Numerical Study on Tackling Microbial Reservoir Souring during Engineered Water Injection

Summary

Sulfate-reducing bacteria (SRB) activity in reservoirs causes several challenges related to reservoir souring during waterflooding. In the present study, a biochemical numerical model was developed to capture a laboratory continuous upflow bioreactor using suitable microbial growth and metabolite kinetic models. The capabilities of modeling microbial souring treatments at the laboratory and field scales during engineered water injection (EWI) were explored. Two 1D laboratory-scale models were built. The first, implemented in CMG-STARS™, captured the mechanistic examinations of SRB and nitrate-reducing sulfide-oxidizing bacteria (NR-SOB) activities in a laboratory bioreactor (data of Hubert et al. 2003). This model was tuned using a division factor and a reaction-rate constant, but was limited by the lack of formulation of the Monod equation. The second, implemented in the MATLAB Reservoir Simulation Toolbox (MRST), was developed to address this limitation and was validated against the continuous upflow packed bed bioreactor experiment of Chen et al. (1994), with continuous sulfate and volatile fatty acid (VFA) feed.

In the CMG-STARS 1D model, the treatment was deemed successful since NR-SOB commenced to grow as the nitrate was injected gradually. This resulted in complete mitigation of the hydrogen sulfide (H2S) generated, supported by the NR-SOB oxidation. The 1D model was tuned using a division factor and a reaction rate constant to better match the experimental data for H2S and sulfuric acid (H2SO4), but it has shortcomings due to the lack of explicit incorporation of the Monod equation. These shortcomings were addressed through a Chen 1D history match using a different simulator. The results showed that the 1D model was successful in history matching the increase in the generated H2S at the end of SRB growth duration with a root mean square error (RMSE) of 12.95 mg/L and a coefficient of determination (R2) of 0.80, which shows a better match than the General Purpose Adaptive Simulator (GPAS) simulation with an RMSE of 13.35 mg/L and an R2 of 0.79. For the 3D model, the thermal viability shell (TVS) model was implemented to capture temperature-dependent SRB activity. The results showed that mixing between cold-injected engineered water (EW) (25°C) and formation brine (90°C) creates a thermal front where temperatures fall within the viable range (20–80°C), enabling SRB metabolism. The findings showed that temperature reduction from mixing between injected and formation waters triggered H2S generation at the injector. Subsequently, it was observed that the front moved until breakthrough and then sharply dropped as most of the VFA was consumed. Colder injection temperatures produced higher H2S concentrations because the resulting mixing zone temperature is closer to the optimal SRB growth temperature. The base case (25°C injection) produced the highest H2S concentration (48.9 mg/L), followed by 40°C (7.2 mg/L) and 50°C (1.8 mg/L). This study incorporates the effects of SRB growth, injected water temperature, and reservoir heterogeneity into H2S production within a unified biochemical model. This approach offers a straightforward yet comprehensive workflow for predicting reservoir souring. By addressing fundamental mechanisms often overlooked, the proposed method advances field operations and broadens understanding of reservoir management during EWI.