(32b) Measurement and Modeling of the Viscosity and Stability of Visbroken Fractionated Bitumen

In Alberta, bitumen production is restricted by the limited pipeline capacity to external markets. The capacity is further limited because the bitumen must be diluted with a solvent to meet the pipelines' viscosity and density specifications [1]. Solvents occupy 25% to 50% of the bitumen transportation pipelines and require a second pipeline to return the solvents [2]. One way to increase the capacity is to use a combination of partial deasphalting and visbreaking [3,4] to reduce the density and viscosity of the bitumen so that less diluent is required. Partial deasphalting uses a poor solvent to precipitate and remove some of the asphaltenes from the bitumen. The asphaltenes are the densest fraction of the bitumen; hence deasphalting reduces the product density but only moderately reduces its viscosity. Vibreaking is a mild liquid-phase thermal cracking method that only moderately reduces the density the bitumen but significantly reduces its viscosity [5,6]. The viscosity reduction that can be achieved in a visbreaking process depends on the intensity (residence time and temperature) of the process. There is an optimum intensity below which there is less viscosity reduction and above which coke forms. Coke formation is undesirable because it leads to fouling. The visbroken product must also be stable versus asphaltene precipitation when blended with a diluent. Therefore, to optimize a visbreaking process it is necessary to predict the viscosity and stability (versus asphaltene precipitation) of the product.

The eventual goal of this project is to examine combinations of visbreaking and deasphalting. As a first step, fractionated oils are examined and the specific objective of this study is to measure and model the viscosity and stability of the products from three types of visbreaking process feeds: a whole oil, a vacuum residue, and a partially deasphalted oil. The feeds are all from or derived from Western Canadian bitumen. A methodology designed for integration with process simulation software and previously applied only to whole bitumen was followed [6,7]. Each of the feeds were reacted in an in-house visbreaker at two different conversions. The composition of the products were measured in terms of distillates and the SARA fractions of the distillation residue. The molecular weight, density, and viscosity of each fraction from the feeds were measured. The density, viscosity, and asphaltene solubility of the products and there maltene and residue fractions were also measured. The Expanded Fluid viscosity model was used to model the viscosity of the products. The Modified Regular Solution model was used to model stability of the products.

The Expanded Fluid model relates viscosity to density using three fluid specific parameters. The model includes mixing rules to determine whole oil parameters from the oil composition and component parameters. In a previous study [6], recommended parameters were proposed for the distillates and SARA fractions from whole bitumen. The density of each fraction (a required input) was related to temperature using two parameters and recommend values were provided. In addition, correlations were developed to predict the viscosity and density parameters as a function of conversion for fractions from visbroken oils. Preliminary results indicate that the recommended parameters and correlations also apply to vacuum bottoms and deasphalted oils except for the asphaltenes from deasphalted oils. Notable property trends, the differences from the previous properties, and the proposed modifications to the correlations will be discussed.

The Modified Regular Solution model is an activity-coefficient-based liquid-liquid phase equilibrium model which requires the molecular weight, density, and solubility parameter of each component in the mixture. In a previous study [7], recommended properties were proposed for the distillates and SARA fractions from whole bitumen. In addition, correlations were developed to predict these properties as a function of conversion for fractions from visbroken oils. Preliminary results indicate that the recommended properties and correlations also apply to vacuum bottoms and deasphalted oils except for the asphaltenes from deasphalted oils. Notable property trends, the differences from the previous properties, and the proposed modifications to the correlations will be discussed.

[1] R. W. Luhning, A. Anand, T. Blackmore, and D. S. Lawson, "Pipeline transportation of emerging partially upgraded bitumen," in Canadian International Petroleum Conference 2002, CIPC 2002, 2002.

[2] A. Hart, "A review of technologies for transporting heavy crude oil and bitumen via pipelines," Journal of Petroleum Exploration and Production Technology. 2014.

[3] M. R. Gray, Upgrading Oilsands Bitumen and Heavy Oil. The University of Alberta Press, 2015.

[4] A. Zachariah and A. De Klerk, "Partial Upgrading of Bitumen: Impact of Solvent Deasphalting and Visbreaking Sequence," Energy and Fuels, 2017.

[5] J. B. Joshi et al., "Petroleum residue upgradation via visbreaking: A review," Industrial and Engineering Chemistry Research. 2008.

[6] A. Marquez, F. F. Schoeggl, S. D. Taylor, G. Hay, and H. W. Yarranton, "Viscosity of characterized visbroken heavy oils," Fuel, 2020.

[7] S. Rodriguez, E. N. Baydak, F. F. Schoeggl, S. D. Taylor, G. Hay, and H. W. Yarranton, "Regular solution based approach to modeling asphaltene precipitation from native and reacted oils: Part 3, visbroken oils," Fuel, 2019.