Research On The Optimal Management Of Energy Storage Based On Distributed Energy Output And Load Forecasting

In the context of the rapid development of energy Internet technology,energy management,energy technology and supply are becoming more and more efficient and clean with the production,allocation,conversion and use of various energies.At present,domestic energy development is under pressure from many aspects,which not only comes from the energy revolution of external environment,but also from the environmental pressure of regional development.In particular,it is of great theoretical significance and practical value to combine energy development with domestic urbanization,economic adjustment and social development and develop sustainable distributed energy systems to improve energy use efficiency,reduce environmental pollution and reduce carbon emissions.In view of this,this paper firstly holds the research background and significance of the subject,summarizes the research results at home and abroad,and lays a theoretical foundation for the subject research.Secondly,the output model of photovoltaic power generation and wind power generation,two typical nonlinear output energies,is constructed,and their economic and environmental benefits are analyzed and evaluated.Moreover,moreover for distributed energy power load forecast,Outlines the electricity load cycle,for periodic electricity load,it can be divided into cyclical in week,day as the unit for the unit with the seasonal periodicity and cyclical,selection method of level processing power load data and error correction,and for distributed energy electricity load forecasting of the main path selection and data error correction model to build.Again,using the neural networks to forecast the power load,in a school building as an example,discusses the neural network prediction model based on WNN the principle,structure of the model are analyzed,and the process of training to detail,finally,to a school laboratory as an example,through April 2019 to collect relevant data,analyzes in the school laboratory in April 2019 days and holidays on photovoltaic power,wind power output data such as parameters,and by using the WNN data for training,building electricity load forecasting model,through the distributed energy WNN forecast results and verification,it is concluded that,WNN can effectively optimize the power load,and can accurately predict the distributed energy power load.The predicted value of the network is infinitely close to the expected value.It is believed that with the continuous increase of training times,the model prediction accuracy will be further improved.In order to ensure the optimization of distributed energy storage,this paper puts forward the energy storage optimization management strategy of opportunity load and distributed energy storage prediction from three aspects:technology supporting energy storage optimization,multiple fusion supporting energy storage optimization and method supporting energy storage optimization.