Non-Invasive Three-Phase Load Modeling And Its Identification Algorithm

Non-invasive load monitoring is the latest intelligent grid advanced measurement technology,which only sampling the voltage and current at the customer's total incoming line,and identified the working sequence of the electric appliance through intelligent algorithm,so as to ensure that there is no need for access to the indoor construction,construction and maintenance cost.The non-intrusive load monitoring system can obtain fine-grained power consumption information such as user-side electrical appliances,start-stop status and power consumption,which can enable residents to understand the household energy consumption in real time,so that residents can reasonably arrange the electric time of the electric appliance for the purpose of saving electricity cost.Meanwhile,the fine-grained load monitoring data of the user also provides data support for the residents' side demand response.First of all,this paper deeply analyzed the load characteristics of various types of power equipment operation,and extracts the identification characteristics of the load operation state,including steady-state active power,steady-state reactive power,steady-state harmonic current,transient power impact and load time characteristics,and puts forward the online extraction strategy of load identification characteristic quantity.Then,a load identification model based on the working principle of home appliances was established,including electric heating model,air conditioning model,kitchen appliance model and washing machine model.On the basis of various load identification models,non-invasive load identification algorithm based on hierarchical classification recognition structure is proposed.The examples of single-phase family users include fixed frequency air conditioning,frequency conversion air conditioning,electric heating equipment,electric rice cooker,microwave oven,electromagnetic oven,refrigerator and other electrical appliances.The test conditions include the typical continuous operating condition of household appliances,the working condition of the family meal preparation scene,the scene working condition of the family after meals and the working conditions at night,the identification accuracy can reach 88%,and the power collection accuracy is about 85%;The examples of three-phase family user include three-phase central air conditioning,single-phase air conditioning(including fixed frequency and frequency conversion),electric oven,electric steam box,dryer and electric heating class.The test condition is the most active scene in the evening,with the identification accuracy of about 85%,and the accuracy of power collection is around 80%.Finally,the research on the non-intrusion load identification of office building users is carried out,and the three-phase load identification model of high energy dissipative appliances in buildings is established,including the model of air conditioner,three-phase hot water heater and new wind turbine model.Combining with the actual operating condition waveform of office buildings,the technical difficulties of non-intrusion load identification of office building users are analyzed in detail.The non-invasive load identification algorithm based on online learning has been designed,and it is indicated that the recognition accuracy of the central air conditioner can reach 50%.