Layer | Fill | Outline |
---|
Map layers
Theme | Visible | Selectable | Appearance | Zoom Range (now: 0) |
---|
Fill | Stroke |
---|---|
Collaborating Authors
Results
Summary. This paper presents two kinetic models for representing the thermal cracking of crude oils, which incorporate the cracking rate parameters and stoichiometric coefficients to correlate experimental data. parameters and stoichiometric coefficients to correlate experimental data. The models presented show that the first-order kinetics generally accepted for pure components are unsatisfactory for multicomponent systems characterized by pseudocomponents. We conclude that three corrections to the existing first-order model are needed for modeling thermal cracking of mixtures. First, the apparent reaction order is always greater than one. Second, the reaction order is a decreasing function of temperature. Third, coke may also be formed from intermediate products. These corrections are incorporated into the models. In the first model, crude oil is split into two pseudocomponents, while in the second model, crude oil is represented by three pseudocomponents. The models can be easily extended to any number of pseudocomponents. The models can be easily extended to any number of pseudocomponents. pseudocomponents. The models successfully correlated experimental data of four systems available in the literature. Furthermore, it was confirmed that coke is not always the same source of the fuel burned in an in-situ combustion process. process. Introduction It is generally believed that in in-situ combustion processes, the combustion zone is preceded by a cracking or processes, the combustion zone is preceded by a cracking or superheated steam zone where coke is formed from the thermal cracking (pyrolysis) of crude oil. The kinetics of the cracking reaction may be a crucial process mechanism affecting the performance of combustion processes because it not only produces solid-like coke for combustion but also upgrades the remaining oil, which affects the vaporization behavior. As a result, the cracking reaction will strongly influence the total amount of fuel available in the combustion zone. The effects of cracking reactions on the fuel deposition mechanism and the fuel composition have been discussed elsewhere. The reaction mechanisms of hydrocarbon cracking are very complex. Even for a pure component, it is almost impossible to describe the mechanism precisely. Nevertheless, it is possible to use simple global rate expressions to represent the reaction rate of the reactant. For pure hydrocarbons, it is well established that the cracking reaction can be properly modeled by a first-order rate expression, although self-inhibition (decreasing first-order rate constant with increasing conversion) was generally observed. This is caused by the formation of olefins, which are known to be good inhibitors of free-radical reactions. As a general rule, the reaction rate constant for normal paraffins increases with increasing carbon number. while the activation energy decreases with increasing carbon number. Global rate expressions were also applied to the pyrolysis of gas oil and crude oils by use of first-order pyrolysis of gas oil and crude oils by use of first-order kinetics. Most of these studies lumped the multicomponent mixture into one oil component, while McNab et al. assumed that only the heavy-oil fraction contributes to the cracking reaction and arbitrarily chose 80% of the residue of the distillation of the original crude as the heavy-oil component. This two-pseudocomponent approach was adopted by Henderson and Weber. In all the above studies, only the reaction rates of the crude oil components were considered, and no attempt was made to correlate the stoichiometry of the reaction. In an attempt to match the product distribution of Athabasca bitumen pyrolysis, Hayashitani et al. constructed a number of complex kinetic models. The oil was divided into three to five components, and six to eight first-order reactions were included in each model. They found that these complex models cannot satisfactorily correlate all aspects of experimental behavior of cracking reactions. They concluded that during the course of cracking reactions, each pseudocomponent might have changed its characteristics, A comprehensive reaction scheme for catalytic cracking of gas oils was recently developed by Jacob et al., who simulated the gas-oil cracking using four pseudocomponents (heavy oil, light oil, gasoline, and gas plus coke). pseudocomponents (heavy oil, light oil, gasoline, and gas plus coke). For each of the light- and the heavy-oil components, the pseudocomponent was further split into paraffinic, pseudocomponent was further split into paraffinic, naphthenic, aromatic, and aromatic substitute groups. As a result, the reaction scheme involves 10 species (components) and 20 first-order reactions. The Arrhenius constants and the activation energies for these reactions were assumed to be universal constants (i.e., independent of gas-oil composition), This model was claimed to work well for a wide spectrum of gas oils. However, the applicability of this model to heavy oils with a wider range of composition remains to be proved. P. 54
- Energy > Oil & Gas > Upstream (1.00)
- Materials > Chemicals > Commodity Chemicals > Petrochemicals (0.88)
Abstract This paper presents the results of numerical simulation of dry, forward combustion tube experiments. The kinetic aspects of in-situ combustion processes also are discussed. The goals of the study are to investigate processes also are discussed. The goals of the study are to investigate the fuel deposition mechanism and to identify the key parameters affecting the performance of in-situ combustion processes. The thermal simulator developed at Gulf R and D Co. was used in the study. It was modified to include the capillary outlet effects for a more realistic description of the oil and water productions. The following experimental data were matched: cumulative water and oil productions, position of the combustion front as a function of time, fuel consumption, position of the combustion front as a function of time, fuel consumption, temperature as a function of time and position, and the pressure drop across the tube. History matches were performed for two crude oils with distinctly different physical properties (gravities of 26.5 and 13 API [0. 896 and 0. 979 g/cm3]). The agreements between experimental data and simulation results were excellent. Results indicate that the component equilibrium K-values and the kinetics of cracking reactions are the most important parameters affecting the fuel deposition, and that the fuel deposition mechanism, the fuel composition, and the locations and sizes of the transient zones depend on the crude oil and reservoir rock properties. Simulation results are always sensitive to the K-values of the light oil component but insensitive to the K-values of the heavy oil component. Results are sensitive to the kinetics of cracking reaction only if the cracking reaction is catalytic or the peak temperature and the fuel consumption are sufficiently high. Furthermore, the fuel available may or may not be solely in the form of coke. Our study suggests that further investigations of the catalytic effect of reservoir rocks and reaction kinetics of the cracking reaction are needed. Also, more than two crude oil components may be required to simulate the evaporation effect of crude oil accurately. Introduction In in-situ combustion processes, many physical changes as well as chemical reactions take place simultaneously or sequentially in the vicinity of the combustion front. It is generally believed that the combustion zone is preceded by a cracking or superheated steam zone, where coke is formed and preceded by a cracking or superheated steam zone, where coke is formed and deposited on the sand grains, and some lighter crude oil components evaporate and move forward with the flowing gas phase. The kinetics of combustion and cracking reactions in the combustion zone and the cracking zone has been discussed widely in the literature. The mechanisms of the physical changes and chemical reactions occurring around the combustion zone can be studied effectively through numerical simulation by using a thermal simulator. Although a number of numerical simulations of combustion tube experiments have been performed with different thermal simulators, no conclusions regarding the mechanism of fuel deposition can be drawn from these studies. The mentioned simulations either neglect the formation of coke from cracking reaction or use a high cracking rate so that no residual oil will be present in the combustion zone. The mechanism of fuel deposition is controlled by two important processes: the evaporation of crude oil components and the kinetics of the processes: the evaporation of crude oil components and the kinetics of the cracking reaction. These two processes determine how much fuel eventually will be burned and how much fuel will be in the form of coke. It has been reported, that low-temperature oxidation can have a significant effect on the fuel deposition and fuel characteristics. However, this reaction is important only when oxygen is available downstream of the combustion front. If oxygen is used completely in a combustion tube experiment, low-temperature oxidation will not play an important role in the fuel deposition mechanism. For a system with a high cracking reaction rate, it is likely that all of the crude oil in the cracking zone will be either evaporated or coked so that coke is the sole source of fuel. However, if the cracking rate is so low that only a portion of crude oil in the cracking zone is evaporated or coked, then some residual crude oil also will be burned in the combustion zone. This is supported strongly by the experimental data of Hildebrand who conducted a number of combustion tube experiments using clean, crushed Berea sandpacks with a variety of crude oils. SPEJ p. 657
Abstract This paper describes a model for numerically simulating thermal recovery processes. The primary locus is on the simulation of in-situ combustion, but the formulation also represents fire-and-water flooding, steamflooding, hot water flooding, steam stimulation, and spontaneous ignition as well. The simulator describes the flow of water, oil, and gas, and includes gravity and capillary effects. Heat transfer by conduction, convection, and vaporization-condensation of both water and hydrocarbons are included. The rigorous but general nature of the simulator is obtained by employing conservation balance equations for oxygen, inert gases, a light hydrocarbon pseudocomponent, a heavy hydrocarbon pseudocomponent, water, coke, and energy. pseudocomponent, water, coke, and energy. Vaporization-condensation is governed by vaporliquid equilibrium using temperature and pressure-dependent equilibrium coefficients. Four pressure-dependent equilibrium coefficients. Four chemical reactions are accounted for: formation of coke from the heavy hydrocarbon component and the oxidation of coke and both heavy and light hydrocarbon components. Formulation details, numerical solution procedures, and computational results are presented. procedures, and computational results are presented. The computational results include both one- and two-dimensional cross-sectional studies. The simulator represents a major improvement in the ability to simulate thermal recovery processes under complex conditions. Introduction Considerable progress has been made in numerically simulating thermally enhanced oil-recovery processes during the last few years. This is particularly true for-processes involving steam, where we have seen a continual improvement of our ability to treat the problem. The most recent contributions provide an analysis capability for steam displacement and steam stimulation recovery methods, accounting for all the important physical mechanisms of these processes. Progress in simulating the performance of in-situ combustion processes is not so advanced. Initial simulation attempts were concerned primarily with the heat-transfer aspects of combustion. The most sophisticated heat-transfer model was developed by Chu. His numerical model considers the energy effects of vaporization and condensation on the temperature distribution, but neglects the accompanying phase changes by assuming constant fluid saturations. More recent heat transfer or heat-wave models for the in-situ combustion process were proposed by Kuo in 1969 and by Smith and Farouq-Ali in 1971. Kuo's model allows two temperature fronts-one at the combustion zone and one at a heat front. The heat-front position is predicted by gas flow that is allowed to have a velocity different from the velocity of the combustion front. The simulator proposed by Smith and Farouq-Ali is designed for proposed by Smith and Farouq-Ali is designed for predicting sweep efficiencies in confined well predicting sweep efficiencies in confined well patterns. Their numerical model accounts for heat patterns. Their numerical model accounts for heat generation by a combustion zone (assuming fixed fuel content all through the reservoir), heat transfer by conduction and convection (single-phase gas flow) in the reservoir, heat losses by conduction to adjacent formations, and different permeability-to-gas (air) flow on either side of the combustion zone. Special cases of the in-situ combustion process were studied by Gottfried and Khelil. These authors examine the heat transfer and oxygen use in reservoirs composed of an oil-bearing layer and an overlying "clean" porous zone containing only gas. These models were designed primarily to investigate the various transport mechanisms present when combustion is initiated in a reservoir present when combustion is initiated in a reservoir containing a gas cap. Because of the many assumptions invoked and the specialized geometry to which they apply, they do not satisfy the need for a general purpose simulator. SPEJ P. 37