A Novel Discrete-Time State-Space Model for Decentralized Dynamic State Estimation of Grid-Forming Inverters

Stiff dynamics continue to pose challenges for power system dynamic state estimation. In particular, models of inverters with control schemes designed to support grid voltage and frequency, namely, grid-forming inverters (GFMs), are highly prone to numerical instability. This paper develops a novel analytical modeling technique derived from two cascading subsystems, namely synchronization and dq-frame voltage control. This allows us to obtain a closed-form discrete-time state-space model based on the matrix exponential function. The resulting model enables a numerically stable and decentralized dynamic state estimator that can track the dynamics of GFMs at standard synchrophasor reporting rates. In contrast, existing dynamic state estimators are subject to numerical issues. The proposed algorithm is tested on a 14-bus power system with a GFM and compared with the standard algorithm whose process model is discretized using well-known Runge-Kutta methods. Numerical results demonstrate the superiority of the proposed method under various conditions.