Hybrid-time Scale Optimal Scheduling Of Integrated Energy Systems With Load Forecasting

With the rapid development of industrialization and the improvement of people’s living standards,the demand for production has increased sharply around the world,and the problems of energy supply and environment have become increasingly clear.How to efficiently utilize the system clean resources,improve energy efficiency,develop new energy technologies and promote the transformation of energy structure has become the focus of research at home and abroad.The integrated energy system covers a variety of energy sources,which contains cold,heat,electricity and gas,which can achieve the deep coupling of different types of energy,promote the rational use of energy resources.However,compared with the traditional single power system,the load of the integrated energy system and the output of renewable energy are susceptible to the influence of climate and environmental factors,and the random fluctuation is higher.Besides,the coupling equipment in the system operates under various conditions,and the coupling characteristics of different energy and flow are relatively fuzzy.Therefore,the system needs to be optimized and analyzed in different aspects.To sum up,based on the multi-energy flow integrated energy system as the research object,deeply analyzes the coupling relationship between different sources,and discusses the load forecasting for various energy sources,hybrid-time scale optimization scheduling and other issues in the system.First of all,accurate load forecasting is an important part of the system design,planning,and its operation.Due to the strong random fluctuation of load,the instantaneous features of nonlinear and non-stationary signals can be obtained by adding data decomposition method,CEEMD,in the process of load forecasting.Pearson correlation coefficient was used to analyze the influencing factors,and the influencing factors with high correlation degree were selected as the input for the comprehensive energy load prediction.The high precision load prediction value is obtained by training and learning the network model of LSTM,which contains special memory structure and gate structure.In addition,the correlation characteristics of the data are analyzed.And the effectiveness of CEEMD decomposition method and LSTM network model is’verified.Second,the demand side response can realize the supply side and demand side of the bi-directional interaction,which can reduce the impact of the new energy output randomicity,and further reduce the operation cost of integrated energy system.This paper models price-based demand response and substitute-based demand response.In terms of price-based demand response,price adjustment is carried out at different times by adding the elasticity coefficient of load electricity price and gas price,so that users can reasonably adjust their demands according to different prices.In terms of substitute-based demand response,users can according to oneself circumstance,convert different energy sources through system equipment,so as to improve energy utilization efficiency and ensure users requirements.Finally,the dynamic characteristics of different energy sources are different,so the corresponding scheduling time is obviously different.This paper proposes a day-ahead,intraday and real-time optimal dispatching model.In the intraday dispatching layer,price-based demand response is added.At the real-time dispatching level,substitute-based demand response is added on the basis of intra-day dispatching plan.Hybrid-time scale optimal scheduling strategy formulation can maximize the revised schedule plan,in order to cut the influence of uncertainty factors step by step.In addition,this way can also improve the access capacity of new energy and achieve the coordinated operation of system network and equipment,at last to ensure the dynamic demands of end users.