A major simulation study was undertaken to examine the sensitivities of various important parameters on the dry forward combustion enhanced recovery process conducted on a field scale. The reservoir fluid is a L1oydminster-type oil having 14 degree API gravity. A fully-implicit thermal reservoir simulator (ISCOM) was used to perform the sensitivity analyses. Among the reservoir properties studied were pay thickness, vertical permeability variation and initial oil saturation. Variations in well pattern size and configuration, air injection rates and injection fluid (air or steam) were also examined. To determine the chemical reaction parameters for the simulations a combustion tube test was matched with ISCOM. A technique was devised to model the combustion process in a large field-size grid. It was found that the vertical permeability variation could be modelled adequately with three layers. Results from areal and cross-sectional models were combined to obtain a three-dimensional analysis of the fire flood performance for each nine-spot field pattern.
Model predictions indicate that there is no technical advantage in using pattern areas smaller than 10 acres for 10 ft. net pay, and 20 acres for 25 ft. net pay. Below these pattern sizes, recoveries of about 40% OOIP and 45% OOIP, respectively, are obtained_ Above these pattern sizes extinction of the fire front significantly reduces areal sweeps and oil recoveries, resulting in maximum project lifetimes of about 4.2 years and 10 years for the two pay thicknesses.
We found that the vertical sweep efficiency is predominately controlled by the vertical permeability variation in the reservoir, and not by steam or air override effects. The detailed simulator output showed that oil vaporization, cross flow between layers and heat loss to the adjacent formations are important in this thermal recovery process.
The objectives of this study were to derive a data base for the simulation of thermal EOR processes in L1oydminster-type heavy oil reservoirs, and to investigate the sensitivity of oil recovery performance of the combustion process to various parameters.
It was necessary to establish a proper data base for the sensitivity study. Viscosity-temperature data, distillation data and literature references were used to characterize the heavy oil. A combustion tube experiment helped determine an adequate reaction model. The methods and correlations used to obtain the physical properties are discussed in detail.
The sensitivity parameters were divided into three functional categories: model, reservoir and operational parameters. The model parameters arise from the discretization of time and space. Reservoir parameters are specific to the reservoir of interest, and include vertical variations in rock properties and initial conditions.
Practical considerations such as well spacing, pattern configuration and injection and production well specifications comprise the operational parameters. The model parameters were optimized to give a balance of accuracy and economy satisfactory for the study of the reservoir and operational parameters. To reduce computing cost the three-dimensional combustion problem was decoupled into areal and vertical cross-sectional ones, whose sweep efficiencies were then recombined in a consistent manner.
With the optimized model parameters, the reservoir and operational sensitivity parameters were then studied and the results were examined.