Combustion tubes are the classical laboratory tools for evaluating in-situ combustion. However, high operational, capital, and maintenance costs, as well as, extensive material and analytical requirements have restricted their application. Recently, inexpensive methods for evaluating in-situ combustion using thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) have been described in the literature. TGA/DSC techniques have been used to determine combustion parameters such as fuel laydown and peak temperature and to quantify the kinetics of low temperature oxidation and combustion. The published TGA/DSC procedures often employ experimental conditions which are significantly dissimilar to those encountered in either a combustion tube or a reservoir process. Yet, limited data has been reported on the sensitivity of combustion parameters determined by TGA/DSC to experimental conditions such as heating rate, oxygen partial pressure and flow rate.
In this paper, the sensitivity of TGA/DSC procedures to experimental/operating conditions are evaluated. First, the operating conditions and techniques used in the literature are briefly reviewed. Next, the results of a laboratory study on the sensitivity of TGA/DSC procedures to heating rate (0.2 to 20 deg. C/min), oxygen partial pressure (0 to 400 psia), purge gas flow rate (75 to 1000 ml/min), matrix materials (0 to 90 wt%), and sample size (20 to 100 mg) are presented. This study details not only the effects of the experimental conditions on the thermal curves but also on the combustion parameters/kinetics calculated from them (fuel laydown, peak combustion temperature, heats of reaction, Arrhenius constants, and the extent of low temperature oxidation [LTO]). The observed sensitivities are examined in terms of heat/mass transfer limitations and reaction kinetics. As a consequence of the sensitivity study, TGA/DSC procedures are proposed that mimic the experimental conditions typically encountered in a combustion tube. Application of these procedures to eight viscous and heavy oils is discussed.
Introduction
During in-situ combustion, a fraction of the reservoir crude is burnt to generate heat and recover the remainder of the oil. Oil recovery results from (i) viscosity reduction due to heating and dissolution of produced gases, (ii) thermal expansion, (iii) distillation, (iv) thermal cracking, (v) steam/hot water drives, and (vi) the increased pressure gradient due to gas injection. The fuel for combustion is a carbon-rich residue (" coke") which is deposited on the reservoir rock as the result of thermal cracking, distillation, and low temperature oxidation of the oil ahead of the combustion front. Design parameters include the injection gas composition (air, oxygen-enriched air, water, and produced gases), the burn direction (injector to producer or vice versa), and the ignition method (electrical heaters, gas burners, spontaneous ignition).
The numerous chemical reactions coupled with simultaneous heat, mass, and momentum transfer, make in-situ combustion one of the most difficult enhanced oil recovery methods to simulate either physically or numerically. For many years combustion tube experiments have been the primary screening tools for evaluating in-situ combustion. Since heat loss is negligible in the reservoir, most combustion tube experiments are conducted under adiabatic conditions. Adiabatic combustion tubes generally consist of a length of thin-walled (less than 2 mm) steel tubing insulated and wrapped with heat tape and fitted inside a pressure jacket. The tube is packed with sand or native core material, saturated at reservoir conditions, and ignited. The progress of the combustion front is monitored by thermocouples placed at regular intervals along the axis of the tube at wall and center positions. The produced liquids and gases are continuously separated and analyzed.
Although combustion tube experiments are simplistic approximations of the reservoir process, they provide critical data for evaluating in-situ combustion. Since the combustion front will not propagate without sufficient fuel, perhaps the most frequent parameter measured is the fuel laydown or the quantity of coke available to the combustion front.
P. 323^
