An online prediction method for high-frequency oscillations using semi-analytical solutions

With the large-scale integration of tightly regulated inverter-based resources (IBRs), high-frequency oscillations (HFOs) have posed significant challenges to power systems. Recent years have witnessed incidents of HFOs above 1000Hz on a high-voltage DC transmission line at Zhangbei HVDC system that caused instability abruptions. Accurately predicting HFOs requires modelling and simulating electromagnetic transients (EMT). Existing numerical methods demand microsecond-level time steps, leading to excessive computational costs that hinder fast estimation. This paper proposes a novel method that solves the differential-algebraic equations about the EMT of a power system based on an online-offline recursive algorithm. The offline stage computes a semi-analytical solution (SAS) based on truncated power series incorporating Padé approximation. The online stage parameterizes the solution using online measurements. The separation of offline-online computations enables more numerically stable performance with longer prediction period. The proposed work is validated based on a grid-forming IBR grid integration simulation case study. Simulation results show that the proposed methods can allow fast and precise estimation at larger step size.