Non-deterministic Single-/bi-level Programming Methods For Multi-scale Energy Systems Management Under Carbon Mitigation Target

Energy is an important material for supporting the development of modern society.With the proposal of the "dual-carbon" goal in China,the development of clean,lowcarbon,safe and efficient modern energy system has been more and more important in energy system design.In order to implement the "dual-carbon" goal,it is necessary to optimize energy systems at different scales.Energy system is a huge system which involves multiple sectors,multiple users,and hierarchical structure,leading to various complexities and multiple uncertainties.Thus,for achieving the modern energy system,it is of great importance to descript the complexities and uncertainties,find the impacts of different emission mitigation measures on energy system,and propose clean,lowcarbon,robust,and reliable development plans for multi-scale energy systems.In response to the above issues,this thesis developed non-deterministic singlelevel programming methods and non-deterministic bi-level programming methods to deal with system complexities and uncertainties such as intervals,random numbers,fuzzy sets,and hierarchical decision-making structure.Then,based on the developed methods,a series of models for energy system management have been established at municipal,provincial,and national scales.Specifically,it includes:(ⅰ)Developed a full-infinite interval two-stage credibility constrained programming(FITCP)method to handle multiple uncertainties represented as functional intervals,crisp intervals,fuzzy sets,and random numbers.Based on FITCP,a power system model has been formulated,in which carbon emission trading scheme(CET)has been introduced to cope with the problem of carbon mitigation.The results show that the CET scheme can bring more economic benefits for power plant.(ⅱ)An interval-fuzzy full-infinite programming(IFFIP)method has been developed to handle the uncertainties represented as crisp intervals,functional intervals,and dual-functional intervals.Then,an IFFIP based energy system management model has been formulated for planning Hebei energy system through considering energy transitions in industrial,transport,and residential sectors.Scenarios related to coal/oil substitution in final energy demand were designed to evaluate the impact of final energy transition.The results reveal that final energy substitution would bring about 8.9%reduction in the share of coal in energy supply over the planning horizon.(ⅲ)An interval-parameter bi-level programming(IBP)method that can be effective for solving hierarchical decisionmaking problems with conflict objectives and uncertainties has been developed.Then,an IBP based China’s energy system model(abbreviated as IBP_CES)has been formulated for mainland China during 2021-2050.In IBP_CES,the higher-level objective of minimizing carbon dioxide emission is given priority,followed by the lower-level objective of minimizing system cost.Results from IBP_CES show that the share of coal would decrease to 33.9%by 2050,the share of non-fossil energy would grow to 38.5%,and the electricity supply would increase by 30.0%.(iv)Developed a leave-one-out based stepwise cluster analysis-factorial analysis(LSCA-FA)method for the energy consumption simulation,and analyzed the changes and spatial distribution of energy consumption in China under different scenarios.Then,a bi-level joint-probabilistic programming(BJPP)method has been developed for planning multi-regional energy system under different mitigation policies and uncertainties.BJPP can handle leader-follower issues in decision-making process as well as examine the risk of violating joint-probabilistic constraints.Based on the BJPP method,a China’s multi-regional energy system(named as BJPP_CMES)model has been formulated to provide optimal schemes for energy system planning of China over a long-term horizon(2021-2050)by synergistically minimizing carbon dioxide emission and system cost.A series of scenarios associated with different carbon capture and storage(CCS)levels and violation risks of energy-demand constraints are examined.Results reveal the highest growth of the renewable supply would occur in Ningxia(rising 47.7%);Sichuan,Inner Mongolia,and Gansu would be the top suppliers of hydro,wind and solar electricity,respectively;Hebei and Shandong would be main contributors of CO2 emission in the future.Summarily,this thesis has developed a variety of non-deterministic single-/bilevel programming methods by incorporating interval-parameter programming,fullinfinite programming,stochastic programming,fuzzy programming and bi-level programming methods,which can deal with uncertainties represented as intervals(i.e.crisp intervals,functional intervals and dual-functional intervals),random values and fuzzy sets,as well as handle the conflicts between decision makers of two levels.Based on the developed methods,a series of multi-scale energy system management models have been formulated to explore the impact of different emission mitigation measures on energy systems,provide sustainable development plans under multiple uncertainties,and give theoretical guidance and technical support for the modernization and transformation of energy systems.