Research On Photovoltaic Power Forecasting And Microgrid Energy Management

As a stable distributed generation(DG),microgrid(MG)can not only be connected to the large power grid and run in parallel,but also operate in isolated islands when off-grid.When microgrid is connected to the grid,it can not only ensure the stable operation of the power grid,but also effectively improve the utilization rate of renewable energy.It is particularly necessary to study the microgrid technology,reduce the overall operation cost of the microgrid,and establish the energy management system(EMS)of the microgrid.This paper mainly focuses on the research of photovoltaic(PV)power forecasting and energy management system based on solar/storage/diesel.Firstly,based on the traditional RBF neural network PV Power Forecasting,the input layer data such as module selection and module cleaning frequency are added to correct the traditional prediction method by taking into account the factors affecting power generation in the actual operation process of PV power stations.At the same time,the idea of combining cloud image with RBF neural network is put forward to make very short-term PV Power Forecasting through more accurate surface radiation,and its accuracy is verified by practical engineering cases.Then,the paper constructs the hybrid energy storage model of lithium iron phosphate battery and Ultra—Capacitor,analyzes the management strategy of hybrid energy storage,ensures that lithium iron phosphate batteries always runs in the optimal SOC region,and ensures the economy of the energy storage system.Finally,this paper studies the energy management strategy under grid-connected and islanded modes,analyzes the operation characteristics of photovoltaic,hybrid energy storage and diesel generator,establishes the objective function with the lowest operation cost of the microgrid system based on soalr/storage/diesel,proposes the management strategy under different operation modes,and obtains the optimal scheduling control model through the solution of genetic algorithm,and verifies that the management strategy in this paper can not only help outer grid with “peak clipping and valley filling”,but also realize the reliable and economic operation.