{"id":"03037afd-5cef-4e0a-9cfa-d77522fca6c7","arxiv_id":"2411.14756","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"KPG 193 is an open, 193-bus synthetic Korean transmission network with 2022 hourly demand and renewable data, validated by unit commitment and AC optimal power flow simulations.","lead":"This paper builds and publicly releases KPG 193, a 193-bus synthetic model of the Korean power grid made from open data, with 122 generators, 407 lines, and hourly 2022 weather and demand profiles. It gives researchers an open testbed for decarbonization and expansion planning in Korea, where real grid data is largely withheld.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The representativeness claim rests on an unvalidated OSM-to-193-bus clustering, and the one quantitative match reported as validation is partly calibrated; independent aggregate checks are needed.","rationale":"The reader's weakest_assumption correctly identifies the clustering of OSM line data onto 193 KEPCO regional offices as the load-bearing step: if that synthetic topology is not representative, every subsequent flow, voltage pattern, and expansion-planning conclusion inherits the distortion. My stress-test agrees with this and sharpens it by pointing to the absence of any independent aggregate check and to the explicit manual adjustment of line connections, which introduces an unquantified, non-reproducible degree of freedom. The reader's rationale also flags validation circularity from cost-parameter fitting; I treat that as a supporting weakness rather than the central one, because the main claimed validation of 'representativeness' is the flow/voltage behavior, which depends on topology more than on cost coefficients. The paper is honest about its limitations and does not claim exact fidelity, so the appropriate disposition remains conditional: the dataset is likely useful, but the fidelity assertion should not be taken at face value until the proposed aggregate topology check is performed and reported. I therefore recommend no change to the reader's CONDITIONAL verdict; my concrete test would either substantiate the topology or expose a need to downgrade the fidelity claim to 'plausible benchmark' rather than 'representative of the Korean grid.'","tokens_in":7097,"tokens_out":5000,"duration_ms":56312,"concrete_test":"Compare KPG 193's aggregate transmission infrastructure against official Korean grid statistics that are publicly available independently of the paper, e.g., KPX/KEPCO annual transmission statistics reporting total circuit length by voltage class (154 kV, 345 kV, 765 kV, HVDC) and total transformer capacity. Specifically, compute the total circuit length at each voltage level from the released MATPOWER case file and compare it with the official 2022 totals. If the model's voltage-class lengths (or their ratios) deviate by more than roughly 20% from official totals, the OSM extraction and clustering are not representative, and the flow/voltage validation should be treated as unverified. A secondary check: rerun the UC/ACOPF with unmodified cost coefficients from [31]/[32] and report the generation mix, to separate calibration from validation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that KPG 193 'captures' Korea's concentrated urban demand, coastal generation, and isolated grid structure, with power-flow and voltage patterns that mirror the real grid. The load-bearing step is the spatial clustering in Sections II-A-1 and II-B-1: OSM-derived transmission line endpoints are assigned to the nearest of 193 KEPCO regional offices, with 'some line connection points manually adjusted' to avoid isolated buses. There is no independent evidence that this procedure preserves electrical distances, transfer capacities, or even connectivity to a meaningful degree. OSM data is itself acknowledged in Section IV as having uncertain update schedules and topology accuracy issues, and the actual Korean grid topology is not public. Consequently, the ACOPF-based flow directions in Fig. 3 and the voltage heatmap in Fig. 4 are outputs of an unvalidated synthetic topology, not comparisons against measured or independently derived grid behavior. Moreover, the only quantitative validation in Table III compares annual generation shares, but Section III states that cost coefficients were explicitly 'modified ... to reduce discrepancies with historical data.' That makes the 1.4% match a calibration result, not an independent validation. Together these leave the fidelity claim under-supported: the model may be feasible and internally plausible, but the paper has not shown that its topology is representative beyond the coarse, input-driven facts that load is concentrated in the Seoul area and generators are coastal.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents KPG 193, a synthetic 193-bus, 122-generator, 407-transmission-line test system for the Korean power grid, constructed entirely from open data sources. The modeling workflow clusters OpenStreetMap transmission-line geometry, KPX/Korea Energy Agency generator data, KEPCO demand data, and LDAPS weather data onto 193 KEPCO regional offices. The system includes hourly 2022 demand, weather, and renewable profiles, and the authors report that daily unit commitment (UC) and hourly AC optimal power flow (ACOPF) simulations converge for the full year. Validation consists of a comparison of annual generation shares by fuel type in Table III and a qualitative snapshot comparison of power-flow directions and voltage magnitudes in Figures 3 and 4. The paper acknowledges limitations including the omission of compensation devices, uncertainty in OSM topology data, and a fixed reactive-power demand ratio of 90% of real power.","tokens_in":7334,"tokens_out":4823,"duration_ms":48900,"significance":"If the fidelity claims are supported, KPG 193 would fill a real gap: a public, spatio-temporally resolved Korean test system for decarbonization and expansion-planning studies, built from reproducible open data. The release of the dataset, the integration of hourly weather-derived renewable profiles, and the demonstrated convergence of year-long UC/ACOPF simulations are concrete strengths. However, the current evidence for representativeness is weaker than the paper's language suggests. The main quantitative validation is partly circular because cost coefficients were fit to reduce discrepancies, and the spatial topology derived from OSM clustering plus manual adjustment is not independently checked against any real or independently derived Korean grid property. The paper needs either additional validation or a more careful framing of what is being claimed.","major_comments":[{"comment":"The validation reported in Table III is partly circular and should be reframed. The text states that generator cost coefficients were \"modified using methodologies from [9] to reduce discrepancies with historical data,\" and Table III then presents the resulting annual generation shares as a successful validation. Since the UC generation mix is a direct output of those fitted cost coefficients, the 1.4% deviation is a calibration result, not an independent test of the model. Please either (a) validate against quantities not used in fitting, such as hourly dispatch patterns, regional generation shares, or an out-of-sample year; or (b) explicitly state that Table III demonstrates calibration quality, and move the fidelity claim to quantities that are not fitted.","section":"Section III, Table III, Eq. (3)"},{"comment":"The representativeness of the network topology is under-supported. OSM line endpoints are clustered to the nearest of 193 KEPCO regional offices, and \"some line connection points were manually adjusted,\" but no evidence is given that the resulting 407-line model preserves electrical distances, transfer capacities, impedance distribution, or even gross connectivity relative to the actual Korean grid. Section IV itself notes that OSM has \"uncertain update schedules\" and topology accuracy issues. Figures 3 and 4 are therefore outputs of an unvalidated synthetic topology, not comparisons with measured or independently derived grid behavior. To support the claim that the model \"captures\" the real grid's flow and voltage patterns, I would need at least one independent aggregate check, for example comparing inter-regional transfer capacities or line utilization with KPX data, or a sensitivity analysis showing that the reported flow directions and voltage patterns are robust to the clustering and manual-adjustment choices.","section":"Section II-B-1, Sections II-A-1 and IV, Figs. 3-4"},{"comment":"The voltage-related fidelity claim is weakened by the paper's own list of omissions. Section IV states that no compensation devices (shunt capacitors, STATCOMs, tap-changing transformers) are included and that nodal reactive power demand is fixed at 90% of real power. Under these conditions, the voltage heatmap in Fig. 4 cannot be presented as a replication of actual Korean grid voltage behavior; it is a stress-test-style result of a model without voltage-control equipment. The authors should either add representative compensation devices or clearly label the voltage pattern as illustrative and qualitative, rather than using it as evidence that the test system \"successfully captured ... voltage patterns observed in the actual grid.\"","section":"Section IV and Fig. 4"}],"minor_comments":[{"comment":"The number of generators is inconsistent: the abstract in the header says 123 generators, while the full-text abstract and Section II-C say 122 generators. Please reconcile this discrepancy.","section":"Abstract and Section II-C"},{"comment":"There is a typo in the conclusion: \"devloping\" should be \"developing.\"","section":"Section V"},{"comment":"The sentence \"The Korean power system is characterized by a concentrated of electricity demand\" contains a grammatical error; it should read \"a concentration of electricity demand.\"","section":"Section III"},{"comment":"The HVDC line is mentioned as manually added, but no electrical parameters or ratings for this line are provided in Table I or elsewhere. Please add the HVDC line parameters or cite where they can be found.","section":"Table I and Section II-A-1"},{"comment":"Table IV covers thermal generator parameters but does not list the renewable-curve constants v_CI, v_R, v_CO, and FF_PV introduced in Eqs. (1)-(2). Please provide the numerical values or a reference for these constants.","section":"Table IV"}],"recommendation":"major_revision","confidential_remarks":"The dataset and computational pipeline are potentially useful, but the circular validation in Section III and the unverified topology clustering in Section II-B-1 are the main gatekeeping issues. The paper would be publishable after either adding independent fidelity checks (e.g., out-of-sample or regional comparisons, topology sensitivity) or substantially tempering the representativeness claims to match the evidence actually provided."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe short version: KPG 193 is the first open synthetic Korean transmission grid with weather-coupled data, and that alone makes it worth a serious look. It is not a field-reshaping method, but it fills a real gap for researchers who need a public Korean test system for decarbonization or expansion planning.\n\nWhat it does well: the construction workflow is transparent and reproducible in spirit, with clear data sources (OSM, KPX, KEPCO, LDAPS), and the dataset is actually released. Running daily UC and hourly ACOPF over a full year is a solid feasibility exercise. The qualitative checks—concentrated demand near Seoul, coastal generation, northward flows, voltage patterns—are consistent with what is known about the Korean grid. The limitations section is honest about missing compensation devices, OSM uncertainty, and fixed reactive demand.\n\nThe soft spot is real and the stress-test note is right: the headline quantitative validation in Table III is partly circular. The authors modified cost coefficients to reduce discrepancies with the 2022 historical generation data, then report the resulting generation shares as validation. The 1.4% deviation is a calibration result, not an independent check. That should be labeled clearly, or better, supplemented with an independent aggregate comparison (e.g., line-length totals, or generation by province from another year). The clustering of OSM lines to the 193 KEPCO offices is also not independently verified; without the real internal bus-level data it is hard to know how well the approximate topology preserves transfer capacities or electrical distances. The paper acknowledges this via the OSM limitations, but the flow and voltage figures should be framed as plausible outputs of the synthetic model, not validated grid behavior.\n\nThese concerns don't sink the dataset. The artifact is useful, the limitations are disclosed, and the methodology follows established practice in PyPSA-Eur and Texas A&M synthetic grids. My recommendation: send it to peer review, and require the authors to (1) reframe Table III as calibration, (2) add at least one independent aggregate check, and (3) release the processing code so the clustering can be inspected. For readers working on Korean energy systems, this is a worthwhile resource. I would cite it and would bring it to a reading group focused on synthetic grid validation.","headline":"First open synthetic Korean grid dataset worth a serious look; the headline validation is calibrated rather than independent, but the artifact and honest limitations support peer review.","tokens_in":7897,"tokens_out":2878,"would_cite":true,"duration_ms":25989,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper introduces a 193-bus synthetic Korean power grid built entirely from open 2022 data, and shows that this coarse model reproduces the real system's fuel mix, coastal generation, concentrated metropolitan load, and northward power…","keywords":["synthetic test system","Korean power grid","decarbonization studies","open data modeling","unit commitment","AC optimal power flow","renewable generation profiles","transmission network clustering"],"falsifier":"Compute the transmission losses or the net power transfer crossing the metropolitan boundary in KPG 193's 2022 ACOPF solution and compare them with the real system's published figures; if the simulated interregional flow direction reverses, or losses differ by more than a few percent, the representative-clustering claim fails.","tokens_in":6839,"feed_emoji":"⚡","tokens_out":7875,"duration_ms":69421,"temperature":0.7,"pith_summary":"The paper introduces KPG 193, a 193-bus, 122-generator, 407-line synthetic model of the Korean transmission grid assembled entirely from open 2022 data. Because Korea does not publish its real grid topology, the authors build the network by clustering open map data, plant locations, demand, and weather around the utility's 193 regional offices. They then demonstrate the model's feasibility by solving daily unit commitment and hourly AC optimal power flow for all of 2022, obtaining fuel-generation shares within 1.4 percentage points of historical data. The resulting model reproduces Korea's defining pattern: coastal power plants, a metropolitan area consuming about 40% of demand, and persistent northward power flows over long 345/765 kV and 500 kV HVDC lines. If the construction is sound, researchers gain a public, reproducible Korean testbed for decarbonization and transmission-expansion studies.","feed_headline":"Open data yields a 193-bus Korean grid that reproduces real flows","feed_subtitle":"Built only from public data, it matches 2022 fuel shares within 1.4% and can drive decarbonization studies.","key_machinery":"The load-bearing construction is clustering: every municipal-level dataset (230 municipal segments) is reassigned to 193 buses placed at the utility's regional offices; line endpoints are snapped to the nearest bus, plants are assigned to the nearest bus, demand is allocated from municipal monthly consumption plus population, and 1.5 km resolution weather fields are assigned per bus. This clustering is what preserves Korea's geographic character, coastal generation, metropolitan load of about 40%, and long transmission corridors, while hiding exact facility locations and keeping the model small enough for year-long simulations. Standard conductor-based line parameters and a manually added HVDC link complete the network.","core_discovery":"The central discovery is that coarse spatial clustering onto 193 buses is enough to make a synthetic grid behave like the real Korean system in the ways that matter for planning: the annual generation mix by fuel matches the historical mix within 1.4 percentage points, unit commitment and ACOPF converge for every hour of 2022, and the peak-summer snapshot shows northward transfers and coastal overvoltage patterns consistent with the actual grid. The authors state this as a feasibility and representativeness result, not as an exact replica: the model is designed to capture Korea's low renewable penetration, concentrated urban demand, and isolated grid structure, and to provide a stable testbed rather than a faithful map.","pith_inferences":["Because the model omits shunt capacitors, STATCOMs, and tap-changing transformers, and fixes reactive demand at 90% of real power, voltage-magnitude results should be read as a structural pattern rather than a level; adding reactive-support devices would be the natural next test.","The same clustering recipe can be rerun on updated or historical open data to produce annual editions of KPG, turning the 2022 snapshot into a family of yearly grids for longitudinal decarbonization studies.","A stronger validation would compare simulated line flows or transmission losses against the operator's measured interregional transfers; such a check could quantify how much of the northward-flow signature is real signal versus an artifact of clustering."],"forward_implications":["Researchers can run year-long unit commitment and AC optimal power flow on a public Korean grid model without needing confidential utility data.","Decarbonization scenarios that raise Korea's renewable share can be tested while preserving realistic geographic constraints, including the long coastal-to-metropolitan transfer corridors.","The 2022 weather dataset bundled with the model lets renewable profiles be recomputed at each bus, supporting storage-siting and transmission-expansion planning.","Because the model matches the historical 2022 fuel mix within 1.4 percentage points, it provides a baseline for calibrating future Korean grid studies.","The construction method transfers to other regions that publish municipal-level demand and generator data but keep their transmission topology confidential."],"supporting_citations":[{"why":"Establishes the accuracy baseline for the volunteered mapping data that supplies the network topology.","marker":"[5]"},{"why":"Provides the query tool used to extract overhead-line and underground-cable geometry for the network.","marker":"[18]"},{"why":"Defines the 193 regional-office locations that become the buses and the service boundaries used for spatial clustering.","marker":"[25]"},{"why":"Supplies cumulative installed renewable capacity by municipality, used to place and size solar and wind generators.","marker":"[15]"},{"why":"Supplies approximate locations and capacities of conventional plants, used to build the conventional generator fleet.","marker":"[22]"},{"why":"Provides monthly municipal electricity consumption used to distribute system demand geographically.","marker":"[16]"},{"why":"Provides the 2022 hourly system demand time series used in the demand-allocation and validation simulations.","marker":"[24]"},{"why":"Provides 1.5 km hourly weather fields (wind, irradiance, temperature) from which renewable profiles are computed.","marker":"[23]"},{"why":"Supplies the wind and solar power profile equations used to convert weather data into renewable generation.","marker":"[29]"},{"why":"Supplies the method for adjusting generator cost coefficients so simulated generation mixes match historical data.","marker":"[9]"}],"fun_headline_variants":["193-bus synthetic Korean grid for decarbonization studies","Open-data 193-bus grid mirrors Korean fuel mix in 2022","KPG 193: synthetic test system for Korean decarbonization","Synthetic Korean grid (193 buses) for decarbonization planning"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that grouping the open, municipal-level maps and statistics onto 193 utility regional offices produces a network whose flows and voltages stand in for the real, undisclosed Korean grid; stale map data or distorted clustering would silently invalidate every simulation based on it.","fun_headline_variants_meta":{"raw":{"variants":["193-bus synthetic Korean grid for decarbonization studies","Open-data 193-bus grid mirrors Korean fuel mix in 2022","KPG 193: synthetic test system for Korean decarbonization","Synthetic Korean grid (193 buses) for decarbonization planning"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000623,"raw_usage":{"total_tokens":2824,"prompt_tokens":820,"completion_tokens":2004,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":436,"completion_tokens_details":{"reasoning_tokens":1932}},"tokens_in":436,"tokens_out":2004,"duration_ms":14090,"temperature":1.0,"reasoning_tokens":1932,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:55:53.421476+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the transmission losses or the net power transfer crossing the metropolitan boundary in KPG 193's 2022 ACOPF solution and compare them with the real system's published figures; if the simulated interregional flow direction reverses, or losses differ by more than a few percent, the representative-clustering claim fails.","supporting_citations":[{"cited_title":"How good is volunteered geographical information? a comparative study of openstreetmap and ordnance survey datasets,","cited_arxiv_id":null,"evidence_quote":"Establishes the accuracy baseline for the volunteered mapping data that supplies the network topology."},{"cited_title":"Overpass Turbo,","cited_arxiv_id":null,"evidence_quote":"Provides the query tool used to extract overhead-line and underground-cable geometry for the network."},{"cited_title":"Kepco introduction, organization, regional offices,","cited_arxiv_id":null,"evidence_quote":"Defines the 193 regional-office locations that become the buses and the service boundaries used for spatial clustering."},{"cited_title":"Supply statistics cumulative supply capacity - regional cumulative supply capacity (basic),","cited_arxiv_id":null,"evidence_quote":"Supplies cumulative installed renewable capacity by municipality, used to place and size solar and wind generators."},{"cited_title":"Transmission of electricity in south korea: Q&a,","cited_arxiv_id":null,"evidence_quote":"Supplies approximate locations and capacities of conventional plants, used to build the conventional generator fleet."},{"cited_title":"Electricity sales volume by city/county/district,","cited_arxiv_id":null,"evidence_quote":"Provides monthly municipal electricity consumption used to distribute system demand geographically."},{"cited_title":"Hourly national electricity demand by city/county/district - 2022-12-31,","cited_arxiv_id":null,"evidence_quote":"Provides the 2022 hourly system demand time series used in the demand-allocation and validation simulations."},{"cited_title":"Local data assimilation and prediction system (ldaps)","cited_arxiv_id":null,"evidence_quote":"Provides 1.5 km hourly weather fields (wind, irradiance, temperature) from which renewable profiles are computed."},{"cited_title":"Optimal renewable resources mix for distribution system energy loss minimization,","cited_arxiv_id":null,"evidence_quote":"Supplies the wind and solar power profile equations used to convert weather data into renewable generation."},{"cited_title":"US test system with high spatial and temporal resolution for renewable integration studies,","cited_arxiv_id":null,"evidence_quote":"Supplies the method for adjusting generator cost coefficients so simulated generation mixes match historical data."}],"review_version":1}