In-Situ Combustion—An Overview

Saskatchewan Research Council

Executive Editor’s Note

The Technical Report session is designated for high-quality, invited technical reports produced by nonprofit Canadian research organizations. Although not necessarily peer-reviewed by JCPT, these reports are meant for updating our readership of what’s new and exciting in the Canadian research and technology development community relevant to the petroleum industry. We believe these are of great interest to our JCPT readers. In this issue, we present a review article by the Saskatchewan Research Council (SRC) updating the status of research and development regarding the in-situ combustion technologies with an emphasis on Canadian opportunities. I hope you will find the article enjoyable and inspirational. Your feedback regarding this session and the technical content of this article are strongly encouraged. Please send your comments to:

Dr. Jian-Yang (JY) Yuan, BSc, MSc, PhD

JCPT Executive Editor


Heavy oil remains a dominant component in Saskatchewan’s surging oil production. The SRC is one of a few technology institutions that work alongside industry to address the technical challenges of its continued production. The most immediate challenge is the relatively low proportion of original heavy oil in place that can currently be produced.

For more than 30 years, SRC has been leading and participating in a large number of enhanced oil recovery (EOR) initiatives. One of the types of EOR processes that it has been helping to develop is in-situ combustion. When it was first tested, this technology fell short of expectations, but with improved understanding and new oilfield techniques, it is again being viewed as a viable EOR process. A general description of this process and how it has changed is given here.

This Technical Report is not peer-reviewed, but highlights a topic that may be of interest to JCPT readers.

The Basics

In-situ combustion is normally considered to be a thermal method of oil recovery, but sometimes it is regarded as a form of gas drive or a displacement process. In practice, it can be all of these.

Implementation of in-situ combustion is technically complex but conceptually simple. As shown in Fig. 1, air is compressed and injected into an oil-bearing formation, and if some of the oil can be burned, the heat and combustion gases displace the rest of the oil toward a production well. The complexity arises because of the large number of mechanisms that contribute to oil displacement: gas drive, viscosity reduction by both solvent effects and heating, steam generation, distillation of light oil fractions, and pyrolysis of heavy-oil fractions.

The critical question in any given reservoir is whether some of the oil can be burned stably. Stable combustion can occur in two methods. If the formation temperature is high enough naturally and the oil contains enough of the components that oxidize easily, then combustion appears spontaneously. This occurs mainly in deep light oil formations, and this form of combustion is often referred to simply as “air injection.”

The term “in-situ combustion” has come to represent the use of air injection with heavier oils, and it is mostly on this topic that the rest of this article is directed.

In heavy oils, combustion will usually not occur spontaneously. Therefore, the oil formation in the vicinity of the injection well must be preheated artificially, usually to approximately 250°C, before ignition can occur. Such temperatures can be achieved with the injection of high-pressure steam, but when it is not feasible to inject steam, a downhole natural gas burner is usually used.

During in-situ combustion of heavy oils, most of the fuel is coke. This coke is formed by pyrolysis of nonvolatile oil components that remain in the residual oil after the movable oil is driven away by the combustion gases and in-situ-generated steam, and after the residual oil concentration is further reduced by evaporation of its more volatile components. The initial concentration of volatiles in the residual oil can be reduced through low-temperature oxidation, which can thereby increase the coke yield.

Early Failures

Coke burns much less quickly than the oil components that are oxidized during light-oil air injection, and therefore it needs to be kept at a fairly high temperature, usually at approximately 400°C or more. Normally, the combustion front needs only a sufficient flux of injected air to be stable, but if the air supply is interrupted or diminished for a prolonged period, it is difficult to re-establish the front. Therefore, it is essential that a high-temperature combustion front be maintained. Not realizing the importance of maintaining the combustion front was the bane of many Canadian in-situ combustion pilots and projects before 1990, when this difficulty was first identified (Tzanco et al. 1991).

A poor combustion front is accompanied by significant oxygen concentrations appearing in the formation and at production wells. Produced oxygen leads not only to the obvious danger of fire or explosion at production facilities, but also to the creation of polar and acidic compounds in the oil, which can cause serious emulsion-treatment and corrosion problems. In contrast, when a combustion front is strong and stable, it does not allow significant amounts of oxygen to pass through, and almost all of these problems are alleviated. The production of large amounts of carbon dioxide can still cause corrosion, and may necessitate the use of downhole corrosion inhibitors, but with healthy rates of oil production, the naturally oil-wetting nature of metal production tubing and casing will generally alleviate this problem.

The Main Advantage

A strength of in-situ combustion is its ability to provide high oil recovery. Where a stable combustion front has passed, only clean formation rock remains. Therefore, any oil that is not burned has been displaced and is available for recovery. Coke/fuel consumption often ranges between 15 and 30 kg per cubic metre of formation. In high-porosity formations, this translates into less than 15%, or sometimes as low as 6%, of the initial oil. In other words, wherever the combustion front passes, more than 85% of the initial oil is displaced and made available for recovery.

In-Situ Combustion Process Variations

The challenge is therefore to ensure that the combustion front sweeps as much of the reservoir as possible, with as low a cost for air injection as possible. When an oil formation is at least slightly inclined, a linedrive from the high end probably provides the best sweep (Panait-Patică et al. 2006). In addition, the injection of water together with the air (a process variation called “wet” combustion) has been shown in field trials to also improve performance (Joseph and Pusch 1980). The injected water scavenges heat from the hot, burned-out formation and delivers it downstream of the combustion front where it enlarges the steam zone.

Both the oil and the combustion gases must be able to flow to production or vent wells fairly easily, and therefore formations with both high permeability and lower oil viscosity are more favourable. Naturally, this can pose a problem in heavier oils with their high viscosities because the oil bank that forms downstream of the combustion front can form a barrier to the combusted gases and may choke off further air injection. If the initial formation temperature is low, then preheating with steam can provide an ideal scenario for later introduction of air injection. Alternatively, the effort to overcome mobility restrictions in more-viscous oils has led to the development of various process variations like THAI™ (Greaves et al. 2008) and COSH (Kisman and Lau 1994), in which well trajectories are arranged to provide simple flow paths for produced fluids.

This same tendency for the combustion front to create a leading oil bank is also a major contributor to success. The oil bank spreads injected air over a large area and greatly improves the sweep; air does not channel nor override with the same ease as would occur if only flue gas were injected.

Operational Controls

Once the well placement and perforation intervals for any particular field have been decided, the main variables that can be controlled are the air-injection rate and, if water is injected, and the water/air ratio. The air-injection rate must be sufficient to sustain combustion, but if sweep is not affected adversely, the rate will be increased normally to its maximum, which is usually constrained by the permissible injection pressure. In theory, the optimal water/air ratio will be the one for which the water nearly reaches the combustion front. If, however, water overruns the front, this leads to combustion-front extinction and disastrous consequences; it is therefore safer to use a more conservative ratio.

In principle, the production wells could also be choked back in an attempt to control the movement of the combustion front. Because increasing the backpressure at the producers tends to lead to lower oil-production rates, and because it is hard to make a compelling case in advance for restricting production rates, this operational control is seldom used.


If one assumes that in-situ combustion will continue to operate stably at the same conditions that were observed to work in a laboratory, then it is relatively simple to prepare a reservoir simulator that will duplicate this performance in a field setting. Such simulations can be used almost routinely for assessing the economic potential for a project. However, genuine prediction of in-situ combustion field performance is difficult. As an example, the limits to well spacing are neither obvious nor predictable, unless determined from previous experience in similar fields. There are many reasons for regarding simulated predictions with caution, but there are two that stand out. First, the combustion fronts that exist during in-situ combustion are too sharp to be described with the size of gridblocks used normally in field simulations. Consequently, the reaction chemistry cannot be used directly in simulations. Coarse gridblocks cannot honour the normal sequence of coking and combustion. However, the advent of dynamic gridding in commercial simulators within the past several years has helped to alleviate this difficulty.

The second reason for cautious regard of field-simulation results is that there are still serious gaps in the reaction models that are available for in-situ combustion. Of the three main categories of reactions that are acknowledged to occur generally (pyrolysis/coking, combustion, and low-temperature oxidation), only two can be modelled reasonably well, and even then only under restricted circumstances. The third category, low-temperature oxidation, although formally included in the most successful of the published reaction models (e.g., Belgrave et al. 1993), has not yet been described well enough for general prediction purposes (Freitag and Verkoczy 2005). Even so, research on simulation techniques and reaction models is continuing, and significant improvements may well appear in the next few years.

Laboratory Screening Tests

The proven standard laboratory test for in-situ combustion is the combustion tube test, in which a reconstituted core is burned under carefully monitored conditions. Experience has shown that the combustion behaviour observed in a combustion tube—whether good or poor—generally corresponds to the same behaviour in the field. Fuel consumption and air requirements generally correspond well, too. The parameters from combustion tube tests can therefore be used for screening and economic evaluation of field projects.

Most other specialized laboratory tests have been developed to provide some measure of the dominant chemical reactions and kinetics. The accelerating rate calorimeter is used to identify the lowest temperature at which oxidation reactions begin to release heat at detectable rates. Ramped temperature oxidation tests give an indication of the rates of reactions that occur at different temperature. Thermal gravimetric analysis and differential scanning calorimetry can provide similar information. However, the previously described tests are often run in the faster, nonisothermal mode, which makes it more difficult to identify temperature-dependent effects or to separate the influence of competing reactions. Running in an isothermal mode, or using isothermal reactors, tends to provide more accurate data, but at the cost of increased time and expense.

Synopsis of Current Projects

Only a modest number of active in-situ combustion projects currently exist in North America, and many of these have not been reported in the open literature. However, the results of several successful air-injection projects in the light oil formations of the Williston basin have been published. Overall, the process is being adopted less tentatively in other parts of the world. The long-running project at Suplacu de Barcau in Romania and the more recent success at the Balol field in India are solid evidence that in-situ combustion is a healthy and viable oil-recovery technique. More recently, this method has also been implemented at various locations in South America, but information on the performance of these projects is still scant.


The discovery that air injection could be successfully applied in deep, low-permeability, light-oil reservoirs was an early departure from the conventional concept about in-situ combustion. Since then, it has become increasingly apparent that in-situ combustion is highly adaptable, and although it cannot be applied in all reservoir conditions, an awareness of its potential in unexpected scenarios and in combination with other enhanced oil-recovery methods has been emerging. One such example is the EnCAID process by Cenovus Energy in 2014, in which residual oil in a gas zone is burned to produce flue gas that displaces the natural gas.

Various strategies have been proposed to include in-situ combustion in steam-assisted gravity drainage or other steam projects, because the preheated conditions offer an ideal environment for maintaining a combustion front. The prize in these situations is that combustion could provide the necessary heat to maintain these projects at substantially lower cost and with significantly less greenhouse gas production. Whether similar benefits can be obtained in combination with solvent vapour extraction (VAPEX) processes similar to VAPEX or in wormholed reservoirs (Miller et al. 2002) is yet to be tested. In general, if the complexities of these process variations can be handled, the reward can be large.


Credit is given to Muhammad Imran and Norm Freitag for their work on this, ably supported by Kelly Knorr.


Belgrave, J.D.M., Moore, R.G., Ursenbach, M.G. et al. 1993. A Comprehensive Approach to in-Situ Combustion Modeling. SPE Advanced Technology Series 1 (1):98–107. SPE-20250-PA.

Freitag, N.P. and Verkoczy, B. 2005. Low-Temperature Oxidation of Oils in Terms of Sara Fractions: Why Simple Reaction Models Don't Work. J Can Pet Technol 44 (03):54–61. PETSOC-05-03-05.

Greaves, M., Xia, T.X., Turta, A.T. 2008. Stability of THAI ™ Process-Theoretical And Experimental Observations. J Can Pet Technol 47 (09):65–73. PETSOC-08-09-65.

Joseph, C. and Pusch, W.H. 1980. A Field Comparison of Wet and Dry Combustion. J Pet Technol 32 (09):1,523-1,528. SPE-7992-PA.

Kisman, K.E., and Lau, E.C. 1994. A New Combustion Process Utilizing Horizontal Wells And Gravity Drainage. J Can Pet Technol 33 (03):39–45. PETSOC-94-03-05.

Miller, K.A., Moore, R.G., Ursenbach, M.G. et al. 2002. Proposed Air Injection Recovery of Cold-Produced Heavy Oil Reservoirs. J Can Pet Technol 41 (3):40–49. PETSOC-02-03-03.

Panait-Patică, A., Şerban, D., Ilie, N. et al. 2006. Suplacu De Barcau Field—a Case History of a Successful in-Situ Combustion Exploitation. Presented at the SPE-100346-MS, Vienna, Austria, 12–15 June. SPE-100346-MS.

Tzanco, E.T., Moore, R.G., Belgrave, J.D.M. et al. 1991. Laboratory Combustion Behaviour of Countess B Light Oil. J Can Pet Technol 30 (5):43–51. PETSOC-91-05-03.

 © Saskatchewan Research Council

Fig. 1—Generic in-situ combustion.