Simulating the Power Electronics-Dominated Grid using Schwarz-Schur Complement based Hybrid Domain Decomposition Algorithm

TL;DR

This paper introduces a two-stage hybrid domain decomposition algorithm combining Schwarz-Schur methods to accelerate power system dynamic simulations, especially with high renewable penetration, ensuring accuracy and stability.

Contribution

It presents a novel hybrid decomposition approach that improves computational efficiency for large-scale power system simulations with renewables.

Findings

Effective in simulating power systems with high renewable integration

Ensures convergence and accuracy comparable to original models

Applicable to large and very-large-scale power system analyses

Abstract

This paper proposes a novel two-stage hybrid domain decomposition algorithm to speed up the dynamic simulations and the analysis of power systems that can be computationally demanding due to the high penetration of renewables. On the first level of the decomposition, a Schwarz-based strategy is used to decouple the original problem into various subsystems through boundary variable relaxation, while on the second level, each decoupled subsystem is further decomposed into subdomains that are solved independently using the Schur-complement approach. Convergence is checked in both stages to ensure that the parallelized implementation of the subsystems can produce identical results to the original problem. The proposed approach is tested on an IEEE 9 bus system in which one synchronous generator is replaced with a solar PV farm through a grid-forming inverter (GFM) with an admittance control method to evaluate its effectiveness and applicability for large-scale and very-large-scale implementations. Since conventional dual-loop GFMs are not stable when connecting to a stronger grid with a small grid inductance, a virtual inductance method is adopted to increase the equivalent inductance connecting the grid to enhance stability.