Fine classification method and system for power multivariate load users

Abstract

The invention discloses a fine classification method and system for power multi-element load users, and the method comprises the following steps: obtaining the historical power consumption data of thepower multi-element load users, and carrying out the preprocessing of the historical power consumption data; performing clustering analysis on the preprocessed data to obtain load characteristic curves of different types; according to the different types of load characteristic curves, using an SVM classifier to carry out training, and generating a classification prediction model; and inputting to-be-classified user historical power consumption data into the classification prediction model to obtain an output classification result. Through data preprocessing, the reliability and accuracy of the data can be improved; the clustering effect obtained by clustering analysis is more accurate; the generated classification prediction model can provide more scientific data reference for electric power decision making; and finally, the users to be classified can be classified accurately, quickly and finely.

Description

technical field [0001] The invention relates to the technical field of electric power big data, in particular to a fine classification method and system for electric power multiple load users. Background technique [0002] In today's information technology field, data mining technology, machine learning and other methods dig deep into the "data ocean" accumulated by power companies in the construction of "digital power", and realize intelligent prediction and processing of problems in power companies. At present, the research on the classification of power users is mainly to improve the stability of the power grid and improve the utilization rate of power resources. With the wide application of data mining technology and machine learning technology, continuously optimized electricity user classification algorithms will be produced one after another. Combining deep learning and clustering methods to deeply mine load data has become an inevitable trend. [0003] With the rap...

AI Technical Summary

Problems solved by technology

But its technical solution has two deficiencies: (1) The data acquisition module of the system directly obtains the user's power consumption

The disadvantage of this method is that the dimensions of general user load data are too large, and there are still problems in the acquisition process due to large-scale power outages, meter failures, and data omissions in the transmission process; (2) The cluster analysis module of the system is based on preset The clustering method uses the size level of the demand for electric power and the rate of change of the demand for electric power as the clustering index to finely classify the users.

This method does not consider the relationship between the user load itself and the load user category; and uses the clustering method to classify the load users, ignoring the error rate in the clustering process, and it is difficult to solve the problem of load user classification prediction

Images