Probabilistic Load Flow Method Considering Dependences Of Photovoltaic Power

Photovoltaic with its inexhaustible,no pollution,safe and reliable,has become the fastest growing green energy all over the world.However,photovoltaic power generation is a typical intermittent energy source,and the PV power depends on the radiation intensity obtained by photovoltaic array.Muti-photovoltaic plants within the same region lie in the same radiation intensity zone,resulting in the PV power shows appreciably correlative.And the correlation makes the PV power more fluctuant.So for probabilistic load flow in system including PV power stations,it is important to model the dependences of Mutiple PV power to analyze the impact of PV power on power system.This paper proposes probabilistic load flow method considering dependences of multiple photovoltaic power.The main work and innovative achievements are as follows:Firstly,considering the asymmetric tail dependence between the PV power,the mixed Copula function is adopted based on the linear combination of the binary Copula functions.A penalized likelihood function is constructed and the EM algorithm is adopted to estimate parameters in the mixed Copula function.Taking two PV power stations in Qinghai area for example,the Simulink result shows the mixed Copula function can describe the dependence structure among PV power more precisely than the single Copula function.Secondly,considering the multi-Copula function can only describe the symmetrical tail dependence among multi-dimensional random variable.Based on vine,pair Copula function is adopted using the binary Copula functions.Pair Copula function considers the variant dependency structure of pairs of PV power.Taking the history data of PV power stations in Texas for example,the Euclidean distance is taken to demonstrate the pair Copula function describe the dependence structure among PV power more precisely than multivariate Gaussian Copula function.Thirdly,probabilistic load flow method is improved and digital interlacing technology is adopted to interlace the higher dimensional quasi random number to lower dimensional quasi random number,eliminating the influence of the non-uniformity of the higher dimensional quasi random number on algorithm convergence and improving the efficiency.Combining the pair Copula function,the probabilistic load flow in IEEE 30-bus system and IEEE 118-bus system is taken to demonstrate the proposed method can improve the calculation speed and get more accurate result in the same sampling scale.