Research On The Sequential Service Restoration Of Resilient Distribution Systems In New Power Systems

The frequent occurrence of extreme natural disasters has brought great challenges to the safe operation of power systems.Especially,distribution systems(DSs)are extremely vulnerable to natural disasters due to fragile grid infrastructures and limited back-up resources,which is the main reason for large-scale power outages.To reduce tremendous socioeconomic losses caused by outages,emergency control of DSs under natural disasters has become a focused issue and trend.The traditional emergency control methods restore critical loads in on-outage arears by changing the network topology,which cannot address the outage caused by failures in main grid.With the development of renewable sources and the proposal of carbon peaking and carbon neutrality goals,the construction of new power systems is being accelerated.Meanwhile,a large number of distributed generators(e.g.wind turbines,energy storages and mobile emergency generators)are integrated into the DS.In this context,the emergency control strategy based on resilient distribution system has been proposed.When an outage accident occurs in the main grid,isolated microgrid(MGs)are formed and distributed generators are used to provide emergency power supply to critical loads,which is of great significance to reduce the scope and shorten the duration of power outages,as well as improve the reliability of power system operation.One key to enhance the resilience of DSs under emergency control condition is to form isolated MGs.A large number of studies have been undertaken to sectionalize the DS into MGs and optimally dispatch energy among MGs.However,MGs may fail to form due to the intermittence of renewable sources and loads.Therefore,for the areas that MGs failed to form,some researchers proposed to restore the DS sequentially after outages.The sequential service restoration(SSR)procedure is to dynamically expand scope of MGs by generating sequential operating instructions for each device.By this means,isolated MGs are formed and critical loads in on-outage areas are picked up.The existing SSR method utilized deterministic power sources such as micro turbines and diesel generators to restore the system.However,such methods cannot be applied to the SSR of DSs in new power systems due to the massive integration of renewable energy sources.To solve the problem,this dissertation studies the SSR methods of resilient DSs in new power systems.Series of studies take into account several characteristics of new power systems including the large number of local distributed generators,uncertainties of renewable sources,and great diversity of power sources.The main contents and contributions are summarized as follows:1)Considering that a large number of local distributed generators are integrated into new power systems,this paper proposed a unified SSR decision method that coordinates the optimization of network sectionalization,restoration paths and restoration status.First,restoration path is modeled by introducing the “virtual energization agent”.Then,the SSR problem that considers the MG sectionalization optimization is modeled as a mixed integer nonlinear programming(MINP)model.Furthermore,the nonlinear constraints are linearized to facilitate solving.The numerical results demonstrate that the proposed SSR method can restore multiple isolated MGs in parallel,and achieve better restoration efficiency compared to the existing SSR methods.2)To deal with the uncertainties on both supply and demand sides in new power systems,this paper proposes a tri-level robust SSR method that considers the source-networkload-storage coordination.First,the wind power generation and load demand are modeled as a polyhedral uncertainty set.Then,a tri-level robust model is formulated to coordinately dispatch power sources,network,loads,and energy storage units.To reduce the computational complexity,the original model is relaxed to reduce the number of integer variables.Then the extended column-and-constraint(EC&CG)algorithm is employed to solve the modified model iteratively.The numerical results demonstrate that the proposed method can guarantee the security of MG operation and achieve a less conservative solution compared to traditional robust optimization method.Moreover,the proposed solution algorithm can reduce the computational complexity when guarantying the exactness of solutions.3)Considering the wide application of mobile emergency generator(MEG)in new power systems,this paper proposes a pre-and post-disaster two-stage MEG dispatching model for the SSR of DSs.In the preventive control stage,uncertain contingencies are considered,and a scenario-based stochastic model is proposed to pre-position MEGs prior to event strikes.In the emergency control stage,wind power uncertainty is taken into account,and a tri-level robust model is formulated to optimize the MEG allocation decisions and SSR schemes.Moreover,the progressive hedging algorithm(PHA)and disjunctive programming-based column-and-constraint(DP-C&CG)method are customized to effectively solve the two-stage models.The numerical results demonstrate that the proposed method can optimize the MEG position when sequentially restoring the DS,and the system outage duration is effectively reduced.Moreover,the proposed solution algorithm can significantly reduce the computational complexity when optimizing the bulk energy system.4)In extreme conditions,traditional service restoration strategies could pose potential security risks to restored services due to subsequent contingencies in succeeding events.To address this challenge,a two-stage SSR method that considers the formation of proactive MGs is proposed.First,considering the uncertain probability distributions of contingency,an ambiguity set is constructed that matches the N-k security criterion.On this basis,a distributionally robust(DR)optimization model is proposed with the purpose of maximizing expected load restoration with regard to the worst-case distribution of contingencies.Then,the proposed DR model is transformed into an equivalent robust optimization model,which can be effectively solved through the C&CG algorithm.Further,considering the cold load pickup issues,the SSR problem is modeled as a multitime step mixed-integer programming problem.The numerical results demonstrate that the proposed DR-based method can reduce the MG vulnerability when faults actually occur in the succeeding extreme events.Moreover,the proposed DR approach can achieve a less conservative solution compared to the robust optimization method.