Coherency Analysis in Nonlinear Heterogeneous Power Networks: A Blended Dynamics Approach
Yixuan Liu, Yingzhu Liu, Pengcheng You
TL;DR
This paper addresses power-system coherency in large-scale, heterogeneous nonlinear networks subject to persistent disturbances. It extends the blended dynamics framework to construct a reduced-order, coherent trajectory $\omega_b$ as a weighted average of nodal dynamics and establishes time-domain bounds on the coherence error $e_c = \max_i |\omega_i - \omega_b|$. The results show two key factors promoting coherence: high network connectivity (large $\lambda_2$ or $k$) and slow disturbance variation (small $\dot{\xi}_i(t)$), with explicit exponential decay rates and long-run bounds that hold for both linearized and nonlinear power flows. Simulations on the Icelandic grid corroborate the theory, illustrating practical implications for model reduction and controller design in modern grids with renewable integration.
Abstract
Power system coherency refers to the phenomenon that machines in a power network exhibit similar frequency responses after disturbances, and is foundational for model reduction and control design. Despite abundant empirical observations, the understanding of coherence in complex power networks remains incomplete where the dynamics could be highly heterogeneous, nonlinear, and increasingly affected by persistent disturbances such as renewable energy fluctuations. To bridge this gap, this paper extends the blended dynamics approach, originally rooted in consensus analysis of multi-agent systems, to develop a novel coherency analysis in power networks. We show that the frequency responses of coherent machines coupled by nonlinear power flow can be approximately represented by the blended dynamics, which is a weighted average of nonlinear heterogeneous nodal dynamics, even under time-varying disturbances. Specifically, by developing novel bounds on the difference between the trajectories of nodal dynamics and the blended dynamics, we identify two key factors -- either high network connectivity or small time-variation rate of disturbances -- that contribute to coherence. They enable the nodal frequencies to rapidly approach the blended-dynamics trajectory from arbitrary initial state. Furthermore, they ensure the frequencies closely follow this trajectory in the long term, even when the system does not settle to an equilibrium. These insights contribute to the understanding of power system coherency and are further supported by simulation results.