Probabilistic Load Flow Calculation of Power System Integrated with Wind Farm Based on Kriging Model

Abstract

Because of the randomness and uncertainty, integration of large-scale wind farms in a power system will exert significant influences on the distribution of power flow. This paper uses polynomial normal transformation method to deal with non-normal random variable correlation, and solves probabilistic load flow based on Kriging method. This method is a kind of smallest unbiased variance estimation method which estimates unknown information via employing a point within the confidence scope of weighted linear combination. Compared with traditional approaches which need a greater number of calculation times, long simulation time, and large memory space, Kriging method can rapidly estimate node state variables and branch current power distribution situation. As one of the generator nodes in the western Yunnan power grid, a certain wind farm is chosen for empirical analysis. Results are used to compare with those by Monte Carlo-based accurate solution, which proves the validity and veracity of the model in wind farm power modeling as output of the actual turbine through PSD-BPA.