Non-intrusive Power Load Decomposition And Analysis Of Relevant Statistical Characteristics Based On HPLC Data

In recent years,with the acceleration of China’s layout in the field of new energy and the enhancement of support,China’s determination to protect the environment and change the energy landscape has become increasingly firm.Due to its characteristics of safety,cleanliness,and easy conversion,the consumption of electric energy continues to increase and will be applied in more fields in the future.However,as one of the main electricity users,household electricity is still in a regulatory blind spot due to limitations in technical management capabilities,with disorderly planning.In current research,non-invasive load monitoring has been gradually promoted due to its practicality.However,further analysis and visualization of the collected load data is still in the initial stage,lacking in-depth analysis of user electrical category information and software system visualization presentation.Since the data collected by non-invasive load monitoring technology is only the total load data of a user,and the specific data information of each electrical appliance cannot be directly obtained,it is necessary to find a suitable algorithm to disassemble the total load into the load information of various electrical appliances,and this process has problems such as complex types of electrical appliances,different user electricity habits,and fast replacement of electrical appliances.To solve the above problems,the following research is carried out in this thesis.Firstly,the relevant theories such as flexible load,non-invasive load monitoring and HPLC smart meter are studied,and the principles of commonly used variable point detection and Kmeans clustering algorithm are introduced.Secondly,based on the application scenario of this thesis,three multi-variable point detection methods are selected: binary segmentation method,segmented neighborhood method and PELT algorithm,and the most suitable algorithm is selected by using Hausdorff distance and F1-score through simulation data: segmented neighborhood method.Then,using the real data set of a user’s monthly HPLC meter in Nanchang for verification,the "meaningful" change points in the household electricity load fluctuation curve are searched by the segmented neighborhood algorithm,and the load decomposition is carried out with the change point as the segmented point,which is divided into multiple load curves,and then the K-means algorithm is used to cluster according to the load characteristics of each curve to finally obtain the user’s electricity consumption information.Finally,this thesis uses python language to develop GUI graphical user interface,build a visual system demonstration platform,package and package model algorithms,and display the analysis process and results in the form of exe program interface to realize local operations on the Windows platform.In order to improve the sensing function of load information of home energy routers,realize home power load decomposition and obtain relevant statistical information,based on HPLC data through non-invasive load monitoring technology,this thesis establishes an algorithm model based on segmented neighborhood method for load decomposition and Kmeans clustering for feature extraction,which reduces the system complexity and equipment cost of existing analysis methods,and finally designs and builds a visual system demonstration program to analyze real household electricity consumption data.The analysis structure is presented through the GUI interface,which verifies the effectiveness of the algorithm and provides a more convenient and intuitive visualization scheme for household electricity information.The HPLC data acquisition cost in this thesis is low,the algorithm model structure is clear,concise and practical,and the model combining segmented neighborhood method and K-means clustering is popularized and applied to the field of power load decomposition,and the developed visualization system software program is simple and easy to generalize.And through the analysis of household electricity load,understanding the use of household appliances is conducive to helping all departments of the power system to make decisions and deployment,help residents improve their own electricity habits and rational electricity use,and then achieve the effect of peak shaving and valley reduction,energy conservation and emission reduction,and provide support for ensuring the basic electricity consumption of society and ensuring the safety and stability of power supply.