The Research On Non-intrusive Power Load Identification

Non-intrusive power load identification can obtain the energy load energy consumption information by collecting and analyzing the load data such as voltage and current of the incoming terminal bus.For power companies and users,it is possible to guide users to participate in the “sharp peak filling” measures of the power grid to improve the efficiency of power usage.It can provide data support for the development of smart grid and has important social and economic benefits.In this paper,a non-intrusive power load monitoring device is designed,and the waveform data of working voltage,current and active power of common laboratory power load are collected.On the basis of analyzing the characteristics of electric load,The power load switching event detection is performed by using a sliding window based CUSUM algorithm.The Power load type identification is performed by using a The ELM model.The main work and achievements are as follows:⑴ A non-intrusive power load monitoring device based on HLW8112 is designed for the real-time collection of power load data.The device collects current,voltage and other electrical load data through the sampling resistor and the HLW8112 single-phase energy metering chip.The STM32F103 microprocessor is used to calculate the power load characteristic value and upload it to the host computer through the WiFi wireless communication module.Experimental results show that the device can monitor single load and multiple load conditions.⑵ The waveform data of instantaneous current RMS,instantaneous reactive power and instantaneous active power of the electrical load are collected,and the adaptive Gaussian filter is designed to preprocess the data.Based on the analysis of the steady state and transient electrical characteristics of the electrical load,six load characteristics were extracted.⑶ The CUSUM algorithm based on sliding window is used to detect the power load switching event.The effects of sliding windows at different locations on the event detection algorithm are analyzed.The setting of the threshold of the cumulative sum of the detection algorithms is improved.The effectiveness of the algorithm is verified by the measured load switching data.⑷ The mapping relationship between load characteristics and load types is studied,and the electric load identification model based on ELM is established.Based on the selection and analysis of model sample data,activation function and related parameters,the power load identification algorithm based on the model is designed.The simulation results show that the algorithm can identify the load type well.