Non-intrusive load identification method based on metric learning

Abstract

The invention provides a non-intrusive load identification method based on metric learning, and the method comprises the steps: mapping load current characteristics to a metric space through a one-dimensional convolutional neural network, achieving the clustering of the characteristics through the triple loss during network training, and carrying out the similarity discrimination of the characteristics of the metric space, thereby achieving the load identification. According to the method, the unknown load can be effectively identified, and the generalization ability is high; on the other hand, metric learning serves as one of small sample learning methods, dependence on training samples can be relieved, and high practicability is achieved.

Description

technical field [0001] The invention relates to the field of non-intrusive load monitoring (NILM), in particular to a non-intrusive load identification method based on metric learning. Background technique [0002] With the rapid development of society, the demand for energy continues to increase, and electric energy, as the main secondary energy, is one of the main energy utilization methods. How to improve the efficiency of electric energy use and realize intelligent power consumption has attracted widespread attention, and it is particularly important to grasp the detailed power consumption information on the user side. Traditional intrusive load monitoring requires the installation of acquisition and communication devices at each electrical load to detect the load status, and requires modification of existing electrical appliances or lines, which is difficult and expensive to implement. The non-intrusive load monitoring technology analyzes the status of each load in the...

AI Technical Summary

Problems solved by technology

First of all, most of these models rely on a large number of labeled samples for model training, but in practical applications, it is often impossible to obtain enough labeled data or the acquisition cost is high; secondly, these models usually assume that all loads in the scene are known, and new loads are added to the scene Afterwards, the original model cannot recognize the new load, and even affects the recognition effect of the original load; finally, these models have poor generalization performance, and often need to retrain the model after changing the scene. Training and building a model for each scene separately will Lead to greatly increased maintenance complexity and cost

Images