Research On Probabilistic Load Flow Method Of Power System Containing Distributed Generation And Electric Vehicle Charging Load

There are many uncertainties in the operation and control of power systems.The large-scale integration of distributed generation based on volatility(wind power,photovoltaic power)and electric vehicles will bring more uncertainties in the operation and control of power system.Probabilistic load flow computation can better reveal the characteristics of the system,find out the weak link and potential risks of the system operation,and can be applied to the economic operation,safety and stability analysis,reliability analysis of the power system.Since there are multiple random variables in the probabilistic load flow calculation of power system containing distributed generation and electric vehicle charging load,a Monte Carlo method based on composite sampling method is put forward to perform probabilistic assessment analysis of voltage quality of power system containing distributed generation and electric vehicle charging load.This method considers not only the randomness of wind speed and light intensity as well as uncertainty of basic load and electric vehicle charging load,but also other stochastic disturbances,such as the failure rate of transmission line.According to the different characteristics of random factors,different sampling methods are applied.Simulation results demonstrate the accuracy,rapidity and practicability of the proposed method.In contrast to the simple random sampling Monte Carlo simulation method,the proposed method is of higher computational efficiency and better simulation accuracy.The variation of nodal voltages for power system before and after connecting distributed generation and electric vehicle charging load is compared and analyzed,especially the voltage fluctuation of the grid-connected point of distributed generation and electric vehicle charging load.In order to accurately describe the influence of the random factors in power systems on the probabilistic load flow,the stochastic response surface method is led in to perform analysis of probabilistic load flow of power system containing distributed generations and electric vehicle charging load.The Hermite polynomial chaos expansion is applied to fit the function relation between the output responses and the input random variables,the optimal collocation points are selected according to linear independent principle,and the stepwise regression analysis is introduced to screen out minor items in response surface expansion terms,the undetermined coefficients in the expansion are decreased efficiently under the prerequisite of ensuring calculation accuracy,thereby the statistical information of the response could be assessed by faster simulation process,the computational efficiency and simulation accuracy are significantly improved.Comparing with the traditional stochastic approaches,such as Monte Carlo simulation method,the proposed method can solve problems that they are facing,like more number of calculation times,long simulation time and need of large memory space.Simulation results demonstrate the proposed method is of higher computational efficiency and better simulation accuracy.To calculate probabilistic load flow of power system containing distributed generators and electric vehicle charging load,a probabilistic load flow analysis evaluation model of power system considering the correlation of between wind power outputs and between photovoltaic outputs is built on the basis of Hybrid Copula Function.According to this basis and based on these drawback that the traditional Monte Carlo simulation method based on Latin Hypercube Sampling is of low accuracy for its incapability of generating sampling sequences of low discrepancy,a Quasi Monte Carlo simulation method is proposed.At first the hybrid Copula is introduced to model the correlation of between wind power outputs and between photovoltaic power outputs.Then,Sobol sequence that can overcome the bottleneck of simple random sampling Monte Carlo simulation method convergence is introduced to generate Low-Discrepancy Sequences.The simulation results demonstrate the validity of the proposed method.In contrast to Monte Carlo simulation method based on Latin Hypercube Sampling,the proposed method is of superior speed and accuracy and robustness.