Research On Energy Management Of Multi-Microgrid Considering Uncertainties

With the increasing penetration rate of renewable energy,our country’s energy structure has been continuously optimized,providing a solid foundation for the realization of the "carbon peak" and "carbon neutrality" proposed by the Central Committee of the Communist Party of China.As a highly flexible power grid topology,microgrid can organically integrate regional loads and distributed renewable energy sources to achieve efficient utilization of clean energy and energy balance locally,which is the key to relieving the scheduling pressure of the main grid.Considering the randomness and volatility of composite factors such as local load,renewable energy and electric vehicles,the stable economic operation of microgrids will be challenged.In order to achieve the balance between the economic and environmental benefits of the microgrid considering multiple uncertainty scenarios and improve its comprehensive interests of society,this researcch carries out the following work:Firstly,a multi-objective optimization model based on charging load prediction is studied for the energy management problem of multi-microgrid with electric vehicles.The prediction module set the time series and temperature as the associated variables of the electric vehicle charging load,then an adaptive long short-term memory deep learning method is proposed to analyze and correct the error curve provided by back propagation neural network training model.A two-layer optimization model including single/multi-microgrid structure is proposed.According to the final prediction scenario,a decomposition optimization method based on the improved Chebyshev framework is adopted at the multi-microgrid layer,while minimizing the system operating cost,network line loss and carbon emission.After a multi-model approximation for the characteristics of micro turbine,an improved consensus algorithm is applied to achieve fast economic optimization at the single microgrid layer.The simulation results demonstrate the superiority of the global search performance and the rapid convergence performance of the proposed improved algorithm.Secondly,in order to solve the multi-period coupling robust optimization problem of electric vehicles in the microgrid system,a two-layer multi-microgrid optimization model based on an improved data-driven approach is proposed.This model includes continuous uncertainties of load and wind power,and integer uncertainties of multi period electric vehicle grid connection quantity.The multi-microgrid layer adaptively iterates the charging constraints according to the information transmitted by the electric vehicle aggregator layer,so as to effectively avoid safety limit exceeding events.A segmented model with multi-period coupling uncertainties is established to obtain a more practical energy management model.In the energy management layer of the electric vehicle aggregator,a multi-objective optimization model is established according to the specific electric vehicle charging characteristics and duration.Then a data-driven uncertainty set construction method based on neural network framework is proposed,which learns from historical data and locate the feature scene set.In this way,fewer scenes can be used to obtain reliable schemes that are highly similar to the operation results based on all historical data,so as to improve the operation performance.The simulation results demonstrate that the model can provide safer and faster energy management solution.Thirdly,in order to broaden the application scenarios of microgrid and improve the solution efficiency of energy management models with multiple types of variables,a optimization model of the combined cooling,heating and power microgrid is proposed.A beta function-based double-layer method for forecasting wind power output and calculating the expected error value is proposed,which can predict the uncertainty of wind power output more accurately and provide a quantified forecast error cost according to the local system environment.A multi-objective energy management model of double-layer combined cooling,heating and power microgrid based on time-of-use electricity price is proposed.And the minimization of operating costs,pollutant emissions and unit asynchrony are set as objective functions.It’s solved by a non-dominated sorting algorithm improved by co-evolution theory and bionics.The solution set is divided according to the operating characteristics of different operating units.Interleaving learning and recording information between individuals of different populations is conducive to preserving ordinary individuals with adaptive potential in the early stage of iteration,and effectively improves the global search ability and convergence performance of the algorithm.Finally,a hierarchical two-stage robust optimization model is proposed for the multi-region combined cooling,heating and power microgrid without sufficient historical data.This model considers continuous and binary uncertainties such as renewable energy,power load,heating load,cooling load and line fault,and obtains a unified uncertainty set based on Mc Cormick envelope.An modified column and constraint algorithm combined with co-evolutionary theory is proposed to improve the solution efficiency.According to the unit constraints and line constraints,the multi-region microgrid system is divided into several blocks,each of which can be expressed as a separate min-max-min mixed integer nonlinear programming model.The rationality and validity of the proposed model and the superiority of the solution performance of the improved algorithm are verified through simulation case studies involving a system composed of four microgrids.By constructing microgrid energy management models under different data scenarios,this paper solves the optimal operation problem of microgrid considering various uncertainties.The proposed measures can effectively improve the model solving efficiency and obtain more economical and reliable operation solutions,providing theoretical support for the safe,economic,and efficient operation of microgrid in green and low-carbon environment.