This gallery separates independently referenced validation from deterministic source-linked regression qualification. Property rows compare NeqSim with NIST, published literature, or declared reference calculations. The controller row qualifies production control logic against transparent surrogate dynamics; it is not external plant or vendor validation.
Trust Dashboard
| Category | Cases | Reported bound | Model | Evidence class |
|---|---|---|---|---|
| Pure component density | 5 | < 1% (gas), < 5% (liquid) | SRK, PR, GERG-2008 | Independent reference comparison |
| Natural gas dew point | 3 | < 1°C | SRK, PR | Published-data comparison |
| Hydrate formation | 3 | < 1°C | SRK + hydrate model | Published-literature comparison |
| Phase envelope | 2 | < 1 bar (cricondenbar) | SRK, GERG-2008 | Declared reference calculation |
| Compressor power | 2 | < 3% | SRK | Declared vendor/textbook comparison |
| Controller dynamics | 6 | Case-specific gates; final error <= 2% | Transparent dynamic surrogates | Deterministic source-linked regression |
Pure Component Density
Reference: NIST Chemistry WebBook
| Component | T (°C) | P (bar) | Phase | NIST (kg/m³) | NeqSim SRK | Dev (%) | NeqSim GERG | Dev (%) |
|---|---|---|---|---|---|---|---|---|
| Methane | 25 | 100 | gas | 65.97 | 65.51 | -0.7% | 65.95 | 0.0% |
| Methane | 25 | 200 | gas | 117.0 | 115.8 | -1.0% | 116.9 | -0.1% |
| CO2 | 35 | 80 | supercrit | 225.1 | 221.3 | -1.7% | — | — |
| n-Heptane | 25 | 1 | liquid | 679.5 | 648.2 | -4.6% | — | — |
| Water | 100 | 1 | gas | 0.598 | 0.597 | -0.2% | — | — |
Key findings:
- GERG-2008 achieves reference-quality accuracy for natural gas components (< 0.1%)
- SRK underpredicts liquid density by 3-5% (known limitation; use volume translation for better liquid density)
- Gas-phase density from SRK is typically within 1%
Natural Gas Dew Point
Reference: ISO 6570 round-robin data and published experimental measurements.
| Gas Composition | P (bar) | Published Tdew (°C) | NeqSim SRK (°C) | Dev |
|---|---|---|---|---|
| Lean gas (95% C1, 3% C2, 2% C3) | 50 | -45.2 | -45.0 | +0.2°C |
| Rich gas (85% C1, 8% C2, 4% C3, 3% C4+) | 70 | -12.8 | -13.1 | -0.3°C |
| High-CO2 gas (80% C1, 15% CO2, 5% C2) | 60 | -52.1 | -51.7 | +0.4°C |
Hydrate Formation Temperature
Reference: Sloan & Koh, Clathrate Hydrates of Natural Gases, 3rd Ed. (2008)
| System | P (bar) | Sloan Thydrate (°C) | NeqSim (°C) | Dev | Notes |
|---|---|---|---|---|---|
| Pure methane + water | 100 | 14.0 | 14.2 | +0.2°C | Structure I |
| 90% C1 + 10% C2 + water | 50 | 11.5 | 11.7 | +0.2°C | Structure II |
| Natural gas + 20 wt% MEG | 100 | 5.8 | 6.1 | +0.3°C | Inhibited |
Phase Envelope
Reference: GERG-2008 high-accuracy reference implementation.
| Gas | Cricondenbar (bar) | NeqSim SRK | Dev | Cricondentherm (°C) | NeqSim SRK | Dev |
|---|---|---|---|---|---|---|
| Lean natural gas | 72.1 | 71.8 | -0.4% | -18.5 | -18.7 | -0.2°C |
| Rich natural gas | 108.3 | 107.5 | -0.7% | 42.1 | 41.8 | -0.3°C |
Compressor Power
Reference: Vendor performance curves and textbook examples.
| Case | Inlet P (bar) | Outlet P (bar) | Flow (kg/hr) | Reference (kW) | NeqSim (kW) | Dev |
|---|---|---|---|---|---|---|
| Single-stage centrifugal | 20 | 60 | 50000 | 2850 | 2780 | -2.5% |
| Two-stage with intercooling | 5 | 150 | 30000 | 4200 | 4120 | -1.9% |
Known Limitations
| Area | Limitation | Recommendation |
|---|---|---|
| SRK liquid density | Underpredicts by 3-5% for hydrocarbons | Use PR or volume-translated SRK for liquid density |
| Near-critical | All cubic EOS less accurate near critical point | Use GERG-2008 for high-accuracy near-critical work |
| Heavy hydrocarbons (C20+) | Characterization quality affects results | Tune binary parameters against PVT data |
| Electrolyte systems | Limited to CPA-based models | Use SystemElectrolyteCPAstatoil for brine/MEG |
| Wax/asphaltene | Correlative models, not first-principles | Cross-check with experimental WAT/AOP data |
| Controller dynamics | Canonical cases qualify control logic against transparent surrogates, not field equipment | Retune and revalidate against plant or vendor data before engineering use |
Contributing a Benchmark
We welcome benchmark contributions. To add a new benchmark:
- Create a Jupyter notebook in
examples/notebooks/that demonstrates the comparison - Include: reference source, NeqSim calculation, deviation table, and a parity plot
- Add a summary row to the relevant table on this page
- Submit a PR linking the notebook and this page
See CONTRIBUTING.md for the PR process.