Research On Demand Side Forecasting And Optimal Operation Of Flexibility For Regional Integrated Energy System

Energy is the basis of human survival and development and is the lifeblood of the national economy.Improving the comprehensive utilization efficiency of various types of energy and reducing pollution emissions have become the core issues to be considered in building a clean,low-carbon,safe,and efficient modern energy system in China.In order to break the existing model of independent planning,independent design,and independent operation of the traditional energy supply system,a concept of regional integrated energy system has been proposed,which aims at taking the best use of all kinds local energy resources through optimal resource allocation at all spatial-temporal scales.A regional integrated energy system often includes multiple energy carriers,e.g.electricity,heat,and gas,as well as their associated energy infrastructure ranging from energy production to distribution and consumption.Based on the application of a series of technologies related to sector coupling and electrification,such as heat pump,waste energy recovery,intelligent control and energy management,etc.,the regional integrated energy system is known for its advantages of flexible operation,low carbon efficiency and high renewable energy consumption rate.This Ph.D.thesis aims at exploring the regional g energy system theory by establishing steady-state models of electro-thermal coupling technology,developing multi-time scale demand forecasting algorithms,and designing flexible energy system scheduling schemes.The specific research work of the Ph.D.study is as follows:(1)The dissertation has developed a comprehensive multi-time scale flexibility model for the regional integrated energy system,covering various sourcenetwork-load technologies.On the source side,operation constraints of the CHP are modeled for exploring the insight of heat-power coupling principle.On the network side,both network constraints within the regional integrated energy system and the safety margins for energy exchange between the regional integrated energy system and its neighoring systems are modeled.On the load side,besides modeling the operation characteristics of various types of the heat loads like heat pump,electric boiler and heat storage,etc.,the time-lag effects of energy demand are explictly modeled.(2)A forecasting method based on time series feature decomposition is proposed for estimating the monthly electricity consumption of the regional energy system,which is sensitive to random factors.By using the decomposition method,the electricity consumption data is decomposed into a seasonal component,a trend component,and a random component.According to the time-varying characteristics of each component,three algorithms,i.e.,vector auto-regression model,least squares support vector regression and average value method are used to predict the three components respectively.The forecasted value of the monthly electricity consumption is then achieved through the projection reconstruction of the three forecasts.The proposed method considers the influence of seasonal inflection point and regional economic factors on monthly electricity consumption.The effectiveness of the proposed method is verified by an example,and the results showed that the method could improve the prediction accuracy of the monthly electricity sales of the studied system.(3)A hybrid forecasting method based on variational mode decomposition(VMD)decomposition and long short term memory(LSTM)network is proposed to forecast the daily electricity consumption of the regional energy system,which is sensitive to random factors.The proposed method uses a framework approach to effectively integrate the VMD and LSTM.The time-series data of electricity consumption is decomposed into several sub-modes through VMD.The non-linear mapping relationship between electricity consumption and the typical daily sum temperature is determined by Bayesian optimization.The original series is expanded according to the correlation,and the subsequences are predicted by the LSTM.The effectiveness of the proposed method is verified by case analysis.(4)A flexibility-based optimal operation scheme for addressing wind power curtailment in the regional integrated energy system is proposed.A variety of flexibility resources including thermal power unit,electric boiler,heat pump,and energy end-users' demand response potential are modeled and dispatched.Multiple objectives functions are included in the optimization problem with a flexibility setup,such as the system's operation cost,wind curtailment cost and cost of energy exchange,etc.while taking into account the security constraints of the power grid.The developed solution is applied to a group of scenario-based studies including seasonal variations in both energy resources and demand.The resulted variation in operation constraints and flexibility potential of various energy components,as well as the associated economic consequences,are well depicted.The results show that the proposed operation scheme has a very good effect on improving system flexibility and reducing wind power curtailment rate.