Prediction of Heavy Oil Viscosity in Current Reservoir Simulators

Abstract

Abstract An accurate estimate of oil mobility is essential for reliable exploitation of oil and gas reservoirs, and for development of new technologies, particularly with the industry's increasing reliance on unconventional resources such as heavy oil. The viscosity of heavy oil decreases rapidly with increasing temperature and with increasing concentrations of light components, and can vary over 2-3 orders of magnitude during typical exploration and productions operations. The accurate prediction of the viscosity is therefore difficult, and simple extrapolation of the current methods used in the petroleum industry for predicting the fluid viscosity in conventional reservoirs is fraught with difficulties. This paper presents a comprehensive analysis of the use of a simple mole-average power law that is based on the Arrhenius equation and used as the default method in some widely used thermal reservoir simulators to predict mixture viscosity. Predictions based on the equation are compared to a set of accurate benchmark data that are based on the best available experimental data for hydrocarbon mixtures. These data cover a temperature range of -175°C to 200°C, and extend to high pressures. The accuracy of the data is better than 5%, sufficient for validation purposes. We summarize the conditions under which the simple power law provides reasonable estimates of viscosity, identify the ranges of pressure and temperature and composition for which large deviations can be expected, and discuss the implications of using the equation to predict heavy-oil viscosity. We demonstrate that the current practice of fitting scarce experimental data and using the viscosity of the heavy fraction as an adjustable parameter is problematic, and that extrapolation in temperature or to similar mixtures is highly unreliable.