Research On Distributed Power Supply Optimization Configuration Based On Probabilistic Power Flow

In recent years,with the energy crisis and increasingly serious environmental problems,the problems brought about by the large-scale power supply mode of traditional centralized power generation have become increasingly prominent.As a new energy utilization method,distributed power sources and electric vehicles have natural advantages in dealing with energy shortages and environmental pollution,and have attracted widespread attention from governments.However,the integration of distributed power and electric vehicles also brings strong randomness,intermittent and other uncertain factors to the distribution network,which brings new challenges to the safe,economic and reliable operation of the distribution network system.The study found that the extent of the impact of distributed power on the distribution network is mainly related to the location and capacity of the installation.Therefore,on the basis of considering the uncertain factors of distributed power supply and electric vehicle,rational allocation of distributed power supply is essential for the environmentally friendly society and the safe and reliable operation of the power system.In this paper,a radial basis function neural network is introduced based on the Latin hypercube Monte Carlo simulation.The traditional Latin hypercube Monte Carlo simulation needs to solve the power flow equation after obtaining the random samples of the distributed power supply and electric vehicle.In order to reduce the computational burden of the algorithm,this paper replaces the traditional power flow calculation equation with the superior performance of the radial basis function neural network in function approximation,and directly uses the generated random samples containing uncertain factors to avoid calculating the Jacobian matrix and partial derivative.The earth reduces the computational time of the probabilistic trend.Then a multi-objective optimization function model including annual investment cost,electricity purchase cost and active loss cost is established.For the model solving algorithm,based on the adaptive differential evolution algorithm based on the success history,this paper introduces the niche technology and improves the defect that the original algorithm population size cannot be adaptively adjusted according to the Euclidean distance-based niche strategy.The crossover rate and scaling factor adaptive mechanism are retained to enhance the diversity of the population while speeding up the convergence of the algorithm and avoiding premature convergence.In addition,in the practical application,the uncertain factors of distributed power supply and electric vehicle are fully considered.The improved Latinhypercube Monte Carlo simulation is used to solve the probabilistic models of wind power,photovoltaic,load and electric vehicles,and the probabilistic power flow calculation is embedded.The improved adaptive history-based adaptive differential evolution algorithm is used to obtain the final optimal configuration scheme.Finally,the improved Latin hypercube Monte Carlo simulation is verified by IEEE14 and IEEE118 node systems.The convergence characteristics of the improved success history-based adaptive differential evolution algorithm are analyzed by CEC2014 and CEC2013 test functions;The proposed algorithm is solved in the distributed power supply optimization configuration scheme of electric vehicle.The convergence characteristics of the proposed algorithm and its comparison algorithm and the distribution of the node voltage after optimization are analyzed.The simulation results show that the proposed algorithm has the advantages of high convergence precision,short calculation time and efficient solution of optimal configuration schemes,and provides a new solution for distributed power planning problems involving electric vehicles.