Non-invasive load monitoring method based on improved spectral clustering

Abstract

The invention discloses a non-invasive load monitoring method based on improved spectral clustering. The method comprises the following steps: acquiring historical active power signals, reactive powersignals, current and voltage on a target home bus by using a data acquisition module, and preprocessing the data to determine a home load feature library; selecting a group of active power signals, taking a differential signal to construct a graph signal, and requiring an action event with all loads of the family; obtaining a switching threshold of a target family by using a spectral clustering method; sending the real-time load characteristics acquired by the data acquisition module to a cloud server by using a 5G communication module, constructing a graph signal model, performing switchingevent spectral clustering by using an obtained family threshold, and performing positive and negative cluster pairing; and acquiring home load information and alarm information by using a feature matching method based on a graph signal, and synchronously updating the information to a client app. According to the method, the transmission problem when the data volume is large and the model trainingproblem when the sample features are few are effectively solved.

Description

technical field [0001] The invention relates to the field of non-invasive load monitoring, in particular to a non-invasive load monitoring method and system based on improved spectrum clustering. Background technique [0002] Home appliances exist in every household and are closely related to everyone's life. Reasonable management of the power consumption of home appliances and detection of the operating status of home appliances will not only help residents better understand their own electricity consumption, but also help Reduce the occurrence of electricity accidents and increase the safety and stability of household electricity. [0003] Non-Intrusive Load Monitoring (Non-Intrusive Load Monitoring, NILM) is to detect the mixed electrical signal at the home bus, use a series of recognition algorithms to analyze the independent operation of each internal electrical appliance, and finally obtain the operating information of the electrical appliance while decomposing the ene...

AI Technical Summary

Problems solved by technology

[0004] (1) It is highly dependent on the data, and the performance of the monitoring algorithm is poor when the amount of data is small;

[0005] (2) Limited by data transmission technology, the implementation cost of monitoring algorithms based on high-frequency features is high, while the accuracy of monitoring algorithms based on low-frequency features is low;

[0006] (3) The recognition effect on multi-level electrical appliances such as dishwashers and washing machines is poor;

Images