PaperlayerView exact revision on arXiv

Source-linked AI summary

An Open Synthetic Test System for the Jordanian Transmission Grid

Muhy Eddin Zater, Bri-Mathias Hodge

arXiv:2609.00388v1eess.SY

TL;DR

Open transmission benchmarks lack a reproducible Middle-Eastern system despite regional questions involving fuel imports, small synchronous grids, and renewable integration. This paper constructs and validates a public-source synthetic model of Jordan’s 2018 transmission grid, reproducing the annual technology-level energy mix while releasing the dataset and supporting scripts for system-level studies.

  • Problem

    No openly reproducible transmission test system exists for a Middle-Eastern grid, although available benchmarks do not represent its fuel-supply, interconnection, and renewable-generation characteristics.

  • Method

    The paper reconstructs Jordan’s transmission topology, restores its 2018-energized state, adds plant-level fleets and hourly renewable profiles, calibrates loads, and validates the model across structural and operational studies.

  • Results

    The model reproduces the annual energy mix at technology level, with the largest producers within 4% and system total within 0.1% of published values.

  • Takeaways & Limitations

    The released case supports reproducible studies of resource adequacy, dispatch and unit commitment, renewable integration, fuel-supply resilience, interconnections, and steady-state operation.

  • Takeaways & Limitations

    The model is synthetic and omits system dynamics, inverter models, plant-specific reactive capability, network constraints in production-cost modeling, and the 33 kV layer.

Abstract

from arXiv · show

Open synthetic test systems are essential for reproducible power system research, yet the available cases represent almost exclusively North American and European grids. No open transmission test system exists for any Middle Eastern country, whose grids raise different questions such as single corridor fuel-supply resilience and high-renewable operation within small synchronous systems, and established benchmarks cannot demonstrate. This paper presents a synthetic test system for the Jordanian transmission grid, assembled entirely from public sources, where the real topology is reconstructed from a published diagram and restored to its 2018 energized state, with plant-level generation and renewable fleets, per-site hourly profiles, and loads calibrated to the values published by Jordan's grid operators. The case is validated through structural statistics against real-grid criteria, power flow and $N\!-\!1$ screening, an energy-weighted loss decomposition, in addition to a full-year production-cost run compared against the published per-plant energy, and cross-solver verification. The model reproduces the annual energy mix at technology level (largest producers within 4\%, system total within 0.1\%), and the released dataset includes the bus-identity key, all scripts, and a post-2019 scenario variant. This test case is designed to benchmark system-level resource adequacy, time-series dispatch and unit commitment, renewable integration, fuel-supply resilience scenarios, interconnection studies and steady-state studies.

I. INTRODUCTION AND MOTIVATION

Open transmission benchmarks largely represent North American and European grids, leaving Middle-Eastern systems without an openly reproducible case. This paper addresses that gap with a public-source synthetic Jordanian system calibrated to 2018 and validated for system-level studies.

  • No openly reproducible transmission test system exists for any Middle-Eastern grid, forcing researchers toward foreign benchmarks or unverifiable confidential data.
  • Existing benchmarks do not combine a dominant import corridor, an effectively isolated synchronous system, and renewable capacity concentrated at the far end of a long network.
  • Fuel supply, interconnections, and fleet availability are represented explicitly so disturbances can be formulated as input changes using public Jordanian sources.
  • The dataset reconstructs topology to its 2018-energized state, provides plant-level renewable fleets and hourly site profiles, and calibrates loads to published system totals.
  • The case is validated through structural statistics, power flow, N −1 screening, loss decomposition, and full-year production-cost comparison with published plant accounts.
  • The system supports system-level dispatch, resource adequacy, renewable integration, fuel-supply, interconnection, and steady-state studies but not actual-NEPCO operational or dynamic analyses.

A. Demand

Jordan’s 2018 system combined occupancy-driven demand with a gas-dominated generation fleet, limited interconnection exchanges, and a transmission network spanning two voltage levels.

  • 3,205 MW was the 2018 winter evening peak, compared with a 3,000 MW summer peak and approximately 1,290 MW minimum load.
  • 5,236 MW of available capacity included approximately 2,740 MW of gas-fired combined-cycle generation and roughly 980 MW of wind and solar capacity.
  • Natural gas supplied close to 88% of annual generation in 2018, with Samra the largest single producer followed by Amman East and Qatrana IPPs.
  • Interconnection exchanges were limited: Egypt supplied 188 GWh, Syria carried no exchange, and Jericho exported 88 GWh in 2018.
  • The system comprised 1,164 km-circuit of 400 kV line and 3,636 km of 132 kV circuit across 62 main substations, with reported transmission losses of 1.97%.

A. Topology Reconstruction

The model reconstructs Jordan’s transmission adjacency from a published single-line diagram, then restores voltage-level structure and 2018 topology using public records and reproducible electrical assumptions.

  • A. Topology Reconstruction: 87 edges spanning 68 numbered buses were recovered from the published diagram, with no islanded nodes and a degree distribution consistent with a real transmission graph.
  • A. Topology Reconstruction: 13 backbone substations are split into co-located 400 kV and 132 kV buses joined by transformers, expanding the model from 68 substation busbars to 81 solved nodes.
  • A. Topology Reconstruction: The 2018 restoration removes the 2019 New Ma’an loop, restores the direct Aqaba–Qatrana corridor, and adds the Qatrana–Amman West connection energized in Q4/2018.
  • D. Line Electrical Parameters: 132 kV lines use single-conductor ACSR Zebra, whereas 400 kV lines use twin bundles of large ACSR sub-conductors.
  • D. Line Electrical Parameters: Line electrical constants are derived from conductor geometry using standard overhead-line theory at 50 Hz, with assumptions summarized in Table II.
  • D. Line Electrical Parameters: Double-circuit corridors are modeled as aggregated branches for power flow and OPF but as parallel single circuits for N −1 analysis.
  • A. Topology Reconstruction: Incomplete public route data are resolved through a three-tiered length-assignment procedure using named project-table lengths and system circuit-kilometre totals.

E. Transformers

Transformers connect the reconstructed 400 kV and 132 kV backbone buses using standard ratings and typical impedances, while lower-voltage networks are aggregated at 132 kV.

  • 400/132 kV transformers use paired 400 MVA standard units, consistent with the documented 2×400 MVA installation at Amman West.
  • The 132/33 kV transformation and distribution systems are omitted, with 33 kV and bulk-supply-point load aggregated onto corresponding 132 kV buses.

F. Load Representation

The load model scales measured bulk-supply-point peaks to a calibrated 3,205 MW system peak, then derives reactive demand and shunt compensation from those active loads. Hourly demand is calibrated to published peak, annual energy, and load-factor values.

  • Spatial load allocation: 3,205 MW system peak is obtained by scaling measured bulk-supply-point peaks totaling 2,928 MW by 1.095.The participation-factor formulation uses each bus’s measured peak relative to the measured system total.
  • Reactive demand: Reactive demand uses a lagging power factor of 0.88, giving tan ϕ = 0.540.Reactive power is calculated as Q_i = P_i tan ϕ.
  • Reactive compensation: Shunt capacitor compensation is placed at larger load buses when P_i > 15 MW, with B_sh,i = κQ_i and κ = 0.75.The compensation is applied to buses exceeding the stated active-load threshold.
  • Bus composition: The 68-bus system contains 51 load buses and 17 pure transshipment or generation nodes, with roughly half the load in the northern region.This distribution describes the spatial structure represented by the load model.
  • Temporal demand calibration: Hourly demand is calibrated to a 2018 peak of 3,205 MW, annual purchased energy of 18,913 GWh, and an implied load factor of 0.674.The temperature-sensitivity coefficient is automatically bisected to match the load-factor target.

G. Generation Representation

Generation is assigned to host substations and represented with quadratic fuel-cost curves derived from heat rates, technology assumptions, commissioning year, and 2018 fuel prices. The resulting merit order places Risha first, imported-gas combined-cycle units next, and the HFO steam plant last.

  • Generator placement: Generators are placed at host substations using the bus-identity key and documented capacities, with Samra designated as the system slack bus.The three independent Al Manakher producers share the Amman East 400 kV substation.
  • Cost-curve formulation: Quadratic cost curves use I(P) = a+bP+cP^2 and C(P) = π_fI(P), with coefficients fitted to full-load and part-load heat-rate conditions.The assumed part-load degradation is K = 0.12, and heat rate is minimized at full output.
  • Technology assumptions: Full-load heat rates range from 6,400–6,700 Btu/kWh for combined-cycle units to 11,000–11,500 Btu/kWh for legacy gas turbines.Reciprocating engines are assigned 7,500–8,000 Btu/kWh, while conventional steam is assigned 10,200 Btu/kWh.
  • Fuel-price assumptions: 2018 fuel prices are set at 2.0 USD/MMBtu for Risha field gas, 5 USD/MMBtu for Egyptian pipeline gas, 8 USD/MMBtu for imported natural gas, 9 USD/MMBtu for LNG, and 12 USD/MMBtu for HFO.These prices reflect the modeled Jordanian supply sources and the Aqaba steam plant’s fuel.
  • Resulting merit order: Risha has the lowest full-load marginal cost, imported-gas combined-cycle units follow at 51–54 USD/MWh, and the HFO-fired steam plant is most costly.These marginal costs exclude capacity payments included in the published average purchased-energy cost of approximately 117 USD/MWh.

1) Renewable Fleet and Profiles:

The 2018 renewable fleet is modeled plant by plant with site-specific hourly solar and wind profiles. Utility-scale capacity matches NEPCO’s transmission-connected accounts, while wind outputs receive a fleet-level bias correction whose ramp implications remain approximate.

  • Fleet inventory: The utility fleet contains 447.5 MW of solar and 280.4 MW of wind, exactly reproducing NEPCO’s transmission-connected capacity accounts.Distributed 250 MW of net-metering and wheeling capacity is excluded.
  • Solar profiles: Solar profiles use site coordinates, NSRDB Meteosat irradiance, fixed-tilt PVWatts, 14% system losses, and a 1.2 DC/AC ratio with inverter clipping.The fixed-tilt representation reflects the predominant configuration of Jordan’s Round-1 fleet.
  • Wind profiles: Wind profiles use MERRA-2 reanalysis through Renewables.ninja with a uniform V112-class turbine at an 84 m hub height.The resulting profiles are generated separately for each plant.
  • Wind bias correction: A fleet-level bias factor of 1.226 scales wind output from the observed-to-modeled 2016–2017 energy ratio.Final output is capped at plant nameplate capacity, with saturation occurring only during the highest wind-speed hours.
  • Profile caveat: Wind ramp statistics are approximate because the 1.226 scaling amplifies ramp rates while reanalysis smooths subgrid variability.The paper states that the net error from these adjustments is ambiguous.

I. Bus–Substation Identity Key

The bus–substation identity key reconstructs unnamed topology buses by triangulating public maps, project tables, distribution-company lists, network structure, and spatial anchors. Structural validation then compares the resulting synthetic network with published real-grid criteria and system totals.

  • Identity reconstruction: Unnamed buses are assigned by combining named substations and lines, project tables, bulk-supply-point lists, network structure, and spatial anchors.Anchors include the Egypt tie at Aqaba, Jerusalem tie at Amman West, Syria tie at Amman North, and the Risha radial.
  • Validation framework: The released model is validated in pandapower and MATPOWER, with Table VI summarizing system dimensions and validation outcomes.Cross-solver agreement is used to check that results are not artifacts of one implementation.
  • Structural statistics: 76% of substations are load-only and 13% contain generation, while the degree distribution decays exponentially with maximum degree 7.The substation composition matches proportions reported for the Eastern Interconnect.
  • Network totals: Modeled 132 kV circuit-kilometres agree with NEPCO’s 3,636 km within +5%.This comparison ties the synthesized network dimensions to a published system total.
  • Structural criteria: The lines-per-node ratio is 1.21, close to the real-grid quota of 1.22 used in the structural criteria.The comparison evaluates the synthetic network against criteria developed from real grids.

B. Power Flow

The reconstructed Jordanian system passes the base-case AC power-flow checks and is largely N−1 secure, while OPF reproduces the expected merit order and identifies stressed corridors under contingencies.

  • B. Power Flow: 0.951–1.033 pu bus voltages remain within limits, with 40.91 MW of series losses and an 83% maximum line loading at the 2018 peak.The peak snapshot has no voltage violations, and the 400 kV and 132 kV buses are distinguished in the voltage profile.
  • C. Optimal Power Flow: $131,478/h AC OPF and $129,288/h DC OPF differ by −1.7%, while renewables dispatch fully and combined-cycle units carry most demand.The marginal CCGT sets a system price of approximately $52/MWh, and the dispatch cost is approximately $45/MWh.
  • D. N −1 Contingency Screening: 196 single-circuit and single-transformer outages yield 94% N−1 security with no islanding, while 12 outages cause post-contingency overloading.The binding cases concentrate on metropolitan 132 kV corridors and the Qatrana–Ma’an circuit serving the southern renewable cluster.
  • E. Time-Series Dispatch (Representative Weeks): 342 MW per hour is the maximum net-load ramp in the winter representative week, while spring solar reduces midday net load and keeps peakers off.The production-cost runs use hourly dispatch, with typical technology parameters for minimum stable levels and startup costs.

F. Annual Production-Cost Validation

The full-year production-cost validation reproduces aggregate generation closely without tuning, while revealing plant-level reallocations and the network location where future congestion-driven curtailment would first emerge.

  • F. Annual Production-Cost Validation: 8,760-hour coupled unit commitment and economic dispatch reproduce the annual energy fleet without adjusting cost parameters for agreement.The modeled annual energy by plant is compared directly with NEPCO’s published 2018 purchases.
  • F. Annual Production-Cost Validation: 3.8% is the maximum deviation among the three largest CCGT producers, while total modeled annual energy differs from reported energy by 0.1%.The modeled gas-fired share is 90%, compared with the reported 88%.
  • F. Annual Production-Cost Validation: The model’s monthly and representative-week outputs capture seasonal dispatch, including the summer cooling-driven peak and solar’s midday net-load reduction.The winter week shows combined-cycle cycling and peaking or steam units at daily peaks.
  • F. Annual Production-Cost Validation: Zero renewable curtailment matches 2018 system-level adequacy, while unconstrained hourly dispatch has no branch overload across 8760 hours and reaches 86% maximum loading.The PCM has no network constraints, so congestion-driven curtailment cannot occur; the Qatrana–Ma’an path is identified as the likely first location under higher penetrations.
  • F. Annual Production-Cost Validation: 3816 GWh modeled versus 3804 GWh reported gives the five-plant flexible segment a 0.3% difference despite internal plant reallocations.The model assigns some energy to cheaper committed alternatives rather than reproducing contracted engine-plant reserve behavior.

G. Transmission Losses

The model estimates series losses at 0.70%, while adding unmodeled components yields total losses of 1.2–1.5%, below NEPCO’s reported 1.97%. Cross-solver results agree to machine precision, but model assumptions and omitted network features limit interpretation.

  • 1.97% is NEPCO’s reported annual transmission loss, including series, transformer, and station losses.
  • 0.70% is the model’s energy-weighted series-loss estimate, corresponding to 132.7 GWh.The estimate uses twenty load levels, AC power flow, average dispatch, and energy weighting.
  • 1.2–1.5% is the model’s estimated total loss after adding transformer, distribution-transformation, and station-consumption components.The model underestimates actual losses by roughly half a percentage point.
  • Uniform conductor resistance, flat voltage setpoints, and estimated 132 kV line lengths contribute to the known loss shortfall.The case also assumes typical conductor and transformer parameters and omits plant-specific reactive capability curves and network constraints in production-cost modeling.
  • 4.9 × 10^-15 p.u. and 1.2 × 10^-14 degrees are the maximum cross-solver voltage-magnitude and slack-referenced angle differences.pandapower and MATPOWER report total system losses of 40.91 MW identically.
  • The synthetic model is intended for system-level studies rather than detailed operational representation.Its stated scope excludes system dynamics, inverter models, system-strength representation, and below-transmission-level analyses.
Loading 2609.00388v1…