Research On Wind-Photovoltaic-ESS And Integrated Energy Robust Planning Based On Confidence Gap Decision

In the era of energy conservation,emission reduction and energy sustainability,a large amount of distributed renewable energy sources(RESs)like wind power and photovoltaic power have been integrated into the power grid on a large scale.However,the intermittency and uncertainties of RESs affected by the environment may have detrimental influence on the stable operation of grid,which will restrict the consumption of renewable energy,bringing serious challenges to the planning of wind-photovoltaic-ESS and integrated energy with high proportion of RESs.Therefore,a series of effective uncertainty handling methods are widely used in the fields of wind-photovoltaic-ESS and integrated energy planning.Robust optimization(RO),stochastic programming(SP),distributionally robust optimization(DRO)and information gap decision theory(IGDT)are the main methods to cope with this planning problems under uncertainties.However,RO commonly uses a symmetrical set to describe the uncertainties,which is difficult to reflect the polymorphism of uncertainties,resulting in rough robustness evaluation,so the result tends to be relatively conservative.SP simplifies broad uncertain intervals into several representative scenarios through clustering algorithm.Still,it is easy to lose the ergodicity of the uncertain intervals,so the characteristic of uncertain variables may not be guaranteed.DRO can achieve a certain balance between economical and conservative performances,but from the perspective of ensuring the optimal solution is feasible in the worst "fuzzy scenarios",this method is still conservative and inflexible.Similarly,IGDT uses a symmetrical box set to describe the maximum fluctuation interval of uncertainties,which is difficult to reflect the actual polymorphism of various uncertainties,and the possible nonlinear relationship between robustness and the upper and lower bounds of the uncertain intervals.In order to compensate for the deficiencies in above methods,multi-scenario clustering analysis and classified confidence intervals of Gaussian mixture model are combined to describe the enormous uncertain scenarios,so the typical multi-scenario deterministic planning is generalized to a multi-scenario confidence interval planning.Furthermore,the robustness idea of IGDT is incorporated,so a novel multi-scenario confidence gap decision theory(MCGDT)is proposed.MCGDT aims to obtain the maximum confidence intervals of uncertainties under the condition that the probability of objective deviation is as small as possible,thus mitigating the detrimental impact of uncertainties to the greatest extent.MCGDT can measure the nonlinear robustness and fully reflect the distribution characteristics and polymorphism of various uncertain variables,without a preset objective deviation factor,so a more reasonable and accurate robust planning can be achieved.Accordingly,confidence gap decision is the theoretical basis in this paper,and in the context of dealing with the uncertainties of RESs,the application of confidence gap decision in the field of wind-photovoltaic-ESS and integrated energy planning has been investigated,and three research contents have been formed: multi-objective robust optimization allocation for energy storage,robust joint planning of Wind-Photovoltaic-ESS and robust capacity planning for IES.However,the established models are very difficult to solve since they come down to a non-convex and non-linear optimization problem,especially with conflicting multiobjective functions.Therefore,inspired by the cross entropy-importance sampling and radar scanning mechanism,a new cross entropy-radar scanning differential evolution algorithm is designed to solve the model,also with the aim of breaking through the common bottleneck of optimization depth and speed of conventional multi-objective swarm intelligent optimization algorithms.Finally,the validity and superiority of the proposed theory and method are verified through case analysis.The research achievements can provide new ideas for power planning under uncertainties and swarm intelligence optimization algorithms.Moreover,MCGDT will also be in favor of optimal scheduling of smart grid and planning of active distribution network with high proportion of RESs,and can also be applied to other uncertainty optimization fields.