| Micro-grid as a useful complement to large power grid has a wide range of applications.The micro-grid integrated load forecasting is an important part of the micro-grid energy dispatching.The forecast result directly affects the source-load balance of the micro-grid.In this paper,the micro-grid phasor measurement unit is developed for the comprehensive load forecasting of micro-grids.The data collected in the micro-grid of the system is used for the experimental verification of the electrification load of electric vehicles and the daily load forecasting model of micro-grid.The two model constitutes a micro-grid integrated load forecasting model.For micro-grid condition monitoring and data acquisition,a micro-grid phasor measurement unit was developed.Design data collection,board and transmission terminal and debug.Acquisition board will collect the data package,through the CAN bus to upload data to the transmission terminal RTU,RTU through the TCP/IP protocol data transmission to the cloud display and storage.The measurement accuracy of the system by Zhejiang Institute of Metrology can reach 0.5 S level,the basic power parameters of the measurement accuracy of 1 level;measured by the FLUKE6100A harmonic error of the system within 1%.Aiming at the problem that electric vehicles gather load into micro-grid and lead to load balance,it is difficult to balance the source load.A load forecasting model of electric vehicle charging station based on copula algorithm is proposed.Firstly,the users are classified.The copula function is selected according to AIC and BIC criteria.The nuclear density function is used to fit the driving law data of electric vehicles.The copula algorithm is used to derive the augmented data including the coupling relationship between the data and the EV load forecast curve.Compared with the Monte Carlo algorithm,the prediction curve obtained by this model is more realistic.According to the daily load forecast of micro-grid,this paper proposes a daily load forecasting model based on empirical mode decomposition,support vector machine and particle swarm optimization algorithm.This model first uses EMD to decompose the micro-grid load time series Step by step into multiple natural modal components,using support vector machine to predict multiple natural modal components respectively,and using particle swarm optimization algorithm to optimize the model parameters.According to different types of micro-grid load data to verify the model,the maximum error of 13.52%,under normal circumstances are less than 10%.In this paper,a phasor measurement unit is developed,and the charging load and daily load forecasting model of the micro-grid are proposed.The model is conducive to achieving the source-load balance of the micro-grid. |
