A Tailored ADMM for Chronological Production Simulation of Flexibly Interconnected Regional Power Systems

In flexibly interconnected regional power systems with high penetration of renewable energy, chronological production simulation encounters combinatorial explosion and computational inefficiency, especially over medium- and long-term horizons with large-scale network constraints. Centralized approaches become intractable due to problem scale and memory limitations, while existing alternating direction method of multipliers (ADMM) variants suffer from convergence instability and high computational cost when regional subproblems involve discrete unit commitment decisions. To address these issues, this paper establishes a network-constrained chronological production simulation model incorporating the operational characteristics of voltage-source-converter-based high-voltage direct-current (VSC-HVDC) transmission and inter-regional flexibility reserve sharing, and proposes a tailored consensus ADMM with network-constrained clustering linearization (ADMM-NCL) for an efficient parallel solution. The ADMM-NCL combines same-type unit clustering, which convexifies the unit commitment formulation for parallel solvability, with redundant network constraint screening thus retaining only the critical constraints. In representative spring and summer periods, ADMM-NCL achieves 12×~18× speedups over the centralized benchmark and 2×~10× over the tested ADMM variants; across four seasonal scenarios (spring, summer, autumn, and winter), it bounds the objective cost deviation within 0.41% while maintaining close agreement in key operational indicators. For the 8760 h case, where the centralized benchmark and several ADMM variants fail due to memory or time limits, ADMM-NCL completes the simulation within an available computational budget, thus demonstrating its scalability with respect to the full-year chronological horizon.