Residential Electricity Consumption Behavior Analysis Based On Non-intrusive Load Monitoring

Non-intrusive load monitoring can give specific information about electricity consumption of various electrical equipment,and further refine the analysis of residential electricity consumption behavior,which creates conditions for residential users to participate in demand response.According to the information of electrical appliances and users'electrncity consumption behavior obtained from non-intrusive load monitoring,it can guide users to change their electricity consumption habits,optimize their electricity consumption behavior,and help the grid to carefully grasp the load composition,provide guidance for grid planning and generation dispatching,and provide services for realizing bidirectional interaction between supply and demand and intelligent electricity consumption.Based on the research of non-intrusive load monitoring technology,the residential electricity consumption behavior is analyzed in this paper.A new event detection algorithm and load identification algorithm are proposed,and a complete system of power consumption behavior analysis using non-intrusive load monitoring method is established.The main results are as follows:According to the fluctuation of the total power signal during switching,an event detection algorithm based on sliding window is proposed.By calculating the variance of the power sequence of sliding window and the steady power difference before and after switching,the switching event can be detected,which can avoid the false detection caused by the normal power fluctuation of other electrical appliances in steady state operation,and can accurately determine the starting point and the end point of power fluctuation.Transient and steady-state processes of switching on and off of distributors.After detecting the switching of electrical appliances,the current characteristics of steady-state load operation are extracted as load characteristics,and a load feature library is established.A sample selection algorithm based on Adaboost is proposed to simplify the load feature library without affecting the accuracy of load identification and improve the efficiency of subsequent load identification algorithms.The A:-NN and kernel Fisher discriminant algorithm are combined to identify the unknown load samples.The simplicity of k-NN is used to identify the unknown load samples in the first round.According to the results of k-NN identification,the types of electrical appliances are selected or the nuclear Fisher is further used to discriminate,so as to improve the recognition accuracy of similar electrical appliances.The transient power sequence of load switching transient process is extracted as load characteristics.The load identification algorithm combined with ELM and PSO-GA algorithm is used to identify single or multi-appliance switching.The designed non-intrusive load monitoring method is applied to the analysis of residential electricity consumption behavior,and the electricity consumption information of various electrical appliances is counted,including the types of start-up and shut-down,start-up and shut-down time,energy consumption,electricity consumption,etc.An example is given to verify that the proposed method can be used to analyze the residential electricity consumption behavior on the basis of identifying electrical appliances accurately.