Decision-making And Optimization Strategies For Power System Restoration

Modern electric power systems are gradually transforming to large,interconnected complex systems.In general,the interconnected large-scale power systems demonstrate advantages in power generation complementation,power system stability and power supply reliability.However,modern power systems are vulnerable to incidents such as human misoperations and extreme weather conditions.If handled improperly,these incidents will possibly lead to cascading failures and further develops into major blackouts,thus the major blackouts are somehow inevitable considering all kinds of unexpectable factors.Although major blackouts are rarely seen around the world,they normally cause enormous economical losses and social impacts once occurred.Thus,a comphrehensive,systemetic and in-depth study in power system restoration after major blackouts will help reduce the negative impacts of power outages,and contribute to reliable power supply to customers significantly.In this context,this dissertation divides the power system restoration process into three phrases,namely black-start,network reconfiguration and load restoration,and investigates the decision-making theories and optimization methods in these phrases concerned.Specifically,the main focuses of this dissertation can be summarized as follows:1)Analysis and optimization methods for black-start group decision-making problems.The proposed planning strategy would offer the investor an opportunity to make a proper coordination between expansion planging of distribution systems and deployment of EV charging infrastructures.An intuitionistic fuzzy distance based black-start decision-making consistency analysis method and a clustering black-start decision-making algorithm are proposed based on the size of decision makers,respectively.To optimize the consistency of black-start group decision-making results,the weights of indexes as well as those of experts are adjusted after each round of decision-making is carried out,until a decision-making result with an acceptable con-sistency is achieved.When the decision-making group becomes very large,group decision-making preferences are obtained through clustering analysis,and the final black-start decision-making results are achieved by combining the weights of black-start indexes and the preferences of the decision-making group.The proposed consistency analysis method and clustering algorithm are able to overcome the possible inconsistence in the decision-making results caused by the increase of members in the decision-making group.2)Sectionalizing optimization strategies for local power system restoration and parallel power system restoration problems.First of all,the characteristics of local power system restoration and parallel power system restoration are discussed,and restoration paths are generated based on the topology of the outage area.Secondly,the restoration paths causing over-voltage problems as well as the unnecessary paths are removed.Finally,an optimization model is developed for determining restoration subsystems and restoration paths based on the generation capacities and characteristics of important loads.Compared with most existing methods in which subjective weights are required for the lines and buses to be restored,the proposed method can attain more reasonable restoration strategies.Thus,the outage durations of important loads and non-black-start generating units will be effectively reduced based on the developed model.3)A distribution system restoration strategy with intermittent distributed energy sources and controllable loads.A restoration model for the isolated island with distributed generation,energy storage,electric vehicles and controllable loads is established first,with an aim of providing continuous and stable power supply to the maximized amount of power load in a given period.The characteristics of different kinds of intermittent energy resources and their output capabilities are discussed next,and the controllable loads are employed to balance the power output fluctuations of these intermittent energy sources.Probabilistic models are utilized to describe the uncertainties of the predicted power outputs of the intermittent energy sources.The proposed distribution system restoration strategy is effective in formulating reliable restoration islands and providing sustainable electricity supply for customers while dealing with DER forecast errors.4)A multi-objective optimal strategies for power system restoration with support from electric vehicles and energy storage systems.Increasing capacity of renewable energy sources and active loads such as electric vehicles and energy storage systems are being integrated intothe power system.In this context,the possible functions of electric vehicle charging/swapping stations and energy storage systems in supporting power system restoration are discussed first.Then,a multi-objective optimization model is formulated with three objectives,i.e.,minimizing the number of restored transmission lines,minimizing the outage durations of the non-black-start generating units and important loads and minimizing the voltage deviations of the restored power network.The proposed multi-objective optimization model will be able to accelerate the power system restoration process and fulfil the other objective concerned in the model by employing flexible charging/discharging operations of the electric vehicles and energy storage systems.5)A load restoration method based on system frequency response and discrete load restoration.First of all,the load restoration model based on system frequency response characteristics is proposed,considering the power system dynamic model,the characteristics of generator ramping,frequency response of generators and power loads.Besides,a restorable load estimation method is proposed employing system frequency data after the network frame has been reenergized,and the discrete restorable loads are employed considering the limited number of circuit-breaker operations and the practical topology of distribution systems.To enhance the estimation accuracy,an estimation error feedback method is adopted using the error between the measured data and calculated data.The proposed discrete load restoration model can make full use of the generator power output capacity during ramping stage,while maintaining the power balance and stability of the system at the same time.The proposed decision-making algorithms and optimization strategies have enriched the state-of-the-art theoretical system and framework for power system restoration problems,and can further provide decision and technical support for power system dispatchers to accelerate restoration speed in case of a major blackout.