(66b) Viscosity Modeling of Visbroken Heavy Oils

Heavy oil viscosity exceeds 1000 mPa.s and bitumen viscosity can be as high as 1 million mPa.s at atmospheric conditions. The viscosity of these fluids must be reduced prior to transport and processing and one option to do so is visbreaking. Visbreaking is a mild liquid-phase thermal cracking process where large hydrocarbon molecules react in order to produce shorter chain molecules [1]. Visbreaking is optimized by modifying the process severity (residence time and temperature) to maximize the production of distillates and viscosity reduction in the residue, while avoiding coke formation [1]. Hence, to design and optimize a visbreaking process, it is necessary to predict the composition and viscosity of the reacted fluid. The overall objective of this study is to develop a method to predict the viscosity of the visbroken crude oil from the feed crude oil properties and reaction conditions. The method is to be structured for implementation in process simulators.

It has been shown that the viscosity of visbroken products cannot be predicted from the properties of the whole product oil [2] but there is potential to predict the viscosity from the changes in the distribution and properties of the oil’s pseudo-components. The proposed methodology is to characterize the oil into a set of pseudo-components that represent the property distributions within the fluid. Each pseudo-component is assigned the set of properties necessary to model the reactions and calculate viscosity. These properties, including molecular weight, density, and boiling point, are determined from a distillation assay, a SARA (saturate, aromatic, resins, asphaltene) assay, and established correlations. The viscosity is to be determined with the Expanded Fluid (EF) viscosity model [3]. The model inputs for the feed can be determined from the molecular weight, specific gravity, and boiling point. To complete the model, a method is required to predict the mass fraction and model parameters of the distillable and SARA pseudo-components of the visbroken products. The specific objective of this contribution is to develop correlations for the model parameters of the visbroken product.

A Western Canadian bitumen sample was visbroken at five different residence times and temperatures in an in-house laboratory pilot plant. In addition, a feedstock and two thermally cracked samples were provided by a Chinese industrial source. The pilot plant is a continuous flow coil visbreaking unit. The feed oil is pumped through a pre-heater at a temperature of 250°C and then to a reactor where it reaches the desired visbreaking temperature. The reactor is held at sufficient pressure to maintain a single liquid phase in the reaction zone. The product from the reactor enters an atmospheric flash separator. The separated gas is vented while the liquid product is collected for further analysis.

The feed and visbroken product were fractionated using distillation and SARA assays. The conversion was determined based on the change in the mass fraction of the 524+°C cut in the product versus the feed. The viscosity, density, and molecular weight of these fractions were measured. The dataset from the Western Canadian bitumen and its visbroken products was used to develop correlations for the Expanded Fluid parameters for the distillables and each SARA fraction based on reaction conditions. The viscosity of the visbroken products was then predicted from the product oil characterization and the correlated parameters. The methodology was then used to predict the viscosity of the visbroken Chinese oil samples.

References:

[1] Joshi, J; Pandit, A; Kataria, K; Kulkarni, R; Sawarkar, A; Tandon, D; Ram, Y; Kumar, M. Petroleum residue upgradation via visbreaking: A review. Ind. Eng. Chem. 2008, 47, 8960-8988.

[2] Rueda-Velazquez, R; Gray, M. A viscosity-conversion model for thermal cracking of heavy oils. Fuel. 2017, 197, 82-90.

[3] Ramos-Pallares, F; Taylor, S; Satyro, M; Marriot, R; Yarranton, H. Prediction of viscosity for characterized oils and their fractions using the expanded fluid model. Energy. Fuels. 2016, 30, 7134-7157.