The Research On Distribution Network Reconfiguration Considering The Uncertainty Of Renewable Energy Generation

In recent years,the consumption of conventional fossil energy such as coal and oil is increasing and the reserves are gradually depleted,which brings greater pressure on the world's energy.The international community's increasing emphasis on energy conservation,environmental protection and the achievements in power technology have led to the development of distributed generation(DG)technology such as photovoltaic power generation and wind power generation.In addition,the DG has attracted experts at home and abroad.However,DG is vulnerable due to environment and weather.And there are characteristics such as volatility,randomness,and discontinuity.A large number of DGs joining distribution networks greatly changes the structural design and control methods of traditional distribution networks.Distribution network reconfiguration can be achieved by adjusting the topology of the network to achieve balanced load,impr ove supply voltage quality,and improve power supply reliability.What's more,it is one of the important means of distribution network optimization.Considering new energy,the traditional deterministic reconfiguration is no longer applicable.Therefore,it is of great practical significance to study distribution network reconfiguration considering DG uncertainty.In this paper,we study the probabilistic power flow with multiple DGs and different strategies for distribution network reconfiguration with DGs.The main work of this paper is as follows:The unscented transformation strategy is used to establish a probabilistic load flow model for distribution networks with multiple DGs,which provides a basis for distribution network reconfiguration.Considering the uncertainty of multiple DGs and load prediction values,this paper us es the unscented transformation to convert the uncertainty flow calculation of a large number of scenes into a deterministic power flow calculation of a small number of scenes,and solves the deterministic problem by the backward-forward load flow.At the same time,the modified IEEE33 system and other distribution network systems are used to simulate three different sampling strategies such as symmetric sampling,minimum skewness single sampling,and hypersphere single sampling.The simulation results and Monte Carlo sampling method are compared.The results prove the effectiveness of the proposed method and the applicable scopes of the three sampling strategies are to analyzed.Considering the known probability distribution function of DG output,a distribution network probability reconfiguration model with active network loss,voltage deviation and load balance as the objective function is established.In order to find the solution which is most satisfactory to decision makers,the membership function and fuzzy set are used to deal with multi-objective problems.The unscented transformation method is adopted and the appropriate sampling strategy is selected to calculate the probabilistic power flow to improve the reconfiguration speed.In addition,in order to improve the global optimization rate and the speed of the algorithm,an improved mine blast algorithm is proposed to solve the optimization problem.The effectiveness of the pr oposed model and the improved algorithm is verified by the distribution network simulation system.Compared with other algorithms,it improves the convergence speed and algorithm stability.Considering the unknown probability distribution function of DG ou tput,combined interval mathematics theory and mathematical linear programming to establish a distribution network reconfiguration strategy based on interval mixed integer linear programming.The common linear transformation method s(SOS-2 piecewise linear method,McCormick convex envelope method,etc.)are described and the conversion error is analyzed to choose appropriate transformation methods.Then the mathematical reconfiguration model based on mathematical programming is established.The expression is linearly transformed using the linear transformation tools described above.Finally,the commercial software CPLEX is used to simulate the distribution system,and the linear conversion accuracy,calculation accuracy and running time is analyzed.