The low-temperature oxidation (LTO) reactions of the SARA fractions separated from two crude oils were studied in the presence of their reservoir sands at temperatures between 130 and 230 °?C. The results indicated that the usual approach to modelling LTO-the use of a very few single-step Arrhenius-rate equations-could not be made to reflect the observed reaction kinetics. Instead, this investigation found that the following reaction characteristics were needed for accurate reaction modelling:a change in the order of reaction with respect to oxygen concentration from ? to 1 as temperature rises;the repression of a saturates oxidation reaction by other fractions; and,a prominent induction period exhibited by the saturates fraction. The compositions and yields of the ultimate LTO reaction products were measured, and these included relatively stable residues with high oxygen contents. Because the LTO reactions play an important role in enhanced oil recovery by air injection methods, the above information is valuable for the simulation and prediction of these processes. Introduction Enhanced oil recovery processes need to be predictable before they can be seriously considered for widespread field application. One of the main problems limiting the development and application of new process variations for air injection or in situ combustion is that their field performance and consequently their technical success or failure can simply not be predicted with any reliability. The most serious questions frequently hinge upon the nature of stability of the combustion/oxidation zones. Many studies have provided valuable knowledge as to the nature of the related chemical reactions, but the usefulness of proposed reaction models for numerical simulation prediction is still limited. Three main types of reaction have been found to govern air-injection EOR processes: pyrolysis/coking, low-temperature oxidation, and high-temperature oxidation (combustion). This study was concentrated on the second category of reaction: low-temperature oxidation (LTO). The free-radical nature of low-temperature oxidation of hydrocarbons has long been known. In a 1958 review, Morton and Bell(1) confirmed that LTO occurs through a free-radical mechanism in which the production of hydroperoxides is an important first step. They also mentioned the role of inhibitors, discussed catalysis by metal surfaces and metallic salts, and described why long induction periods could occur before the onset of significant oxygen consumption was observed. Initially, most studies of the chemical mechanisms conducted both before and after Morton and Bell's review used pure compounds. The results varied between compounds, and could not be used directly to describe the oxidation of complex mixtures like crude oils in a petroleum reservoir. In 1968, Bousaid and Ramey(2), while investigating high-temperature oxidations, carried out three low-temperature oxidation tests on a heavy oil between 23 and 52 ° C. They reported very low values of oxygen consumption, with rates that were correlated with an activation energy of 53,200 J/ gmol. Later, Dabbous and Fulton(3) published much more extensive results for LTO of two whole oils on crushed Berea sand over the temperature range of 121 to 246 ° C.

A thermogravimetric analyzer (TGA) was used to obtain information on the pyrolysis and combustion behavior of both crude oil (Neilburg) and its asphaltenes, each mixed with reservoir sand. Of all the saturate, aromatic, resin, and asphaltene fractions, asphaltenes contribute the most to the formation of coke (fuel). Temperatureramped as well as the isothermal pyrolysis experiments on whole oil and asphaltenes were analyzed to determine the temperature at which coke formation was maximized. Furthermore, isothermal combustion curves for coke derived from whole oil and asphaltenes were obtained to provide reliable data for calculating the kinetics of the reactions. The classical Arrhenius model was applied, and the activation energy for the combustion of coke formed from pure asphaltenes and from the whole oil was calculated. The results showed that the Arrhenius model fitted the data well in the entire range of temperatures the experiments were conducted. The source material for the coke led to modest differences in its reactivity. The observed activation energy for asphaltenes was 117.7 kJ/mol, and for the whole oil it was 129.5 kJ/mol, which indicates that they were in close agreement. Also, the combustion of coke from asphaltenes showed a reaction order of 0.4 at 375°C, which gradually increased to 0.9 at 525°C. For whole oil, it increased from 0.5 at 375°C to 0.7 at 500°C.

Summary The lack of an accurate reaction model for petroleum-oxidation rates is a serious hindrance to the simulation of oil-recovery processes that involve air injection. However, the chemical literature on hydrocarbon oxidation contains many examples of possible reaction mechanisms that could serve as guides. These mechanisms were screened to identify generally accepted reaction paths that could help reveal how oxidation occurs in petroleum reservoirs. It was found that there are at least eight groups of fundamental reactions that can seriously affect oxidation rates of crude oils or their pyrolysis products. These eight reactions are as follows: two that lead to hydroperoxide formation; “branching” by hydroperoxides; two reactions governing the negative temperature coefficient (NTC) region; oxidation inhibition; at least one rate-controlling reaction at very high temperatures; and the combustion of coke that is produced by pyrolysis. Each of these groups exerts an influence within a separate, identifiable range of conditions. These reactions, and the conditions under which they become important, are outlined in this paper. Various oxidation behaviors that were reported for both light and heavy crude oils were then compared and aligned with the eight identified reactions. The result was a framework for selecting pseudoreactions that can facilitate the prediction of the oxidation kinetics under a wide range of oilfield conditions. Some of these pseudoreactions involve the direct representation of free radicals or other chemical intermediates, which is a departure from conventional practice for in-situ-combustion simulation. The new reaction framework is expected to serve as a reliable guide to the construction of predictive reaction models and, consequently, improved simulation of both in-situ-combustion and high-pressure-air-injection (HPAI) processes.

No abstract