Comparisons of Correlated Random Variables Decorrelation Methods in Probabilistic Power Flow Calculation

Xiaochuan Wei,Zheng Zhu,Qin Xu,Yuanhai Gao,Zheng Yan

DOI: https://doi.org/10.1109/psgec54663.2022.9881134

2022-01-01

Abstract:Large-scale renewable energy generations introduce complex correlated random input variables in power systems, which create challenges of uncertainty for operation and control. Probabilistic power flow (PPF) is a quantitative approach to analyzing power system uncertainty. To calculate PPF considering correlated random input variables, Orthogonal, Nataf, and Rosenblatt transformations are widely used for decorrelation. However, errors caused by the three transformation methods for different PPF calculation methods are still unclear. This paper investigates Orthogonal, Nataf, and Rosenblatt transformations in PPF. First, description methods of correlated random variables are reviewed, and the limitations of correlation coefficients are clarified. Second, the ideal conditions for using Orthogonal, Nataf., and Rosenblatt transformations are analyzed. The third is the discussion of PPF calculation implementations that consider correlated random input variables using different methods. The feasibility of the three transformation methods in different PPF methods is given. Finally, an IEEE 33 system integrated with real measured correlated wind powers is used to analyze PPF errors caused by the three transformation methods. The numerical simulation results show that the impacts on PPF error from decorrelating transformations are different in different PPF calculation methods. Recommendations for transformation methods in different PPF calculation methods are presented.