Kinetics of In Situ Combustion Reactions: Laboratory Experiments and Numerical Analysis

In-Situ Combustion (ISC) is an Enhanced Oil Recovery (EOR) method for heavy oil production with many open research questions. The objective of this work is to characterize the chemical kinetics properties of crude oil to predict ignition and the combustion reaction path following a successful ignition. It is important and necessary to predict ignition because it is the first step of ISC. Without successful ignition, ISC will not occur. The significance of reaction path prediction is to extrapolate the use of kinetics for further predictions, such as upscaling the simulation from kinetics cell scale to combustion tube scale, or even field scale.

Ramped Temperature Oxidation (RTO) experiments are conducted to measure crude oil oxidation kinetics. The selected heating rates for RTO experiments are 1.5, 2.0, 2.5, 3.0, 5.0, 10.0, 15.0, and 20.0 degrees Celsius per minute. The small heating rate cases capture the details of reaction, and large heating rate cases spread the range of heating rates for further analysis. Isoconversional analysis is the initial step to interpret crude oil kinetics, and history matching of effluent gas concentrations calibrates a reaction model to represent a particular type of crude oil.

A system automation algorithm is developed to interact with commercial software for iterative computations. This is applied to the history matching and ignition prediction workflows. The calibrated reaction model from history matching is therefore implemented in numerical models to predict ignition under various operational conditions (e.g. heat loss rate, air flow rate, etc.). There are four different ignition criteria implemented and compared in this work. The impacts of the ignition criteria on the final predicted ignition envelopes are investigated and compared. The uncertainty quantification results provide a range of confidence of predicted ignition regime for reference to field operations.

With successful ignition, the reaction conversion path is predicted through two approaches. The reactivity mapping method takes experimental measurements directly to predict conversion, with an arbitrary temperature schedule provided. In comparison, a history matching method predicts reaction path indirectly. The experimental data is used to calibrate the reaction model, and then the simulator predicts reaction under an arbitrary temperature design. The performance of predictions through these two approaches are compared and discussed.