[Submitted on 11 Jun 2026 (v1), last revised 29 Jul 2026 (this version, v3)]

View PDF HTML (experimental)

Abstract:No model of the Korean transmission system at native resolution is publicly available, hindering reproducible research on one of the world's most distinctive grids - an islanded interconnection with extreme separation between generation and the Seoul Metropolitan Area load center and heavy reliance on extra-high-voltage transmission. Working strictly from public data, we present the GIST Korea test system, a geographically grounded synthetic model of the Korean grid. Unlike fully synthetic cases, whose lines match no real corridor, and aggregated public Korean models, it derives its 345 and 154 kV layout from the OpenStreetMap power layer by a multi-source shortest-path reassembly of overhead-line and underground-cable geometry, gap-fills unmapped substations by a geographic minimum spanning tree, and calibrates aggregate circuit length to published statistics (94/100/104% at 765/345/154 kV). The current release spans 2265 buses, 614 generation and renewable sources (151 GW), 3185 AC circuits with explicit underground sections, four HVDC converter links, 3413 transformers, and reactive resources, in a PSS/E-compatible CSV schema with zero-sequence and grounding data for unbalanced-fault studies. The model is distributed as a frozen operating point - taps, setpoints, and bus voltages settled once offline - so a single deterministic Newton-Raphson pass reproduces a 91GW evening peak snapshot anchored to the observed 2025 summer peak (2.3% losses, no overloads), consistent with the public KPG-193 model. Standard-parameter dynamic, protection, and unit commitment layers, and a library of 24 hourly operating points derived by security-constrained unit commitment, extend the dataset beyond a single power flow case. The dataset, maps, and tooling are released as a citable, continuously maintained platform for power flow, planning, and decarbonization studies.

Submission history

From: Yun-Su Kim [view email]
[v1] Thu, 11 Jun 2026 01:15:14 UTC (321 KB)
[v2] Mon, 15 Jun 2026 06:36:05 UTC (341 KB)
[v3] Wed, 29 Jul 2026 06:52:01 UTC (773 KB)