Research On Intelligent Home Electricity Behavior Analysis Method Based On Power Consumption Model Constraints

With the continuous deepening of information technology,the Internet of Things and intellisense devices have gradually penetrated into people’s daily lives and have continuously changed people’s lifestyles and behaviors.Smart Grid is one of the most successful applications of Io T technology.How to use smart grid technology to better improve people’s quality of life has gradually aroused the attention of academia and industry.Unfortunately,although the smart grid technology has achieved great success in many fields such as power dispatching and fault detection,how to use the power model information contained in the smart grid data to further tap into the information based on the user’s individual power consumption behavior in order to realize privacy protection,smart recommendation,intelligent security and other functions are still important problems to be solved.This article focuses on the problem of user power mode decomposition,that is,how to decompose individual power consumption information from total power consumption information.In this paper,a method based on non negative matrix decomposition is proposed.This method is trained by the model of intelligent home data which has detailed electrical information.In the condition of obtaining the total power of the new user,the intelligent household electricity mode is obtained.This paper will provide a new way of thinking for researchers in this field.The research work and major contributions of this article are as follows,1.How to build a unified decomposition model on the basis of NMF method?Decomposition is a top priority in decomposing systems.A high-performance decomposition model is a prerequisite for successful application.How to establish a unified intelligent family data decomposition model is another urgent problem after fully studying the basic algorithm of NMF.This model should make full use of the existing intelligent home power data,and have the basic decomposition performance.The decomposition results can basically meet the user’s needs and are accepted by the user.2.How can we add more data in the model to improve the performance of the algorithm? In real life,there is a general relationship between users.These relationships are widely used in social media,information recommendation and collaborative filtering.In this decomposition model,using these relations information to improve the accuracy of decomposition results is also one of the key research directions in this paper.3.How to establish a unified framework for smart home data decomposition? The smart home data decomposition method should contain a whole set of decomposition processes from data acquisition to data feedback.It is a problem to be solved in this paper to establish a unified framework of decomposition that meets users’ needs and meets the needs of users.In order to solve the main problems mentioned above,this paper combines all factors and proposes an intelligent home power analysis method based on electricity mode constraints.The main functions of this method are as follows,1.In this paper,based on the basic algorithm,the iterative relationship in the data decomposition is fully considered,and the training process and the decomposition process are effectively fused,and a unified decomposition model is established.In this model,the training data is refactored after decomposition,and the test data takes full advantage of the result of the training during the decomposition,and the results of training data and test data in the iterative process promote each other to improve the decomposition performance.2.In order to take advantage of the potential relationship between data,this paper proposes a homogeneity hypothesis applied to social media,that is,there is a similar relationship between the individual electrical appliances when the total electricity consumption is similar.Based on this assumption,by introducing the laplasse matrix containing this information,the effective homogeneity regularization constraint is added to the decomposition model to improve the performance of the decomposition model.3.In order to establish a unified decomposition framework,based on the successful research of decomposition algorithm,this paper fully considers the relationship between the main elements and elements in the system,and establishes a unified framework for the whole process of data acquisition to data feedback.Finally,repeated experiments show that this method has better decomposition performance.Experimental results show that the accuracy of the proposed method is much higher than that of other methods.This method can be applied to families conveniently and quickly.The unified decomposition model proposed in this paper can add more data information through regular entries to improve decomposition performance and have better scalability.