Simulation of Laboratory In-Situ Combustion Data and Effect of Process Variations

Abstract

ABSTRACT A fully implicit thermal simulator was validated by history matching laboratory combustion tube data. The validated simulator was used to identify important parameters that influence the predicted results and to study the effects of process variations. Measured cracking reaction parameters were used to minimize the number of adjustable parameters in history matching. Results indicate that the simulator can history match the laboratory data when the relative permeabilities are modified. Matched experimental parameters include cumulative oil and water production histories, location of combustion front as a function of time, and peak temperatures along the tube. The simulation results were insensitive to the kinetics of combustion, grid size, sand thermal conductivity, and equilibrium coefficient of the heavy oil pseudocomponent. The results, however, were sensitive to relative permeability, capillary pressure near the irreducible water saturation, and equilibrium coefficient of the light oil component. Predicted trends for oxygen injection, wet combustion, and variation in injection rates agree very well with published experimental results. Injection of oxygen instead of air (for nearly the same combustion front velocity) caused a significant increase in the fuel consumption and peak temperatures. Wet combustion generally caused a decrease in the fuel consumption and the average peak temperature, and an increase in the combustion front advance and production rates. The velocity of the combustion front was proportional to the air flux at the combustion front, except at very high injection rates. This implies that at field-like injection rates, the combustion process is air (oxygen) flow rate controlled.