Probabilistic power flow for multiple wind farms based on RVM and holomorphic embedding method

Wind power generation provides a new route for the sustainable development of energy. However, there is a correlation between wind speed in different regions, which will impact wind power and the stability of the power system. Therefore, it is necessary to estimate the steady-state characteristics of power systems and calculate probabilistic power flow (PPF) by considering this correlation. However, the traditional PPF calculation method based on the Newton–Raphson method may suffer from divergence or slow convergence with a high penetration of distributed energy resources. In this study, a new approach based on the relevance vector machine (RVM) was proposed to construct the multivariate copula in order to calculate the multivariate distribution of wind speed. By using the Rosenblatt transformation, the independent wind power variables could be obtained. After that, the cumulant method based on holomorphic embedding was proposed, which could acquire more accurate probability distribution results. The proposed novel method is evaluated on small, medium, and large power flow test cases and compared favorably with the popular PPF methods(such as NR method) on the same platform. Besides, the obtained results illustrated the advantages of the proposed method.