Study On Distribution System Optimal Operation And Service Restoration Methods Based On Distributed Generators

Due to the continuous exploitation and consumption of fossil energy such as oil and coal,people are facing increasingly serious energy shortage which lead to environment degradation and frequent occurrence of extreme disasters.With the development of distributed generation technology,distributed generations(DGs)are widely connected to the distribution system with the advantages of high reliability and flexible installation.On the one hand,DGs can effectively utilize renewable energy and alleviate energy shortage to some extent.On the other hand,the rational use of DGs can enhance the load restoration ability of distribution system after extreme events.However,the inherent intermittency and fluctuation of renewable energy brings challenges to the safe operation of distribution system.Meanwhile,the unbalanced characterize of distribution system also increases the complexity of the distribution system load restoration model considering DGs.Therefore,this thesis studies from the optimal operation and load restoration of the distribution network aspects,and proposes the day-ahead optimal dispatching model of the park distribution system considering the uncertainty of photovoltaic(PV)and the distribution system load restoration model considering unbalanced power flow.The main work of this thesis is as follows:Firstly,based on the assumption of small line loss and balanced node voltage in distribution system,the single-phase and three-phase unbalanced linear power flow model are derived,which can be easily applied to the optimal operation and load restoration in distribution system.Numerical simulations show that the linear approximation method has high accuracy for systems with smaller network loss.Then,on the basis of the single-phase linear power flow model,using gaussian mixture model to describe uncertainty of PV and a chance constrainted programming model for the day-ahead optimal dispatching of the park distribution network is established,which takes into account the uncertainty of PV and the probability of voltage over-limit.Then utilizing the sample average approximation to transform the opportunity constraint into the deterministic constraint and a mixed-integer linear program(MILP)for the day-ahead optimal dispatching of the park distribution network is established.Case studys show that the GMM has a good fitting effect on PV,and the day-ahead optimal dispathching model can maximize the economic benefits of the park distribution system under the condition of satisfying the safe operation of the system.Finally,based on the three-phase unbalanced linear power flow model,a mixed-integer linear program for distribution system load restoration model is established.Considering that the failure may last for a long time,in order to ensure that the critical load can supply power for as long as possible and achieve the optimal allocation of local limited power generation resources,the one-snapshot restoration model is expanded to a multi-period model.The simulation results show that the linear model has higher accuracy in the system with small network loss.By adjusting the output of DGs in each period,the critical load can be guaranteed to be supplied continuously in the whole restoration period.