Load cluster control method and system based on support vector machine

Abstract

The invention provides a load cluster control method and system based on a support vector machine. The load cluster control method and the load cluster control system comprise the following steps: collecting user power consumption data and preprocessing the user power consumption data to obtain a load data sample; Training data set and test data set are established according to the load data sample; The support vector machine classifier is constructed by training data set and testing data set, and the support vector machine classifier is perfected after the support vector machine is perfected.Input the real-time load data to the improved SVM classifier, and output the classification results of the real-time load data. The invention adopts the support vector machine algorithm to classify the user loads in the electric power system, and on the basis of the classification, the load cluster centralized control is carried out, so that the load classification has scientific theoretical support and technical support, and the load classified into the same class is guaranteed to have high consistency.

Description

technical field [0001] The invention relates to the technical field of load classification control and analysis, in particular to a load cluster control method and system based on a support vector machine. Background technique [0002] At present, the controllable load connection of the load control system generally adopts the principle of proximity, and there is no in-depth research on the load connection level and load connection quantity. The load cluster control maintains the stability of the frequency and voltage of the power grid without affecting the characteristics of the power grid. , which is of great significance to the grid frequency stability control. Load control not only emphasizes control time requirements, but more importantly, it can achieve effective cluster control of loads that do not affect the characteristics of production activities; among them, load classification management can make load cluster control more reasonable and effective. Therefore, load...

AI Technical Summary

Problems solved by technology

[0003] At present, the load classification method has not been unified. The power sector often classifies the load according to the characteristics of the production activities of the users. However, the characteristics of the production activities of the users are complex and varied. There is a certain degree of subjectivity in the process of load classification, and the load classification results are not very accurate. , which will directly affect load control management and dynamic characteristic analysis of grid frequency

The current load classification method has defects: users in the same industry may have different load characteristics, which cannot fully reflect changes and differences in grid frequency, resulting in inaccurate classification results and affecting further applications on this basis, etc.

The above clustering algorithms are basically aimed at linearly separable data spaces, so they are not suitable for processing those linearly inseparable data. It is difficult for them to obtain a suitable clustering profile and the desired number of clusters, and the effect of clustering is not enough. Stablize

[0005] To sum up, the objective shortcoming of the prior art lies in the lack of load classification and load clustering methods

Images