Investigation of Different Reference Fluids for Viscosity Prediction of Raw Bitumen and Bitumen Mixture Using Corresponding-States Method

Abstract

Bitumen consists of different alkane and aromatic compounds and has a greater viscosity than conventional oils. Due to its high viscosity, the mobility of bitumen and, thus, oil should be increased significantly using methods such as solvent-based and thermal recovery oil production processes. Extraction techniques, such as steam-assisted gravity drainage (SAGD) and cyclic steam stimulation (CSS), play an important role in recovering unconventional oil, such as bitumen. A thermal and solvent based method such as expanding solvent steam-assisted gravity drainage (ES-SAGD) technique has been developed to reduce the amount of steam consumed in this process. In all of the aforementioned methods, viscosity plays an important role in the design of the processes. Viscosity is a key transport property in well-designed pipeline and process simulations and also in prediction of production rates.

In this research, the focus is on the prediction of viscosity for raw Athabasca bitumen and bitumen mixtures with solvents, such as toluene, xylene and n-decane. In the literature, viscosity corresponding states methods have been used for the prediction of viscosity, with the bitumen characterized into different pseudo-components.

Methane has been used as a reference fluid, which is quite inappropriate for bitumen viscosity calculations. This study investigates different reference fluids, such as alkane, cycloalkane and aromatic hydrocarbons. Aromatic hydrocarbons show a closer trend to bitumen viscosity. Higher molecular weight aromatics, such as 1,1 diphenylheptane, can predict Athabasca bitumen viscosity without much deviation.

The effects of temperature, pressure and solvent concentration on the prediction of viscosity are examined. The model is tested for the prediction of the viscosity of bitumen mixtures with toluene, xylene and n-decane; and, the resulting absolute deviations from the experimental data are 8.40, 8.05 and 8.54%, respectively. The Lindeloff corresponding states model is also investigated with different reference fluids for the prediction of bitumen viscosity; and, again, 1,1 diphenylheptane results in less deviation from the experimental data than the other studied reference fluids.