American Institute of Mining, Metallurgical and Petroleum Engineers Inc.
Considerable experimental and theoretical effort has been devoted to understanding the heat flow and stoichoimetric problems associated with the underground combustion process of oil recovery. Results of these efforts have been vital to success in the engineering of field applications of underground combustion.
Continued success in applying this process to diverse petroleum reservoirs would be aided by a quantitative understanding of the associated fluid flow phenomena. Semiquantitative notions about the oil displacement mechanisms and the nature of the "oil bank" are presently employed in the evaluation of underground presently employed in the evaluation of underground combustion prospects. A quantitative delineation of the fluid flow aspects would facilitate better engineering predictions of:
Injectivity history (compressor requirements),
Fuel content dependence on air flux.
A reasonably complete theoretical treatment of the in situ combustion process would require simultaneous solution of the heat flow, distillation, and three-phase fluid flow problems. Further, this should be done for a multidimensional medium which is not necessarily homogeneous or isotropic. While it is perfectly feasible to develop the necessary equations for such a model, their general solution was much too ambitious an undertaking for the computing capacity available when this work was begun.
We therefore confined this effort to exploring the numerical methods required and to becoming familiar with the displacement mechanisms that would be evidenced in the solution of a greatly simplified mathematical model. Specifically, the following restrictions were adopted:
Two-phase fluid flow (oil and air only).
No distillation effects.
Temperature profiles obtained from an independent study of the associated heat-flow problem.
To provide a basis for evaluating the results of the model study, a special laboratory tube run experiment was conducted. Here the sandpack was initially saturated with oil only, omitting the usual "connate" water. During burning, the only water present was the water of combustion. The experiments were conducted under psuedo-adiabatic conditions for which temperature distributions are known.