(38d) Multi-Scale Modeling of Heavy Oil Upgrading in Near-Critical/Supercritical Water

Raghavan, A., Massachusetts Institute of Technology
He, P., Lamar University
Ghoniem, A. F., Massachusetts Institute of Technology
The upgrading of heavy oil using near-critical or supercritical water can be pushed towards better outcomes (higher light liquid product yield and lower coke formation) by suppressing the condensation reactions of light hydrocarbon species to heavier hydrocarbons and coke precursors. A potential method to achieve this is to ensure a much lower residence time for the light hydrocarbon species in the high-temperature reactor zone once they are formed by thermal cracking reactions of the heavy hydrocarbons.

This difference in residence times between the light and heavy species in the oil can in turn be realized in practice in a reactor configuration like the extractive upflow reactor which exploits gravity effects and the higher solubility of light hydrocarbons in the water phase. The oil phase being 2-8 times heavier than the water phase (depending on the operating pressure and temperature) circulates at the bottom of the reactor, while the water phase injected from the bottom flows through the bed of oil and exits the reactor at the top, in the process, extracting the light hydrocarbons soluble in the water phase. While experiments demonstrating the efficacy of the extractive upgrading process of heavy oil can be found in the literature, there remains a need to model and understand the complex coupling between the chemical reactions of interest and the interfacial mass transfer of hydrocarbon species and evaluate their effect on the final product yields.

With a view to address this need, we employ a two-phase well-stirred reactor (WSR) model with lumped chemical kinetics to study the overall coupling of the chemical reactions and the oil-water phase equilibrium in an extractive upflow reactor at P = 30MPa and T = 663,683 and 703K. Using this low-fidelity model, we demonstrate the improved yields of the upgrading process at higher inlet flow rates of water directly corresponding to higher rates of extraction of the light species from the oil to the water phase. This is in qualitative agreement with experimental demonstrations of the extractive upgrading configuration in literature. The WSR model also shows that the total interfacial species mass transfer rates in the reactor must be higher than the rates of retrograde reactions (condensation and coking reactions) for the best outcomes. This necessitates the need to develop methods to obtain realistic estimates for the interfacial species mass transfer rates in a realistic extractive upgrading reactor configuration.

In order to obtain a realistic range of values for the interfacial species mass transfer rates for different operating conditions of interest we use a multi-scale modeling approach, wherein the total interfacial species mass transfer rates in the reactor are estimated as the product of an average total interfacial area within the reactor and an averaged interfacial mass flux per unit area of the light hydrocarbon species.

The total average oil-water interfacial area in the reactor is calculated by simulating the macro-scale mixing process in a realistic lab-scale reactor geometry (diameter=30mm, height=100mm and 20g of oil feed) using two-phase flow simulations in 2-D with the Volume of Fluid (VoF) method with fluid properties like density, viscosity and interfacial tension estimated at the different given P, T conditions. The total average interfacial area in the reactor scales as the square root of the Weber number for a range of different operating conditions (pressure, temperature and inlet velocities). This is a useful intermediate result which can help engineers design extractive upgrading reactors for specified target light species extraction rates.

The average interfacial mass flux of light hydrocarbon species is estimated using 1-D droplet mixing simulations in spherically symmetric co-ordinates which account for the detailed oil-water phase equilibrium and diffusive fluxes at the interface. The heavy oil is modeled as a mixture of two pseudo-components, one corresponding to the light fraction and another corresponding to the heavy fraction in the mass ratio 1:1. The average interfacial mass flux of the light pseudo-component at P=30MPa increases from T=623K to 703K by more than an order of magnitude which is consistent with the rapidly increasing solubility of the light oil fraction in water with increasing temperature.

The total interfacial mass transfer rates estimated at P=30MPa and T=663-703K for inlet water mass flow rates in the range of 2.5-10 g/min and 20g of initial oil feed are an order of magnitude higher than the rates of the lumped cracking and coking reactions at these temperatures. This implies that the extractive upgrading reactor has the potential to effectively remove the light hydrocarbons from the oil phase as soon as they are formed and transport them safely out of the high temperature reaction zone. Although, this medium-fidelity approach to estimating the interfacial species mass transfer rates does not account for complex effects like the enhancement of the species fluxes due to sharpening of gradients in a flow field or the reduction in interfacial area due to interphase mass transfer, it nevertheless yields a reasonable first-order estimate which can be used to aid the design process of extractive upgrading reactors.