Abstract
Toe-to-heel air injection (THAI) is an in situ thermal-enhanced oil upgrading and recovery technology that makes use of a horizontal producer well for mobilised partially upgraded oil production to the surface. To design operation of the THAI process requires up-scaling lab-scale numerical model to field-scale reservoir numerical model. There are a number of up-scaling methods that were developed. However, these previous up-scaling methods only concentrated on the qualitative nature of the physicochemical processes. Furthermore, these methods have deficiencies ranging from assumption of 100% oxygen utilisation, failing to compare oil production rates for up-scaled reservoir models with and without fuel pre-defined, not reporting on the dynamics inside the reservoirs with and without fuel pre-defined, etc. As a result, this work provided in depth comparative studies of both the qualitative and quantitative physicochemical processes in numerical models containing Arrhenius-based coke deposition thermal cracking reaction (i.e. model A) and that having pre-defined fuel concentration (i.e. model F) as an initial parameter respectively. It is found that oil production started earlier in model A compared to in model F even though they have approximately similar trajectories. It is shown that by the end of the combustion period, oil recovery of 28.5% OOIP is achieved in model F which is greater than that in model A by an additional of 4.8% OOIP. It is found that the peak temperature, which indicates how healthy the combustion is at a single point only at different times, is generally considerably higher in model A than that in model F. It is further discovered that in some cases the combustion at all points might be stable and rigorous but in other cases the combustion might not be stable and rigorous at all points thereby implying that the peak temperature alone does not give the measure of the efficiency and spread of the combustion front. It is shown for the first time that pre-defining fuel concentration throughout the reservoir as is done in model F resulted in faster downward advancement and expansion of the combustion front compared to in the reservoir where the fuel is deposited based on the Arrhenius kinetics parameters (i.e. like in model A). It is also discovered that higher areal sweep of the reservoir is achieved in model F when compared to that achieved in model A. Overall, it then follows that higher volume of the reservoir of model F is combustion-swept, heat-affected, and oil-depleted compared to that in model A. These findings explain the reason why more oil is produced during the combustion period in model F than in model A. Conclusively, this first-of-a-kind study shows that pre-defining the fuel has no effect on the qualitative nature of the physicochemical processes, but has considerable influence on the quantitative parameters such as the rate of and cumulative oil production, peak temperature, areal and volumetric sweeps by the combustion zone, size of the oil-depleted zone, and size of the heat-affect zone.
•Full analysis of predictive ability of THAI model F with pre-defined fuel is reported.•Model F predicted unrealistic high oil production rate & thus cumulative oil recovery.•Model F has faster downward-advancement & general expansion of the combustion fronts.•Higher reservoir volume is combustion-swept, heat-affected, & oil-depleted in model F.•Qualitative behaviours in model F are similar to those in the Arrhenius-based model A.