Skip to the content.

Consolidated Benchmark Results

This page summarizes benchmark validations from all task-solving studies in task_solve/. Each entry links to the source task for full details.

For the curated trust dashboard with acceptance criteria, see index.md.


1. Thermodynamic Property Validation (vs NIST)

Property Fluid Conditions NeqSim EOS Max Error Source Task
Density Methane 200-350 bar, 0-50 °C SRK < 5% 2026-03-07_cng_tank_filling
Cp Methane 200-350 bar, 0-50 °C SRK < 8% 2026-03-07_cng_tank_filling
Viscosity Methane 200-350 bar, 0-50 °C SRK < 10% 2026-03-07_cng_tank_filling
Thermal conductivity Methane 200-350 bar, 0-50 °C SRK < 15% 2026-03-07_cng_tank_filling
Density Natural gas Pipeline conditions SRK < 3% 2026-03-07_npv_subsea_tieback
Z-factor Natural gas Pipeline conditions SRK < 2% 2026-03-07_npv_subsea_tieback

Key finding: SRK EOS performs well for gas-phase density and compressibility at pipeline conditions (< 3% error). Transport properties (viscosity, thermal conductivity) show higher deviations at extreme pressures — consider PR or GERG-2008 for custody transfer accuracy.


2. Phase Equilibria and Solubility

System Property Reference Data NeqSim EOS Points Pass Rate Tolerance Source Task
CO2–Water CO2 solubility in water Published experimental SRK-CPA, Electrolyte-CPA Multiple Good 2026-03-09_h2s_co2_distribution
H2S–Water H2S solubility in water Published experimental SRK-CPA Multiple Good 2026-03-09_h2s_co2_distribution
CO2–Water Water solubility in CO2 Wiebe & Gaddy (1941) SRK-CPA 13 69% (9/13) 30% 2026-03-09_water_solubility_co2
CH4–Water Methane solubility in water Culberson & McKetta (1951), Lekvam & Bishnoi (1997), IUPAC SRK-CPA 18 Good NCS conditions 2026-04-09_mimee_methane_emissions

Key finding: SRK-CPA handles gas solubility in water well for engineering purposes. Water-in-CO2 predictions at high pressure show larger deviations (30% tolerance needed) — use Electrolyte-CPA for improved accuracy in CCS applications.


3. Process Simulation Benchmarks

Equipment Method Reference Max Error Source Task
Adsorber (mercury removal) NTU-efficiency Analytical formula < 1% 2026-03-08_mercury_removal_lng
Packed bed Ergun equation dP Hand calculation < 2% 2026-03-08_mercury_removal_lng
Packed bed Bed lifetime Mass balance estimate Good agreement 2026-03-08_mercury_removal_lng
Pipeline Pressure drop (Beggs & Brill) Textbook correlation < 5% 2026-03-07_npv_subsea_tieback
Sulfur deposition Arrhenius S8 solubility Literature correlation Reasonable 2026-03-09_draupner_sulfur
Compressor train Polytropic efficiency Booster benchmarks < 3% 2026-03-21_compressor_train

Key finding: NeqSim process equipment models agree well with hand calculations and textbook correlations. Pipeline pressure drop with Beggs & Brill shows < 5% deviation from published correlations.


4. Cost Estimation Benchmarks

Category Reference Data Agreement Source Task
SURF cost (NCS) Industry benchmarks Within typical range 2026-03-07_npv_subsea_tieback
Norwegian tax model Hand calculation Exact match 2026-03-07_npv_subsea_tieback
Well mechanical design API 5C3 formulas Verified per standard 2026-03-07_npv_subsea_tieback

5. Plant Data Validation

Facility Measured Property Data Source Agreement Source Task
Generic platform Methane emissions (dissolved gas) Historian data Good agreement at NCS conditions 2026-04-09_mimee_methane_emissions
Production facilities Emission factors Plant data comparison Within operational range 2026-04-17_mimee_code_update_verification

Key finding: NeqSim methane solubility predictions align with plant historian data at typical Norwegian Continental Shelf conditions (50-150 bar, 30-80 °C, produced water systems).


6. Renewable Energy and Concept Studies

Study Benchmark Type Reference Source Task
CNG tank filling Temperature prediction Churchill-Chu natural convection 2026-03-07_cng_tank_filling
CNG tank (composite) Density predictions NIST WebBook 2026-03-18_composites_cng_tank
Floating wind Concept study metrics Industry benchmarks 2026-03-19_utsira_nord_wind

Summary Statistics

Category Tasks Total Data Points Typical Accuracy
Thermo properties vs NIST 4 ~30 < 5% (density), < 15% (transport)
Phase equilibria 3 ~40 < 10% (gas-in-water), < 30% (water-in-gas)
Process simulation 4 ~15 < 5%
Cost estimation 1 3 Within industry range
Plant data 2 ~20 Within operational tolerance

Contributing Benchmarks

When completing a task with benchmark validation:

  1. Add benchmark_validation key to results.json
  2. Add a row to the appropriate table in this file
  3. Include the source task folder name for traceability
  4. State the reference data source, number of points, and tolerance