Assessing the Consistency of Measurements and Models for the Kinetics of In Situ Combustion

Abstract

One method to access unconventional, heavy-oil and natural bitumen resources is to apply in situ combustion (ISC) to oxidize in place a small fraction of the hydrocarbon thereby providing heat to reduce oil viscosity and pressure that enhances recovery. ISC is also attractive because it provides the opportunity to upgrade oil in-situ by increasing the API gravity and decreasing, for instance, sulfur content. The complex nature of petroleum as a multi-component mixture and the multi-step character of oxidation reactions complicates substantially the kinetic analysis of crude-oil oxidation. Isoconversional techniques provide model-free methods for estimating activation energy and naturally deconvolve multi-step reactions. The isoconversional fingerprint is, essentially, the profile of activation energy as a function of temperature (from low to high) of the sample. As such, isoconversional methods are also useful as a screening tool to categorize the burning characteristics of different oils. In the literature there are substantially different kinetic models for the representation of reactions during in situ combustion. Some models feature multistep reactions in series whereas other models combine series and parallel reaction steps. Interestingly, there is no study comparing these models with each other as well as consistent experimental data. In this study, the performance of different models proposed in the literature and a new model is tested using kinetic cell simulations and compared to new experimental data. As the basis of the comparison, the isoconversional fingerprints associated with the models existing in the literature are compared. Results indicate that most of the models proposed are unable to reproduce the typical isoconversional fingerprint measured for crude-oil combustion. This lack of agreement explains, in part, the poor predictability of reactive transport simulations of ISC. Importantly, this work illustrates a methodology for assessing the consistency of a kinetic model with experimental data and identifies physically consistent models.