D6.2 CO2 migration modelling on basin scale, Trøndelag Platform

Ane Lothe (SINTEF), Maarten Pluymaekers (TNO), Arnt Grøver (SINTEF), Ji-Quan Shi (Imperial), Joachim Rinna (SINTEF)

CO₂ migration modelling has been carried out for the Trøndelag Platform area, offshore Mid-Norway. The Garn Formation seems like a promising reservoir for CO₂ storage activities. The north-westwards dipping structure is characterised by a high sand content only moderately buried. Consequently porosity and permeability are excellent for CO₂ storage purposes. Formation thickness has been estimated to 100-150 m, and the number of faults is low on the Platform. In addition, the Garn Formation is overlaid by thick shale sequences further reducing fault leakage risk and also suggesting a low risk for cap rock leakage. The modelling has been carried out with different modelling tools:

• Migration modelling using PetroCharge Express (PetroMod), with no loss included
• Migration modelling using SEMI with implemented loss function for residual and density induced convection within the traps
• Pressure modelling using Eclipse 

A static geological model of the Trøndelag Platform was built incorporating the available data from wells, seismic interpreted depth horizons and fault polygons. The potential reservoir (Garn Fm.) was simplified and modelled as homogeneous sandstone. For three potential injection sites CO₂ migration scenarios were developed, using Petromod and SEMI. For SEMI additional loss functions were implemented, mimicking the physic-chemical behaviour of the (supercritical) CO₂ phase on the migration path and within trap entities. Pressure build-up was simulated using Eclipse in order to assure that injection scenarios not causing a seal rock fracturing risk.

In general, storage unit quality and capacity is so good that under industrial demo size injection volume scenarios of 1 Mt/a (over a period of 40 years) no CO₂ migrates out of the enclosures. If injection volumes are increased 5-fold to 5 Mt/a (also over a period of 40 years) the CO₂ phase starts to spill from the initial entities. As no loss for residual trapping is applied, and no density induced convection implemented in Petromod, migration distances are higher in PetroMod than in SEMI where such loss functions are implemented. However, loss functions may require tuning of the intrinsic function Ψ(ξ) respectively history matching. Pressure build-up is low to moderate and is not considered to cause any major risk for CO₂ storage on the Trøndelag Platform.