An In-Situ Combustion Reservoir Simulator With a New Representation of Chemical Reactions
- Pierre A. Le Thiez (Inst. Francais du Petrole) | A. Lemonnier (Inst. Francais du Petrole)
- Document ID
- Society of Petroleum Engineers
- SPE Reservoir Engineering
- Publication Date
- August 1990
- Document Type
- Journal Paper
- 285 - 292
- 1990. Society of Petroleum Engineers
- 5.8.5 Oil Sand, Oil Shale, Bitumen, 1.2.3 Rock properties, 5.4.6 Thermal Methods, 5.5 Reservoir Simulation, 5.4.2 Gas Injection Methods, 4.1.1 Process Simulation, 5.2.1 Phase Behavior and PVT Measurements, 6.5.2 Water use, produced water discharge and disposal, 4.3.4 Scale
- 0 in the last 30 days
- 401 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 12.00|
|SPE Non-Member Price:||USD 35.00|
Numerical simulation of in-situ combustion processes poses numerical problems related to the representation of chemical reactions in the burning zone. This paper presents a new approach to the description of the combustion front. The fully implicit thermal compositional model developed to simulate oil recovery by wet or dry forward combustion is formulated to handle three dimensions, three phases, gravity and capillary forces, heat transfer by convection and conduction within the reservoir, and conductive heat loss to adjacent strata. Two options are available. The first includes four components and one fuel combustion reaction. The second allows any number of components and several reactions. The first option emphasizes description of the combustion front with a new formulation that uses a heat-release curve to improve numerical stability and to give accurate temperature distribution with large gridblocks. The paper compares different representations of chemical reactions and includes results of a sensitivity study on the mesh size. Thermal-balance considerations are given for wet or dry combustion to provide calculated values of the peak temperature and steam-plateau characteristics. The thermal-balance results agree well with those predicted by the numerical model.
The main difficulty in the numerical simulation of in-situ combustion processes is the representation of the reaction zone. Most published combustion simulators are based on a complex approach involving a set of reactions and kinetic equations. Three types of chemical reactions are generally considered: oxidation of oil components, formation of a coke deposit by oil-cracking reactions, and combustion of the coke. A minimum of three chemical reactions therefore take place, and more occur if the oil is considered as a mixture of several pseudocomponents undergoing chemical reactions with different kinetic parameters. Such a reaction mechanism, which gives a good description of the phenomena. is associated with a large number of parameters (Arrhenius-type kinetic parameters, stoichiometric coefficients, and reaction enthalpies) that are very difficult to obtain through laboratory experiments. Another shortcoming of this approach is that the thickness of the burning zone is smaller than the size of the gridblocks generally used for simulation on the reservoir scale. As such, a kinetic representation that accounts for average values of concentrations and temperature in gridblocks may lose all pertinence when large gridblock sizes are used. To prevent the extinction of combustion when large blocks are used, Coats introduced an "activation-temperature concept" in Arrhenius terms and obtained a reactant-controlled model like Youngren's. Ito and Chow introduced a pseudokinetic scheme for field-scale simulation. The reaction rate is defined by a formula that accounts for the gridblock temperature and specified values of maximum reaction rates and minimum and maximum reaction temperatures. Another method consists of a simplified description of the reactions in which an algorithm is used to approximate the overall effects of the phenomena with few parameters (fuel availability and air requirement). This second approach is especially suitable for simulations with large gridblocks. This type of model, however, cannot account for any incomplete oxygen utilization or describe problems of extinction or spontaneous ignition.
A general-purpose numerical model, SARIP, was developed to simulate thermal oil recovery processes such as hot-water injection. steam injection in heavy- or light-oil reservoirs, steam and gas injection and dry or wet combustion. It was designed to handle field, pilot, single-well. and laboratory applications. Two options are available. The first. described later, includes four components (water. oil. oxygen, and inert gas) and one fuel combustion reaction. The second allows any number of components and chemical reactions. The numerical formulation is fully implicit with the variable substitution technique.
|File Size||569 KB||Number of Pages||8|