Reactivity of carbonate reservoir rocks at Weyburn

CO2 is being injected into the Midale Formation carbonate reservoir as part of an enhanced oil recovery project at the Weyburn oilfield, Saskatchewan, Canada. As this is an active and mature oil field, a lot of wells are present, core samples are available and fluid samples can be taken at almost any time. A three-pronged approach has been used by the North American and European teams as part of the IEA CO2 monitoring and storage project (Wilson and Monea, 2004) to study the impact of CO2 upon reservoir geochemistry: monitoring changes in actual reservoir fluids from deep boreholes; laboratory experiments to simulate in-situ conditions within the reservoir; and predictive modelling of evolving conditions within the reservoir over the short to long term. A selection of results is presented below.

10"" 10:' 101 1CT 10' 10" 10s 106 107 10° 10:' 10"-Time (s)

Figure 3. Batch thermo-kinetic modelling reacting a CO2-rich Weyburn brine (1 molal CO2) with reservoir minerals at 50°C over 1000 years: Evolution of pH and total amount of carbon in dissolved and mineral forms, expressed in mol/REV (Durst et al., 2003). A REV is a Representative Elementary Volume of 3.85 l corresponding to 1 l of fluid and 2.85 l of rock.

The potential for long term mineral trapping in the Weyburn reservoir has been assessed by numerical modelling. Figure 3 presents the results for the case where a typical Weyburn brine (50°C, salinity 81 g/l) saturated with 1 mole of dissolved CO2 reacts with the reservoir minerals in a closed system over 1000 years. Calculations were carried out using the PHREEQC code. Key results from this modelling are as follows (Durst et al., 2003):

• A rapid thermodynamic re-equilibration of the CO2-rich saline solution with respect to carbonates, sulphate and sulphide is predicted. The resulting quick dissolution of carbonates induces an increase in dissolved carbon, whilst the amount of carbon in mineral form decreases.

• slow dissolution of alumino-silicates (Fe-chlorite, illite and K-feldspar), allowing the precipitation of dawsonite and siderite which trap CO2 in mineral form. After about 3 years of simulated time, the predicted carbon concentration in solution continuously decreases while more and more carbon is trapped in mineral form.

• as the system is closed, the total amount of carbon in the system is constant.

• it is predicted that 50% of the CO2 initially dissolved in the brine will be trapped in mineral phases (siderite, dawsonite) after 1000 years,

• however, thermodynamic equilibrium is still not reached after 1000 years.

• no significant variation of porosity was predicted as the system was closed (batch calculations). However for open systems with significant advective flow, carbonate reactions could lead to noticeable porosity changes.

Extrapolation to the reservoir scale was attempted by Perkins et al. (2004). Planned CO2 injection is on the order of 20-25 million tons. In the long term, the Weyburn Midale reservoir has the potential to store all of the injected CO2 through solubility and mineral trapping mechanisms, but assuming that every part of the reservoir is in contact with dissolved CO2.

A series of laboratory experiments has been undertaken where samples of well-characterised borehole material from the Midale Formation were reacted with both CO2 and synthetic reservoir formation waters under simulated in-situ conditions (60°C, 15-25 MPa). The duration of the experiments ranged from 1 week to 6 months. Changes in fluid chemistry, mineralogy, porosity and permeability were measured. The basic layout of the batch experiments is reproduced in Figure 4. Core flooding experiments with a fluid velocity of 15-29 cm/day were also carried out. Interpreted results of these experiments, based on observed changes and numerical modelling using the SCALE 2000 and PHREEQC codes, are described below (Azaroual et al., 2004; Riding and Rochelle, 2005).

CO2 h2o inlet outlet

CO2 h2o inlet outlet

Figure 4. Layout of the batch experiments used to react rock samples with CO2-saturated brines under reservoir temperature and pressure conditions. Vessel volume is 100-150 ml.

In the closed system batch experiments the Midale lithologies (limestones, dolostones and anhydrite) showed variable amounts of carbonate mineral dissolution, some precipitation of gypsum incorporating the Ca released from carbonate minerals (Figure 5), 'inhibition' of anhydrite dissolution, and poor reactivity of alumino-silicate minerals over the timescale of the experiments.

Figure 5. Gypsum crystals up to 2.5 mm long after 8 weeks reaction of Weyburn reservoir rock with CO2 in batch experiments under reservoir conditions (60°C, 15 MPa) (Riding and Rochelle, 2005).

In the CO2 flooding tests observed dissolution features were confined to the inlet end, where CO2-saturated brine entered the sample (Figure 6). Calcite and dolomite showed various states of corrosion. Porosity and matrix gas permeability increased. No precipitating minerals were observed in the CO2 flooding experiments.

The 1D PHREEQC reactive transport code was used to simulate the long-term evolution of Midale material under CO2 injection conditions in the Midale Formation of the Weyburn reservoir (150 bar and 54°C). The effect of mineral dissolution / precipitation reaction kinetics coupled to advective / diffusive / dispersive transport processes was predicted over a simulated 10,000 year time period. Varying flow rates were considered, from the undisturbed natural aquifer flow (0.25 m/year) up to higher values close to the injection zone (50 m/year). Simulations show that main reactions occur close to the injection zone. Calcite dissolves at the very entrance of the system whereas a calcium sulphate phase (anhydrite or gypsum) precipitates. Feldspars dissolve over a long distance while dawsonite precipitates. Clays show two reaction fronts, dissolving close to the injection zone but precipitating further on. Chalcedony, a silica phase, has the opposite behaviour. Finally, as a consequence of these dissolution / precipitation reactions, the porosity is predicted to increase. In the case of a flow rate of 0.25 m/year, the initial 20% porosity is expected to increase in the first 10 meters of the 1D column by up to 20.3% after 1000 years, and by up to 25% after 10,000 years.

Figure 6. Inlet end of the Weyburn reservoir rock sample after the CO2 flooding experiment carried out at GEUS. Numerous small pits are present in the sample surface after the experiment. They represent dissolved calcite grains. (Riding and Rochelle, 2005; photograph reproduced with kind permission of GEUS).

Figure 6. Inlet end of the Weyburn reservoir rock sample after the CO2 flooding experiment carried out at GEUS. Numerous small pits are present in the sample surface after the experiment. They represent dissolved calcite grains. (Riding and Rochelle, 2005; photograph reproduced with kind permission of GEUS).

Further investigations will be carried out in the Weyburn Phase II project. More accurate estimates coupling the CO2 plume displacement, the CO2 dissolution process as well as the geochemical interactions will be necessary to investigate the ultimate fate of the injected CO2.

Was this article helpful?

0 0

Post a comment