Transitioning to Inverter-based Resources: Stability of AC Power Systems considering Current Constraints, Small-signal Sensitivities, and Large Loads

Abstract

The large-scale integration of inverter-based resources (IBRs) has fundamentally altered the dynamics of alternating-current (AC) power systems. Traditionally founded on the physical dynamics of synchronous machines, the grid is transitioning toward a power-electronics-dominated system driven by renewable generation and the proliferation of inverter-based loads from data centers and electrification efforts. While digitally controlled IBRs provide programmable flexibility, they are bound by rigid hardware constraints that limit their capabilities. Ensuring grid stability now requires generalized approaches that move beyond synchronous-machine-centric models to account for the unique dynamics of power electronics. This dissertation investigates IBR control and stability across three critical domains: individual device constraints, system-wide network stability, and large load regulation. First, we address the nonlinear control challenges posed by inverter current magnitude limits, which are essential for protecting semiconductor devices, but inhibit IBRs ability to provide a strong voltage source response during transients. We characterize the geometry of the feasible output of current-limited inverters and derive Lyapunov stability conditions for devices with direct current limiters. We then propose a safety filter approach that minimally modifies desired control actions to maintain safety, and finally propose a safe trajectory gradient flow controller that iteratively optimizes the future trajectory to remain within a defined safe set. Second, we examine the impact of IBRs on system-wide small-signal stability. Leveraging equivalent admittance modeling of systems, we propose a sensitivity-based metric that quantifies how perturbations to power flow parameters, such as bus power injections or network admittances, impact system eigenvalues. We demonstrate the insights these metrics provide for identifying and mitigating stability challenges in high-IBR networks. Lastly, we investigate leveraging IBRs within data centers to mitigate load volatility. We demonstrate how reactive power control and energy storage dispatch can be optimized for voltage regulation, reducing fluctuations and violations caused by the load variations typical of LLM training tasks. Collectively, this work provides an overview of a range of stability-related challenges that must be considered when integrating IBRs into modern, large-scale power systems.