Key Technology Research And Prototype System Implemetation Of Non-intrusive Load Identification Module Based On Smart Meters

Smart grid is the development trend of power grid in the future,which has the characteristics of strong and reliable,cost-effective,clean and environmentally friendly,transparent and open,and friendly interaction.Smart meters are a key part of building a smart grid,which are widely used in smart grids.As one of the important components in the advanced metering infrastructure system of smart meters,load monitoring has extremely high research value.Traditional load monitoring method is an intrusive monitoring method,which mainly involves installing relevant sensors on the sockets of each household electrical appliances of residential users and using the sensors to record the use of various electrical appliances.The monitoring data of this method is accurate and reliable,but The cost of equipment and labor is high,which is not conducive to large-scale popularization and application.Non-intrusive load monitoring only needs to monitor the total voltage,current and power consumption at the entrance of the residential home in real-time to estimate the use of specific household appliances in the user’s home and provide real-time monitoring data to the residential households,let users understand their own power consumption habits,including the type of electrical appliances used,use time and energy consumption,etc.let users spontaneously control some power-hungry appliances use frequency,to achieve effective and efficient energy saving.At present,domestic non-intrusive load monitoring is in the theoretical development stage,and there is a lack of mature products related to non-intrusive load monitoring in the market.This project is cooperating with a smart meter company to set up a non-intrusive load identification and verification laboratory.Based on the in-depth study of the non-intrusive load monitoring system,a multi-feature non-intrusive load identification based on smart meters is designed.This module can be applied to the new generation of smart meters of State Grid,and has been tested in the laboratory.The result shows it has a good load recognition effect.The main contributions of this paper are summarized as follows:Firstly,on the basis of the operation characteristics of the electrical appliances,the load characteristics can divide into steady-state characteristics and transient characteristics,and the load characteristics and extraction methods of typical household appliances in the laboratory are studied in depth.The analysis results establish a load feature library for non-intrusive load identification.Secondly,the overall architecture and key technologies of non-intrusive load monitoring system are studied in depth.An event detection algorithm based on transient current was proposed to detect the load switching process.Based on the analysis and introduction of the algorithm of template matching,decision tree and support vector machine,we propose a multi-feature multi-layer non-intrusive load monitoring classification model.Using MATLAB simulation platform,the algorithm model proposed in this paper is used to simulate the single load identification and mixed load identification of laboratory electrical equipment.The results show that the multi-feature multi-layer non-intrusive load monitoring classification model proposed in this paper has higher accuracy of load identification rate,and load identification effect is better.Finally,based on the research results of the above algorithm,the appropriate ARM chip was selected for the hardware design and software design of the non-invasive load identification apparatus,the sample of the non-invasive load identification apparatus was designed,and the preliminary test of the embedded platform was carried out,which laid a foundation for the subsequent algorithm transplant and the apparatus test in the grid.