Energy Supply-Demand Forecasting And Optimal Operating Techology Of Local-area Integrated Energy System

The energy industry has under gone great changes as a result of the rapid development of technology and the economy since the beginning of the twenty-first century.Exhaustion of fossil energy,environmental pollution,and large-scale renewable energy applications have proven to be the key drivers of the development of the modern energy system which must satisfy various requirements concerning low-carbon emissions and efficient utilization of friendly sustainable energy sources.However,different energy systems have been designed,planned and operated independently in traditional style,in an approach that neglects the coupling correlations relating different types of energy and greatly limits the flexibility of energy system.In this case,integrated energy system emerges as the times require.As a indispensable element of energy internet and ubiquitous power internet of things,local-area integrated energy system covers electricity,heat and gas sub-system is deployed to realize the coupling of different types of energy sources from the aspect of source,network and load module,which apply several advanced techology,including power electronics,information exchange and energy mangagement system and etc.Thus,people have attached great attention to this field by noticing the outstanding advantage,such as,operational flexibility,low-carbon,high-efficiency of renewable energy usage.This dissertation studied the energy supply-demand balance mechanism and operating strategy of local-area integrated energy system from aspects,including the related standards of microgrid,the forecasting technology of energy supply and consumption,the calculation method of multi-energy flow,and the dispatching optimization method of local-area integrated energy system.For supply-side and demand-side forecasting issue,the accurate forecasting has significant impact on dispatching and optimal operation of integrated energy system.Taking distributed PV forecasting as example,the PV power forecasting method based on feature selection and deep learning algorithm is proposed in this dissertation.Considering that the gains in the random forest,the features are selected from input information.Deep architecture of PV power forecasting mode is introduced which learns the structure internal characteristics by deep belief network(DBN)based on restricted boltzmann machine to extract deep features as unsupervised learning process,and the supervised BP neural network is used as fitting layer.The algorithm validity and accuracy of power forecast approach for PV system are verified by the simulation using actual operating data of PV system.This dissertation proposes a method of short-term hourly load forecasting for various energy sources based on deep multi-task learning.The algorithm architecture is introduced which consists of a DBN at the bottom and a multi-task regression layer at the top.The DBN can extract abstract and effective characteristics in an unsupervised fashion,and the multi-task regression layer above the DBN is used for supervised prediction.Then,a two-stage load forecasting system based on off-line training and on-line prediction is deployed subject to the condition of practical demand and model integrity.The validity of the algorithm and the accuracy of the load forecasts for integrated energy system are verified through the simulations using actual operating data from load system.For electrical load forecasting issue,the progress of machine learning and artificial intelligence technology provides an effective approach to improve forecasting accuracy.A load forecasting method based on multi-model combination under Stacking framework is proposed in this dissertation,associated with the frontier theory research of artificial intelligence.Firstly,the mechanism of Stacking ensemble learning is introduced.The XGBoost algorithm constructed by tree model and the deep learning algorithm represented by long and short memory network are presented.Then considering the difference of data observation and training principles,the Stacking based load forecasting model embedded various machine learning algorithms is proposed to utilize their diversified strength.Finally,we use the Swiss load data in ENTSO to verify the effectiveness of the algorithm.The use case shows that the XGBoost,the gradient decision tree and the random forest model can quantify the contribution of the input data through the gain of their own model.The load forecasting results are more accurate when each of base-learner has lower correlation coefficient in Stacking.The results indicate the Stacking ensemble learning based on multi-model has better prediction performance compared with the traditional single model.For consecutive daily peak load issue,the load forcasting model based on sequential-parallel ensemble learning is proposed considering the forecasting error distribution of statistical learning algorithms.Firstly,the decoupling process of generalization error is analyzed,the mechanism of XGBoost sequential ensemble learning,bagging parallel ensemble learning and particle swarm optimization are introduced.Then,the load forecasting based on XGBoost algorithm under bagging framework are deployed,associated with the distribution of bias and variance at the training stage.The PSO is used to cross validate the hyperparameters of XGBoost model.Finally,the effectiveness of the algorithm is verified by using EUNITE competition data.The results show that XGBoost algorithm calculates the characteristic importance to aid to select the valuable features.The particle swarm optimization algorithm effectively shorten duration of hyperparameters optimization.And the Bagging-XGBoost algorithm has better forecasting accuracy compared with several conventional models.The forecasting results indicate that sequential-parallel ensemble learning method have higher application value in engineering application for load forecasting.For multi-energy flow calculation issue,this dissertation proposes an extended Newton-Raphson multi-energy flow calculation method,which is suitable for integrated energy system containing electricity,heat and gas.First,the models of electricity network,heat network and natural gas network in integrated energy system are established.In view of the complexity of traditional method to solve the gas network including compressors,an improved practical method is put forward based on different control modes.On this basis,the multi-energy flow model of integrated energy system is built and the Jacobi matrix reflecting the coupling relationship of multi-energy is derived considering the grid connected and island mode of power system.The validity of the proposed method in multi-energy flow calculation and analysis of interacting characteristics is verified by numerical cases.For optimal dispatching issue,a basic framework of integrated energy system with various energy-hub(EH)is proposed,which matches the scenario with multiple autonomous decision-making energy entities.Then operation characteristics and operation constraints of EH and integrated energy system are analyzed,and a day-ahead low-carbon economy scheduling model for distributed integrated energy systems is established based on ADMM(alternating direction method of multipliers).Finally,simulation is carried out by using practical data from several aspects,including operation of sub-region energy-hub,scheduling results of energy storage equipment,effects of carbon emissions punishment for system,and validity of distributed algorithm,verifying practicality and accuracy of the proposed model and method.