Analysis And Application Of The Residential Non-intrusive Load Monitoring Cloud Data

Residential non-intrusive power load monitoring technology is a technology to obtain the detailed operation characteristics of electric appliances of power users through advanced measurement schemes.It is the basis of the development of smart grid and plays an important role in improving the power consumption efficiency of users,promoting demand response,and realizing energy conservation and emission reduction.By analyzing the operation characteristics of household electrical appliances,non-intrusive load monitoring can obtain residents’ electricity usage habits and further refine their electricity usage behaviors,which lays a foundation for the power company to promote residents’ participation in the demand response.On the basis of non-invasive load monitoring terminal monitoring the user’s electricity consumption behavior,this paper analyzes the cloud data of non-invasive load,proposes the analysis method of cloud load characteristics,studies the new load identification and improvement technology,and uses the non-invasive load data to realize the application of electricity behavior analysis and energy efficiency optimization.The main results are as follows:First,this paper systematically analyzes the general situation of non-intrusive load cloud data.Operating characteristics such as user’s electrical equipment type,total electric power consumption,peak power,start and stop times,start and stop times,running time,etc.By means shift algorithm clustering,the user’s historical electrical cluster is obtained,and the distribution of the characteristics of each historical electrical cluster is statistically analyzed.The identification characteristics of cloud load are analyzed,and the load time characteristics,load power characteristics and running time distribution are proposed as the basic characteristics of cloud load identification.The paper summarizes the cloud load identification characteristics of each kind of electrical appliances,and determines the types of historical electrical appliances according to the different load identification characteristics of various electrical devices in the cloud.Then,after the user’s feature bank of historical appliances is established in the cloud,the operation feature probability model of the same historical appliance cluster is established through the multivariate gaussian distribution.The cloud load identification algorithm is proposed to determine the category of the unknown electrical appliances by calculating the operation characteristics of the unknown electrical appliances corresponding to the occurrence probability of the electrical appliances in each user’s history library.The algorithm program was tested by the actual data of a certain special user and the users of the whole village.The recognition accuracy of the algorithm was improved by 38% on a single user and 8% on the user of the whole village,which verified the engineering practicability of the algorithm.Finally,based on the characteristics of electrical appliances and consumption habits,the users’ behavior parameters are established,and the k-means clustering algorithm is used to cluster power users to form a comprehensive picture of user behavior.The energy efficiency optimization schemes of transparent electricity,green electricity and safe electricity are put forward,and the potential of residents’ controllable load is calculated,and a load control model of residents’ participation in demand response is established.The practical application of non-intrusive load monitoring technology in cloud is demonstrated through practical engineering examples.