A Field Scale In-Situ Combustion Simulator With Channeling Considerations
- Yoshiaki Ito (Gulf Canada Resources Ltd.) | Allan Kwok-Yuen Chow (Gulf Canada Resources Ltd.)
- Document ID
- Society of Petroleum Engineers
- SPE Reservoir Engineering
- Publication Date
- May 1988
- Document Type
- Journal Paper
- 419 - 430
- 1988. Society of Petroleum Engineers
- 4.3.4 Scale, 5.4.6 Thermal Methods, 5.6.5 Tracers, 5.3.4 Reduction of Residual Oil Saturation, 5.5.8 History Matching, 2.4.3 Sand/Solids Control
- 0 in the last 30 days
- 267 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 5.00|
|SPE Non-Member Price:||USD 35.00|
Summary. Many thermal simulators have been developed in the past few years, most based on the assumption that all the important mechanisms of the combustion process would remain intact when a laboratory-scale simulation study is scaled up to a field-scale study. In most cases, the simulators provide the users with the capabilities to achieve good history matches of laboratory results when small gridblocks (i.e., 5 cm [2 in.]) are used. when the simulators are used in a field-scale study in which gridblocks are usually greater than 10 m [33 ft], however, many problems have been encountered, including high fuel consumption and misrepresentations of the chemical reactions.
A new and numerically stable algorithm has been developed to achieve the desirable fuel consumption. This algorithm involves an oil-flow-enhancement scheme. Arrhenius-type reactions have been found to be inadequate in field-scale combustion simulation. A new pseudokinetic scheme is introduced to field-scale studies in which the reaction zone is not simulated accurately because of the large size of the gridblocks.
In some in-situ combustion projects, viscous fingering and channeling phenomena can affect the efficiency of the combustion process. A new phenomena can affect the efficiency of the combustion process. A new formula of molecular diffusivity of oxygen has been introduced to represent the contact efficiency of the reactants. A new concept that represents the nonequilibrium thermal condition between the injected fluid and the resident materials (i.e., poor exchange of heat energy) is discussed.
With the rapid advances in computer technology, data processing expenses have become less costly over the years. As a result, many numerical simulators have been developed and used as helpful tools in predicting reservoir performance. In the in-situ combustion process, however, the actual field performances are often different process, however, the actual field performances are often different from those predicted by simulators. The differences are usually caused by the failure of simulators to simulate the combustion process properly. process properly. Many problems are encountered when scaling up from laboratory conditions to the field conditions in a numerical simulation study. One of these problems occurs in the high-temperature zone where most of the chemical reactions take place. In general, the length of this zone is believed to be 10 to 20 cm [4 to 8 in.] in either laboratory or field conditions. Hence it is too small to be represented properly by the large gridblocks used in the field-scale study. properly by the large gridblocks used in the field-scale study. The reaction zone can be simulated in one to two gridblocks in a laboratory-scale study. In a field-scale study, however, the zone would appear in only a fraction of a single gridblock. Therefore, the temperature profile obtained in a laboratory-scale simulation is very different from that in a field-scale study. It is not currently possible to obtain a correct temperature distribution using thermal possible to obtain a correct temperature distribution using thermal simulators with the field-scale gridblocks. Thus, the kinetic and other data obtained from the history matches of laboratory studies should not be used for field-scale studies. Achieving meaningful results under the field-scale condition would require some forms of modifications. These modifications would involve all the temperature-dependent variables, such as oil viscosity, oil volatility, temperature-dependent relative permeability characteristics, and chemical reaction kinetics. However, all these variable changes would add such an undesirable complexity to the computer simulator that practical applicability is essentially destroyed. We believe that the accuracy of the simulators in field-scale combustion simulation can be increased significantly with minimal complexity by (1) controlling fuel consumption and (2) introducing a pseudokinetic scheme. pseudokinetic scheme. The grid-size problem is a very well-known concern and has been addressed by several other authors. If the gridblock used is too large, it may even cause the fire to extinguish. Coats introduced the activation-temperature concept to ensure that the fire would be sustained by an increase in the reaction rate when the block temperature is below the activation temperature. Hwang et al. found that using large gridblocks will result in higher fuel consumption than that observed in the laboratory. They resolved this fuel problem by specifying an amount of oil to be burned per unit bulk problem by specifying an amount of oil to be burned per unit bulk volume of formation. The burning front continuously pumps oil out from the burning zone to the neighboring blocks.
Grid-size problem is only one of the major problems encountered in field-scale simulation. In many in-situ combustion projects, viscous fingering and/or channeling phenomena can seriously affect the efficiency of the combustion process. The presence of these phenomena is indicated by certain characteristics in the injection phenomena is indicated by certain characteristics in the injection and production histories or in the flue-gas analysis. One such characteristic is the earlier-than-expected oxygen breakthrough as a result of poor contact between the reactants.
Although the presence of channels may be detected, the size, shape, and number of the channels are difficult to determine. They may vary with location, injection rate, or the type of injection fluid used. Some indications regarding the changes in size, shape, or number of the channels can be obtained from the flue-gas analysis and the fluid production history. However, a detailed or accurate description of these parameters cannot be obtained. Therefore, the channeling phenomenon will be difficult to duplicate in the laboratory. We believe that this complex phenomenon can be incorporated into the numerical simulator by adding (1) contact efficiency of the reactants (diffusivity of oxygen) and (2) nonequilibrium thermal condition.
Another problem encountered in field-scale simulation is the type of reaction. Because some chemical reactions take place along the channels, the heat energy can be transferred readily from the channels to the surrounding oil zone because of the large contact area. This will result in a lower temperature in the combustion reaction zone. The chemical reaction type will be shifted to those favoring low temperatures, such as dehydrogenation and/or low-temperature oxidation. Several articles have been published indicating that the performances of the combustion process depend on the low-temperature performances of the combustion process depend on the low-temperature oxidation reactions accompanying the in-situ combustion operations. Hence, low-temperature oxidation and dehydrogenation reactions have been included in the modified simulator.
|File Size||1 MB||Number of Pages||12|