Power flow studies using principal component analysis

This paper employs principal component analysis (PCA), a data mining technique, to study power flow. A simulation-based process is established to perform the study, which consists of steps such as system linearization, training data construction, PCA analysis and results interpretation. The PCA results are presented in a straightforward manner, and interpreted from power system perspective. The conclusions not only are consistent with the well-known facts such as PQ decoupling, but also discover hidden facts such as correlation pattern among input variables and state variables. The proposed power flow study method is not only a helpful tool for power system operators in practice, but also beneficial for engineering students in study.