Load identification method and apparatus

By creating RGB images and combining them with deep neural networks to identify load categories, the problems of high computational cost and poor accuracy in the NILM method are solved, achieving efficient load identification.

CN117452111BActive Publication Date: 2026-06-26HEBEI UNIV OF TECH
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
HEBEI UNIV OF TECH
Filing Date
2023-10-29
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Existing NILM methods for load identification are computationally intensive and have poor accuracy. In particular, methods based on deep learning algorithms rely on long-term electrical appliance data features and require training different neural network models for different types of appliances, resulting in poor load identification performance.

Method used

By extracting the current time-domain data of the target load, an RGB image is created. The R, G, and B channels are used to represent the reactive power, power current, and power factor grayscale images, respectively. A deep neural network is then used to identify the load category, reducing the amount of computation and improving accuracy.

Benefits of technology

It reduces the computational load of load identification, improves the accuracy of load identification, and achieves efficient load category identification.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117452111B_ABST
    Figure CN117452111B_ABST
Patent Text Reader

Abstract

The application provides a load identification method, and relates to the technical field of load identification, which comprises the following steps: extracting current time domain data of a target load; acquiring voltage time domain data of the target load; creating an RGB image based on the current time domain data of the target load and the voltage time domain data of the target load; and determining the category of the target load based on the RGB image. According to the application, the load category can be identified based on the current characteristics of the load to be detected after the current is separated, a mixed color image obtained by superimposing three mixed pixel matrices, which represents the amplitude change of the load current under different harmonics, represents the characteristics of the reactive power, the power factor and the power current, and does not require a large amount of data, thereby reducing the operation amount of load identification and improving the accuracy of load identification.
Need to check novelty before this filing date? Find Prior Art