a single curve on bad data is reduced, therefore, the method is more suitable for discovering potential regulation of large-scale data, and can be used for characterizing the whole regulation better.

Description

technical field [0001] The invention relates to the field of power systems, in particular to a cluster-based typical daily load curve selection method and a cluster-based typical daily load curve selection device. Background technique [0002] In the power system, the load curve refers to the curve of the power load in the power system changing with time, the abscissa is time, the ordinate is generally active power, and the daily load curve refers to the load curve within a day. figure 1 2 shows two daily load curves in the unified load data of a certain province in a certain year, and the sampling interval is 15 minutes, so each daily load curve consists of 96 points. The daily load rate is an index that characterizes the characteristics of the daily load, defined as the ratio of the daily average load to the daily maximum load. [0003] The typical daily load curve is the most representative curve among the daily load curves within a certain period of time. It is the basi...

AI Technical Summary

Problems solved by technology

The daily load rate is a characteristic index of the load curve, but it cannot characterize the entire curve shape

by figure 2 For example, the two curves shown in the figure are the daily load curves of the unified load of a certain place in a certain year, and the corresponding daily load rates are very close, both between 0.81 and 0.82, but the shapes of the two curves are very dissimilar , so there are limitations in selecting only the daily load rate or other single daily load characteristic indicators

Images