Data Driven Supply And Demand Forecasting And Day-Ahead Scheduling Of Multi-energy Complementary Systems

The prediction of renewable energy output and energy demand is the core and difficulty of self consistent comprehensive energy system scheduling and operation.Accurately predicting the energy on the supply and demand sides can improve the economy of scheduling and the stability of system operation.However,the output of renewable energy has high uncertainty,and a large amount of information is contained in historical time series;In demand prediction,the impact of main activities should be fully considered for living buildings,and the laws of main activities reflect strong individual characteristics.Therefore,this article grasps the important characteristics of energy prediction on the supply and demand sides mentioned above.Based on data-driven methods,the CNN-LSTM method is applied to energy system supply side prediction,and the subject modeling method is applied to energy system demand side prediction.The main research content is as follows:(1)Renewable Energy Output Prediction Based on CNN-LSTM.Historical time series data contains information on the patterns of renewable energy output,which is a part that needs to be carefully studied in output prediction.Therefore,this article focuses on the characteristics of renewable energy output and uses CNN-LSTM to construct a model based on historical data,mining patterns while reducing model complexity.(2)Energy demand prediction based on agent modeling.The energy consumption form in living buildings is highly related to the main activities,which heavily rely on the laws of living habits and external conditions.Therefore,this article constructs an agent model that models and simulates daily life behaviors of different types of individuals,and links behavior and energy consumption,Improve the reliability and explanatory power of predictions by grasping the above characteristics.(3)Energy system scheduling optimization strategy based on energy hub method.By analyzing the scheduling results of the system,the general energy consumption patterns of the energy system are discovered.The cold,thermal,and power loads and supply sides of a multi-state energy system have various forms.Therefore,this article focuses on different seasons,considers the impact of holidays,and aims to achieve optimal economic performance.It simulates the optimal scheduling of energy systems in six scenarios,analyzes general energy consumption patterns and the effects of various factors,and guides the planning and design of new power systems.