Research On Active Distribution Network Reconfiguration And Fault Recovery Based On Short-term Load Forecasting

The reconfiguration of distribution network takes advantage of the opening and closing state of switches,recombination of these switches so as to make the distribution network operate in the best state,thus providing an important assurance for the normal operation of distribution network to reduce network loss,improve the operation economy,and fast recover power supply and keep continuity after fault.In recent years,wind power and photovoltaic penetration of distribution generation in power grid is increasing year by year,and the demand for load is also uncertain,therefore,through the optimization of network reconfiguration is of great practical significance to improve the stability of network operation and to solve the problem of the important load power supply after failure.This paper aims to research on two aspects which are active network reconfiguration optimization and fault recovery problem containing distribution generation and load uncertainty,the following work has been carried out:1)Establishing the model of the uncertainty of wind power and photovoltaic power generation,In view of the uncertainty factors in the process of power flow calculation,proposal two points estimation to improve the efficiency of power flow calculation.Considering the uncertainty of load power in the study of normal operation reconfiguration and dynamic fault reconfiguration through the improved shuffled frog leaping algorithm to optimize the parameters of the least squares support vector machine,Establishing a more accurate short-term load forecasting model.2)Research on the wind power,photovoltaic and load uncertainty ofactive distribution network reconfiguration,For load uncertainty,through the information entropy to make the optimal scenario partition number.By using Wasserstein probability distance metric to deduce the fractile formula of photovoltaic,through the existing probability fractile formula model of wind power,a multi-scenario based on the expected network loss value of active distribution network was established.This paper proposes“no-repeat” coding strategy,combining with the improved firework algorithm to optimize the scenario that simultaneously adapts to source and load uncertainty.3)In the process of active distribution network fault recovery reconfiguration,the problems is divided into two parts,that is,islanding operation and residual network reconfiguration,division the isolated island by the maximum power of the distributed power at each time interval.Islanding partition model is established under minimizing the line loss as far as possible circumstance to maximize load recovery.In order to divide the dynamic isolated island region,the depth priority search algorithm is adopted to optimize the solution,and the regional fault recovery scheme and global recovery scheme are coordinated by the multi-agent method.Established a reconfiguration model of considering active power loss for the main distribution network system outside the island.Application of improved firework algorithm to solve the recovery reconfiguration of distribution network outside the island.