To accurately analyze operational risk of large-scale wind power integration system

we need to build a accurate model of probability distribution functions（PDFs） of input variables in probabilistic load flow（PLF） analysis. Therefore

we proposed a PLF method based on spline reconstruction and quasi Monte Carlo simulation. This method can establish PDF of input variables according to variable torque information

and obtain the relevant input variables sample using Nataf conversion which is based on spline reconstitution

then quasi Monte Carlo simulation based on Sobol sequence is adopted to obtain the probability distribution of the output variables. Simulation on IEEE 30 bus system and a real power system demonstrate the validity of the proposed method. The results suggest that the proposed method not only has the advantages of modelling input variables accurately and fast computation speed

but also can deal with correlation with convenience.