User typical load curve construction method based on density clustering

Abstract

The invention discloses a user typical load curve construction method based on density clustering. The user typical load curve construction method comprises the steps of (1) inputting the daily load time sequence of an object in a certain time interval, (2) conducting density clustering on the daily load time sequence by setting the searching radius epsilon and the minimum object number M according to the density variation of distribution of the normalized daily load time sequence in vector space, (3) dividing the object of the daily load time sequence into one or more time sequence clusters and one or more noise time sequence classes according to the results of the density clustering, and (4) extracting one or more typical load time sequences of a user. According to the characteristics of complex load changes of the user, the user typical load curve construction method uses the density variation of distribution of the normalized daily load time sequence in the vector space, the typical load time sequence of the user is extracted by setting reasonable density search parameters, the cluster number is not required to be appointed, noise loads can be removed, and the user typical load curve construction method is much suitable for extracting the one or more typical load time sequences in the user stage.

Description

technical field [0001] The invention belongs to the technical field of power system management and control, and in particular relates to a method for constructing a typical user load curve based on density clustering. Background technique [0002] With the continuous development of the economy, the power load at home and abroad has continued to grow in recent years, and the way to match the load growth by increasing the installed capacity of the unit is subject to many restrictions such as grid structure and line transmission capacity. Side management, demand response, etc. are used as the main means to balance load and reduce load peaks. [0003] Mastering the power consumption habits and load characteristics of power users is the prerequisite for effective demand-side management. However, due to the relatively lagging development of the distribution side of the domestic power system for a long time, it has caused the phenomenon that power user load data is difficult to obt...

AI Technical Summary

Problems solved by technology

The Chinese patent with the publication number CN102156814A proposes a typical daily load curve extraction method based on k-means partition clustering, which needs to input the number of clusters k in advance, and repeatedly iteratively optimize the load object set, but In the case that the object contains noise load, this method cannot distinguish, but classifies the noise load into a certain category, so it is not accurate and efficient

Images