Research On Non-intrusive Residential Load Feature Extraction And Intelligent Power Utilization Research On Non-intrusive Residential Load Feature Extraction And Intelligent Power Utilization

As an important part of smart grid system,the development of smart power technology will directly affect the economic operation and orderly power consumption.In order to acquire more precise load monitoring data and develop intelligent power technology,intrusive monitoring method is mainly adopted at present,that is,sensor devices are installed on every electrical appliance in the user's home to record its usage.Because it will affect the life of the user and the cost of investment is high,it is difficult to popularize.Non-intrusive load monitoring technology can effectively avoid the above problems.In this thesis,non-intrusive residential load monitoring is studied.Through the study of load feature extraction and identification,user load data mining and other aspects,non-intrusive load monitoring technology can be used to obtain more detailed power load data.This thesis mainly divides into three aspects: non-intrusive load monitoring feature extraction and load identification method,non-intrusive load decomposition method and typical application of non-intrusive load decomposition oriented to intelligent power consumption.Firstly,the feature extraction methods and load identification methods commonly used in non-intrusive load monitoring are introduced.Aiming at the problem that the existing methods generally use the transient characteristics of load to analyze,which requires high accuracy of data acquisition equipment and is difficult to apply in practice,a new method of load feature extraction is proposed.This method not only considers the electrical load characteristics of the electrical equipment itself,but also fully excavates the details hidden in the historical data.To solve the problem of low efficiency in load identification,a non-intrusive residential load identification method based on neural network is proposed.The method uses the pattern recognition technology of neural network to realize accurate non-intrusive load identification.To solve the problem of non-intrusive load decomposition,a non-intrusive household load decomposition method based on work pattern recognition of electrical equipment is proposed.Firstly,the steady-state characteristics of current waveform,active power and reactive power of household appliances are extracted.On this basis,a new method to determine the working mode of household appliances is proposed.Then,the artificial neural network is used to recognize the working mode of electrical appliances,and the load of household appliances is decomposed accurately.Finally,the simulation results show that the proposed method can effectively identify and precisely decompose the load.Aiming at the typical application of non-intrusive load decomposition for intelligent power consumption,based on the research of non-intrusive load decomposition,this thesis further studies the deep mining of state data after load decomposition,and analyses the application of non-intrusive load decomposition in the analysis of power consumption behavior and power demand side management business of intelligent power consumption technology.