Distributed AC Optimal Power Flow: A Scalable Solution for Large-Scale Problems
Xinliang Dai, Yuning Jiang, Yi Guo, Colin N. Jones, Moritz Diehl, Veit Hagenmeyer
TL;DR
This work tackles the scalability of large-scale AC OPF by developing a two-level distributed optimization framework that uses a barrier method to handle inequality constraints at the upper level and ALADIN to solve smoothed equality-constrained problems at the lower level. By condensing derivatives with the Schur complement and introducing distributed inertia correction, the approach reduces communication overhead while preserving convergence guarantees. The method demonstrates rapid convergence and competitive solution quality compared to centralized IPOPT on large benchmarks, highlighting its potential for privacy-preserving, scalable power-system optimization under the SPDM paradigm. Practical results show favorable initialization speed and robust performance across diverse network decompositions and operating scenarios, underscoring the method’s applicability to real-world large-scale grids.
Abstract
This paper introduces a novel distributed optimization framework for large-scale AC Optimal Power Flow (OPF) problems, offering both theoretical convergence guarantees and rapid convergence in practice. By integrating smoothing techniques and the Schur complement, the proposed approach addresses the scalability challenges and reduces communication overhead in distributed AC OPF. Additionally, optimal network decomposition enables efficient parallel processing under the single program multiple data (SPMD) paradigm. Extensive simulations on large-scale benchmarks across various operating scenarios indicate that the proposed framework outperforms the state-of-the-art centralized solver IPOPT on modest hardware. This paves the way for more scalable and efficient distributed optimization in future power system applications.