Modular Linear Power Flow Model Against Large Fluctuations
Jiarui Long, Zhifang Yang, Junbo Zhao, Juan Yu
Abstract
The linear power flow model based on an empirical operating point is generally employed in power generation dispatch. However, with the increasing penetration of renewable energy sources (RESs) and frequent topology changes, the expansion of the operating region causes considerable modeling errors in the static linear power flow model. In this paper, the sources of linearization error are thoroughly analyzed, and a modular linear power flow modeling framework is proposed. It is found that the key features that affect the accuracy of linearization are the truncation error and its first derivative. This motivates us to cluster historical scenarios based on selected features. The key approach is to individually select a linear power flow model for power system operating conditions with similar linearization error features. This allows us to derive a refined linear power flow model for each cluster to avoid excessive linearization errors using a uniform linear power flow model. Numerical results for several IEEE and Polish test systems show that the linearization error of the proposed method is reduced up to 52.98% compared to current methods. The results demonstrate that the proposed method can use a small number of linearization segments to deal with topology changes.