Indicative CO₂ pipeline dispersion modelling

1. Animation results

Each panel shows the modelled CO₂ concentration at 0.5 m above ground level (typical breathing height) overlaid on the OpenStreetMap basemap of the local area. Colour scale is logarithmic from 5,000 ppm to 300,000 ppm. The blue star marks the assumed release point. Title states the simulation time, peak concentration in the frame, and the modelling assumption set used.

For each location, two animations are presented (v1 and v2) bracketing the realistic operator design envelope — see §3 below for the details of the difference.

2. Methodology overview

The dispersion chain comprises three sequential physics modules, each calibrated against published reference data and large-scale field experiments where available:

1. Stage 1 — Pipeline blowdown: lumped control-volume model of CO₂ depressurisation from a hypothetical guillotine rupture, integrated over the inventory between the upstream and downstream emergency shutdown valves (ESDVs). Provides the time-varying mass flux ṁ(t) at the orifice.
2. Stage 2 — Equation of state: Span & Wagner (1996) reference EOS for pure CO₂ via CoolProp, supplying density, enthalpy, entropy and sound speed at every (T, p) state encountered during decompression.
3. Stage 3 — Atmospheric dispersion: TWODEE-2 (Folch et al. 2009) shallow-layer dense-gas model, run on terrain-resolving meshes derived from EA Open LIDAR, with surface roughness classified per cell from raw LAS point-cloud data.

Each stage is independently calibrated; the integrated chain has been benchmarked against Spadeadam test data for the blowdown trajectory and against laboratory and field dense-gas experiments for the dispersion stage.

3. The two source assumptions: pipeline-uncertainty bracketing

Two engineering assumption sets — a baseline (v1) and a more realistic operating envelope (v2) — were run for each location to bracket the uncertainty in operator design parameters that have not yet been publicly disclosed:

The v2 assumptions reflect typical industry practice for high-pressure CO₂ pipelines in densely populated corridors (DNV/IGEM/GASNOVA joint guidance, 2021) and the operating realities of SCADA leak detection plus electric actuator valve travel. The factor-2 difference in total released mass between v1 and v2 is the single largest source of consequence-modelling uncertainty prior to formal Development Consent Order disclosure of the operator's design.

4. Spadeadam validation

Spadeadam (Cumbria) is the UK's national large-scale gas safety test facility, operated by DNV on behalf of HSE, industry and academic research programmes. Two consecutive joint-industry programmes have established the empirical reference dataset for dense-phase CO₂ release behaviour:

• COOLTRANS (2010–2014): 84 controlled high-pressure CO₂ releases through orifices of 6–150 mm at storage pressures up to 150 bar. Measured discharge rate, plume temperature, concentration profile and dry-ice fraction.
• CO2PIPETRANS / CO2-FREE (2014–2019): full-bore pipe ruptures and leak experiments at near-realistic CCS pipeline conditions. Provided the primary validation dataset for current CO₂ dispersion codes.

The blowdown model used here was independently validated against the Spadeadam-derived discharge bands at five pressure conditions (40, 60, 80, 100 and 90 bar): peak ṁ, decay envelope and total inventory release matched the reported bands within experimental scatter for all five cases.

5. Blowdown model

The Stage 1 blowdown model treats the inventory between the two ESDVs as a single lumped control volume, integrating mass and energy conservation forward in time as gas escapes through the rupture orifice. Real-gas thermodynamics throughout the depressurisation are obtained from CoolProp's implementation of the Span & Wagner reference EOS, with the homogeneous equilibrium model (HEM) and Wood's sound speed used at the choked orifice to capture two-phase flow as the pipe pressure crosses the saturation curve.

For the scenarios presented here:

6. TWODEE-2 dispersion model

TWODEE-2 is the open-source shallow-layer Eulerian dense-gas dispersion code developed at HSL (HSE's research laboratory) by Hankin & Britter (1999) and subsequently re-engineered and released by INGV (Folch et al. 2009). The shallow-layer formulation correctly captures the dominant physics for ground- hugging dense gas — gravity slumping, terrain channelling, surface friction and entrainment — at a small fraction of the computational cost of full three-dimensional CFD.

Data sources

Model resolution

The animations above use 10 m horizontal grid spacing across each domain. At this resolution individual streets and buildings are resolved, gravity slumping along valleys is captured, and the LIDAR-derived roughness field correctly represents Davenport class 7–8 (UK market-town) fabric in the population centres.

7. Run setup

Three population centres along or near the proposed pipeline route were modelled at 10 m resolution: Macclesfield (8 × 8 km asymmetric domain with 1 km of footprint to the west and 7 km to the east of the release point, to capture the eastward westerly wind drift), Whaley Bridge (7 × 7 km symmetric), and Disley (7 × 7 km symmetric). Each run spans 3 hours of simulation with 30-second output cadence and uses a 50 m × 50 m source footprint at the release point.

Wind conditions were chosen to push the cloud toward each town centre under realistic but stress-test directions (light westerly through Macclesfield; light northerly through Whaley Bridge; light westerly through Disley village).

8. Forthcoming runs

A wider Wirral peninsula domain (18 × 18 km centred on the proposed export terminal area, modelling the higher-pressure 140 bar export-end blowdown) is currently being prepared and will be added to this page once complete. A formal Monte Carlo over the joint wind-direction × wind-speed × stability × release-point distribution (~50,000 runs at 30 × 30 km, 50 m, 2-hour sims) is planned as the next phase, to produce probabilistic individual and societal risk contours at regulator-grade resolution.

References

• Span, R. & Wagner, W. (1996). A new equation of state for carbon dioxide covering the fluid region from the triple-point temperature to 1100 K at pressures up to 800 MPa. J. Phys. Chem. Ref. Data 25(6):1509–1596.
• Bell, I. H. et al. (2014). Pure and pseudo-pure fluid thermophysical property evaluation and the open-source thermophysical property library CoolProp. Ind. Eng. Chem. Res. 53(6):2498–2508.
• Folch, A., Costa, A. & Hankin, R. K. S. (2009). TWODEE-2: a shallow layer model for dense gas dispersion on complex topography. Computers & Geosciences 35(3):667–674.
• Hankin, R. K. S. & Britter, R. E. (1999). TWODEE: the Health and Safety Laboratory's shallow-layer model for heavy gas dispersion. J. Hazardous Materials 66(3):211–226.
• Britter, R. E. & McQuaid, J. (1988). Workbook on the Dispersion of Dense Gases. HSE Contract Research Report No. 17/1988.
• DNV (2014–2019). COOLTRANS / CO₂PIPETRANS / CO2-FREE Joint Industry Programmes — Spadeadam dense-phase CO₂ release tests. Report series.
• UK Environment Agency (2024). National LIDAR Programme — 1 m DTM/DSM composite and classified point-cloud data. Open Government Licence.
• Skamarock, W. C. et al. (2021). A description of the Advanced Research WRF Model Version 4.7. NCAR Technical Note NCAR/TN-556+STR.