Geologic carbon sequestration is a promising solution to reducing atmospheric CO2 and mitigating climate change. Simulation technology will play a key role in assessing a site's viability and storage capacity. However, conventional reservoir simulators will need to be augmented with new physics to fully address the challenges of carbon storage, and will require exceptional performance to manage the additional computational burden
By the mid-2030s roughly 8.5 billion people will require housing, food, heating, and cooling, and a sizeable part of that population will strive for a life beyond poverty, while the rest aspires to improved living standards. These anticipated developments will require energy, indeed vast amounts of energy, and a substantial portion of that energy will still come from burning fossil fuels that create greenhouse gases. While there is more than one greenhouse gas, the majority of heat-trapping emissions are CO2, and the rising concentration of the gas in the earth’s atmosphere has become a critical focal point for scientists, policymakers, and energy companies around the globe who are coming to the opinion that reducing emissions in itself may not be sufficient to stop rising global temperatures. The most direct (but not necessarily easiest) approach is to take CO2 out of the atmosphere and store it away in a safe place underground. That is called geologic carbon sequestration.
Geologic carbon sequestration is performed by injecting and trapping CO2 into underground formations such as saline aquifers, basalt formations, or depleted oil and gas reservoirs. It is assumed to be the most effective way to sequester greenhouse gases, the other methods being biological and technical carbon sequestration. The successful application of this technology - successful meaning that we store away enough CO2 to impact climate change - requires the trapping of gigatons of gas for millennia. Humanity is presently emitting about 37 gigatons of CO2 per year and hence, achieving net-zero means all 37 gigatons must be captured and stored away annually; an unparalleled effort that will require huge investments in research, technology, and operations on a global industrial scale. While the technology to capture CO2 directly from the atmosphere is less developed, rather expensive, and energy-intensive, the science to store the gas underground is better understood. The science of fluid flow through porous media with accompanying chemistry and thermodynamics is the domain of reservoir engineering, a core competency of energy companies for decades.
The U.S. Department of Energy (DOE) has previously identified four pillars of research associated with geological carbon sequestration: (1) wellbore integrity; (2) sub-surface stress and induced seismicity; (3) permeability manipulation and flow control; and (4) new subsurface signals. While the first two are self-explanatory, the latter two require some explanation. Permeability manipulation and flow control address techniques such as fracturing the rock or manipulating flow paths by the use of cements, mud, or polymers up to the point of complete blockage if necessary. For example, in hydrocarbon production reservoir engineers adaptively change fluid flux magnitude and flow pathways to guide hydrocarbon-rich fluids to producer wells, while in nuclear waste disposal and carbon sequestration flow paths to the surface need to be blocked. The term ’new subsurface signals’ relates to the development of new sensors, automated data acquisition and monitoring, and better data analytics - mainly to ensure that trapped greenhouse gases don’t escape confinement.
The Carbon Storage Atlas from the DOE provides a coordinated update of carbon capture and storage potential across the United States and other portions of North America. Figure 1 shows a partial map of CO2 related geologic validation projects. In addition to US efforts, other nations have also started programs to assess technologies and evaluate possible storage sites. For example, project Greensand, funded by the Danish government, which has cleared the validation stage and is expected to inject 500,000 to one million metric tons of CO2 yearly into a 2 km deep offshore reservoir in the North Sea by 2025, gradually increasing the capacity to 4-8 million tons by 2030. To date, the amount of actual CO2 stored worldwide is minuscule compared to the projected needs or to the existing storage capacity of either prospective or validated CCS sites, which are estimated to have a total capacity between 10,000 and 80,000 gigatons.
Figure 1: Screenshot from NATCARB Viewer.
The technology is in its incipient stages, but it’s already clear that simulation will play a key role in modeling the physical processes for the assessment of a proposed site’s in-situ capacity and longevity. While new research simulators are under development, they do not have the performance nor do they integrate with workflows and tools now in use. With deep experience managing large capital projects, surveying geology, drilling wells and operating fields, the oil, and gas industry is well-positioned with the right skills and tools to play a key role in this emerging challenge. Likewise, mature industry modeling standards, conventions, and workflows from seismic to geo-model building to reservoir grid and simulation can be leveraged to provide a firm foundation and a huge head-start to CO2 storage modeling. However, new physics, discussed in more detail below, must be added to conventional simulators to address these emerging challenges.
There are roughly speaking four categories of trapping that operate on timescales from a few years to thousands of years and whose simulation require additional features beyond those in common use today by reservoir engineers. Let us abbreviate the time required for a significant (50% or larger) trapping contribution by “TTC” to indicate the timescale at which the different mechanisms work.
While one or another mechanism may dominate at a certain storage site over a given timescale, considering all of them (or at least the first three) seems to be essential to provide scientists and engineers with more realistic and accurate results for the actual storage capacity, maximum rate of injection, and the capacity for trapping CO2 over millennia in deep underground formations. Following is a brief overview.
Structural or stratigraphic trapping occurs when injected CO2, due to its buoyancy, rises in the pore space till it gets trapped by an impermeable layer of cap rock. Commercial multi-phase compositional reservoir simulators should be able to model the initial transport problem of moving injected super-critical CO2 through the matrix if both convection and diffusion are considered, but complications like fractures, the dissolution of carbonate rock, the change of effective diffusion coefficients, as well as restrictions on plume migration due to capillary trapping and immobilization of the injected gas can introduce inaccuracies in the computed distribution of gas saturation if the model data is inaccurate, or the numerical engine is not sophisticated enough to simulate additional geochemical effects.
Residual or capillary trapping occurs when injected CO2 is disconnected from the plume during secondary water imbibition and forms small bubbles inside the pore space that get trapped due to capillary pressure (snap-off effect). Depending on the storage environment, capillary trapping can achieve stable CO2 saturations between 10% and 30% and can account for 80% of the overall storage capacity in some cases. While the numerical methods to simulate the trapping mechanism are known in principle, the actual choice of three-phase relative permeability model and hysteresis has been shown to have a large impact on long-term storage forecasts and are therefore still subject to research and experimentation.
Solubility trapping refers to the dissolution of CO2 in deep saline aquifers, a process that evolves over time in the pore space and causes long-term, CO2 saturated water to sink to the bottom of the formation. The storage potential is rather impressive. For instance, under standard reservoir conditions at the Sleipner storage site in the Utsira formation (North Sea) one can achieve a storage ratio of up to 50 [kg/m3] of CO2 for an estimated total capacity of 600 Billion tons of CO2 for the whole formation. The process to be modeled starts with the injection of less dense supercritical CO2 into the formation, leading to structural trapping. Over time, the gas mixes with the brine, making the fluid denser, and the brine with the dissolved gas sinks to the bottom of the formation. The downward convective flow accelerates over time and drives lower-density brine upwards, leading to a circular convection-diffusion process. Well trajectories and injection rates have a known impact on the trapping mechanism and need to be studied for a given site.
Mineral trapping finally occurs when CO2 is sequestered by the precipitation of secondary carbonates in the matrix rock. This in-situ mineral carbonation happens when the injected and dissolved CO2 reacts with mineral oxides present in the formation. The successful simulation of mineral trapping requires new features that most commercial reservoir simulators only possess in part. The feature list includes but is not limited to the simulation of kinetically-controlled mineral dissolution and precipitation; computing the sensitivity of the solubility of CO2 w.r.t. pressure, salinity, and temperature in the formation, as well as historic mineral dissolution; modeling of new components and ion solubility together with a solid phase, and finally equilibrium calculations for the chemical reactions that lead to the forming of secondary carbonates.
The new physics adds significant computational burden to conventional simulators. In addition, models will cover a larger region including large aquifers, the grid resolution will be higher in order to resolve important small-scale physical phenomena and simulated time periods will be measured in millennia not decades. Furthermore, the lack of sufficient history and experimental data to calibrate models makes it difficult to validate results, increasing the need for ensemble studies. All of these factors emphasize the need for an ultra-fast engine operating at the highest performance levels. Looking forward, vast amounts of computer time will be required to simulate CO2 sequestration in different depositional environments under uncertainty.
With this in mind, we consider ECHELON, the industry’s fastest reservoir simulator, to be an excellent foundational platform on which to layer additional physics for carbon storage modeling. The design and implementation of the required feature set is a combined research and development activity that calls upon an interdisciplinary team of reservoir engineers, experts in physical chemistry, applied mathematicians, and high-performance computing specialists. The ECHELON Consortium, a partnership between energy companies and Stone Ridge Technology (SRT) to advance the development of ECHELON, provides exactly that - a diverse talent pool comprised of an international team of specialists with broad experience in research, simulation, and HPC software development.
ECHELON already supports the thermodynamics of CO2 injection and includes geomechanical coupling, an important aspect of modeling to predict microseismic events and potential well failures. Logical next steps include enhancing our existing feature set to treat fluid/brine interaction, plume migration, and diffusion. This needs to be followed or accompanied by the development of improved three-phase relative permeability models suitable to support the modeling of capillary trapping. Finally, geochemical simulation code that uses an advanced equation of state and is able to model chemical reactions between CO2, the brine, and the rock, as well as code for kinetically-controlled mineral dissolution and precipitation, will be required for the simulation of long-term mineral trapping.
At SRT we are working each day to build the fastest, most robust, most stable and numerically advanced simulator in the industry, capable of running the largest models to completion in an astonishingly short time. Speed and performance are an integral part of our culture and that speed will be needed to address the most challenging problems emerging from the energy transition.
Klaus Wiegand is Senior Scientific Advisor at Stone Ridge Technology. Prior to joining SRT Klaus developed numerical reservoir simulation software at ExxonMobil where he spent 22 years. Klaus has a Masters degree in Computational and Applied Math from Rice University.
What we are doing to help improve the reservoir simulation industry.
ECHELON now supports AMD Instinct acceleratorsRead article →
For our first benchmark study on the NVIDIA Hopper architecture, we present ECHELON performance numbers for 10 asset models executing on the NVIDIA Volta, NVIDIA Ampere, and NVIDIA Hopper architectures. The longstanding trend of performance scaling with memory bandwidth continues.Read article →
Leveraging modern GPUs and neural computing libraries, scientists are using deep learning technology to solve differential equations that emerge from a diverse set of physics-based problems. This blog introduces the topic and explores potential applications and limitations of the technology.Read article →