Model reduction of power systems with preservation of slow and poorly damped modes

In this paper a recently proposed variation of the Krylov subspace method for model reduction is applied to power systems. The technique allows to easily enforce constraints on the reduced order model. Herein this is used to preserve the slow and poorly damped modes of the systems in the reduced order model. We analyze the role that these modes have in obtaining a good approximation and we show that the order of the reduced model can be decreased if the “right” modes are preserved. We validate the theory on the 68-Bus, 16-Machine, 5-Area benchmark system (NETS-NYPS).