[Paper Review] Scaling up FluidFlower results for carbon dioxide storage in geological media

Anthony R. Kovscek, Jan M. Nordbotten+1 more|arXiv (Cornell University)|Jan 24, 2023

CO2 Sequestration and Geologic Interactions4 citations

TL;DR

This paper develops scaling relationships for CO2 storage in geological formations by extending results from the FluidFlower large-scale visualization experiment. It shows that an hour in the lab experiment scales to hundreds of years in real reservoirs, with convective mixing and CO2-brine mass transfer occurring rapidly during injection, significantly reducing the time to onset of viscous fingering and enhancing early mixing and dissolution.

ABSTRACT

The partial differential equations describing immiscible, but soluble, carbon dioxide (CO2) displacement of brine are developed including local mass-transfer effects. Scaling relationships for characteristic time among laboratory and representative storage formation conditions are found upon assumption that free-phase CO2 transport during injection is dominated by convection. The implication is that an hour in the FluidFlower (large-scale visual model) scales to hundreds of years of elapsed time in the storage formation. The scaling criteria permit extrapolation of the effects of changes in parameters and operating conditions. Interphase mass transfer allows CO2 to saturate the brine phase and such mass transfer is a significant nonequilibrium phenomenon. Significant mixing of CO2 dissolved into formation brine with original brine is found experimentally and is also predicted. The magnitude of onset time for downward migrating fingers containing CO2 is typically only a fraction of the duration of CO2 injection and in general agreement with theoretical analysis in the literature. Predictions for onset time of convective mixing at representative storage formation conditions likewise teach that the onset time for viscous fingering is significantly less than the duration of CO2 injection in some cases. The implications of this observation include that mixing of CO2 with brine and the subsequent settling due to gravity are relatively rapid and coincide with the period of active CO2 injection.

Motivation & Objective

• To establish scaling laws that translate laboratory-scale FluidFlower experiment results to real geological CO2 storage formations.
• To quantify the time scaling between lab experiments and field-scale reservoir conditions, particularly focusing on CO2 injection and convection dynamics.
• To investigate the role of interphase mass transfer and nonequilibrium effects in CO2 dissolution into brine during displacement.
• To determine the onset time of convective mixing and viscous fingering under representative reservoir conditions.
• To assess whether mixing and dissolution occur predominantly during active CO2 injection or afterward, based on experimental and theoretical analysis.

Proposed method

• Derives partial differential equations that model immiscible, soluble CO2 displacement of brine, including local mass-transfer effects.
• Applies scaling analysis under the assumption that free-phase CO2 transport is dominated by convection, enabling time scaling between lab and field conditions.
• Uses experimental observations from the FluidFlower setup to inform and validate the scaling relationships.
• Incorporates theoretical analysis of convective mixing onset times, comparing them to injection duration.
• Models interphase mass transfer as a significant nonequilibrium process, accounting for CO2 saturation in the brine phase.
• Predicts the onset time of downward-migrating CO2 fingers and compares it to the duration of CO2 injection.

Experimental results

Research questions

• RQ1 How do time scales from the FluidFlower experiment scale to real geological CO2 storage formations?
• RQ2 What is the onset time for convective mixing and viscous fingering in representative reservoir conditions, and how does it compare to injection duration?
• RQ3 To what extent does interphase mass transfer between CO2 and brine influence mixing dynamics during injection?
• RQ4 How rapidly does CO2 dissolve into brine, and is this process primarily active during or after CO2 injection?
• RQ5 What scaling criteria can be derived to extrapolate laboratory results to field-scale CO2 storage operations?

Key findings

• An hour in the FluidFlower experiment scales to approximately hundreds of years in actual geological storage formations due to convection-dominated CO2 transport.
• The onset time for downward-migrating CO2 fingers is typically a fraction of the total CO2 injection duration, indicating early development of convective mixing.
• Interphase mass transfer is a significant nonequilibrium phenomenon, leading to rapid CO2 saturation in the brine phase and enhanced mixing.
• Experimental and theoretical results show that convective mixing and dissolution occur predominantly during active CO2 injection, not afterward.
• The scaling criteria derived allow reliable extrapolation of parameter and operational changes from lab to field conditions.
• Predictions confirm that viscous fingering onset times are significantly shorter than injection durations in many cases, supporting rapid early mixing and dissolution.