CO2 Storage in Saline Aquifers: Dynamic Modeling For Risk Management
Updated: Dec 9, 2022
By: Tariq Siddiqui
Risk analysis and dynamic simulation models can help guide geologic storage project by providing information on potential risks associated with CO2 leakage and other adverse impacts. The calibration of models with history match data provides confidence that available models can predict safe storage over the life of project. Models also helps in planning and designing the Monitoring, Verification and Accounting (MVA) plans during site characterization, injection and closure phase.
In this article we discuss quantitative risk assessments associated with the storage aspect of CCS projects and address the potential impact of subsurface technical risks within the physical domain across various categories (e.g., injectivity, storage volume, seal failure, induced seismicity, etc.). Although non-technical, above ground risks are equally important in strategic domain they are not discussed here. The focus here will be on:
Response Surface Modeling (RSM) guiding quantitative risk assessment
Risk Identification & Assessment Analysis (Impact & Severity)
Risk Mitigation & Consequence Management
1. Response Surface Models (RSM) for Quantitative Risk Assessment
Monte Carlo (MC) simulation is a well established method in managing subsurface uncertainties, computationally, however, it is resource prohibitive. RSM can substantially reduce the number of models to replace the fully coupled geological model. It has been successfully used in oil & gas projects as well as CO2 research for riskmanagement; especially for the top two subsurface uncertainties in CO2 storage that may lead to seriuous risks:
CO2 injected plume migrates out of Area of Review (AoR) - see fig below
Pore Pressure increasing during CO2 injection causing seal failure
Here we will illustrate use of RSM to quantitatively predict CO2 plume movement.
RSM is a 5-step process (see title figure)
1- A fully coupled geo-cellular model is constructed first.
2- A parameter sensitivity analysis is done using the model to rank top uncertainties/predictors for the plume movement (see Tornado chart). Here we assumeonly top three predictors are significant (permeability, porosity and Kv/kh).
3- An RSM or Design of Experiments (DoE) software is used to generate response data from the model for the calibration of Model Response Equation (MRE) in next step by generating number of experiments necessary (15 in this case) using top 3 predictors.
4- Regression analysis is then run in DoE to generate a MRE that is calibrated with data (previous step). A high R2 indicates a a high degree of confidence in predicting model response; in this case the CO2 plume migration distance.
5- Finally Monte Carlo simulation is run on calibrated MRE by providing probability distribution (input) to each of three predictors to generate Probability Density Function (output) or PDF for predicting expected plume distance (Low-Mid-High).
This quantitative probability assessment of plume distance is fed to experts in next step for Risk Assessment Matrix (RAM), providing a high degree of confidence.
3. Risk Identification & Assessment Analysis (Impact & Severity)
There are two broad classification of risks in CO2 storage projects
Site Specific Risks; Result in potential loss of storage integrity resulting in unplanned release of CO2 from its storage (usually ; spatial, site-specific quantitative and numeric). They can be predicted using simulations model. Examples of risks may include (not limited to):
Rise in pore pressure from injection causing induced seismicity
Plume migrates beyond the AoR and/or migrates to sensitive area
Contamination of groundwater by CO2
Undetected features like faults and fractures
Project Related Risks : These risks are generally, non-spatial and qualitative (Economic, Commercial Operational or Political risks). These risks although very important, are not quantified by dynamic models and are not detailed here. These could be:
Regulatory and permitting closure. Due to induced seismicity risk
Mechanical failures in the well (packer) requiring work-over
Accident while operations
The Risk Assessment Matrix (RAM) is a tool used in oil & gas industry that is well adapted in CO2 storage business for risk analysis (see title slide). The subject Matter experts (SME’s) are responsible for:
Identify the consequence of ‘Hazard’ if released (see next section).
Estimate the 'Severity ' (1- 5) of the consequence for each hazard
Estimate ‘Likelihood ' ( 1- 5) of each consequence
Example; likelihood of plume is quantitatively estimated using RSM models
Rank each hazard according to impact score
Impact Score (Hazard) = Severity x Likelihood
Score = 20 - 25 (RED) : Non-Operable
Score = 10 - 16 (ORANGE) : Intolerable
Score = 5 - 9 (YELLOW) : Undesirable
Score = 1 - 4 (GREEN) : Acceptable
Risk in all events/hazard must be mitigated to a ‘Acceptable’ level (GREEN), or demonstrated As Low As Reasonably Possible (ALARP) for YELLOW.
4. Risk Mitigation & Consequence Management
The Bow-Tie method can be used to analyze all risk sources, communicate risk scenariosand risk mitigation strategies. Broadly it has four element:
Threats - A physical characteristics of CO2 sites (wells, faults, equipment) that can pose apotential hazard resulting in top event or hazard.
Top Event / Hazards - The point in time when a discrete event results in a loss of control over the hazard, generally in short-term. Following are the top event / hazards orfailure modes that are commonly identified in risk assessment of CO2 storage projects:
Lateral Migration : CO2 plume moves out of planned AoR
Vertical Migration : CO2 moves out of confining zone (via faults fractures) into overburden rock and surface.
Wellbore Leakage : Movement of CO2 behind poorly cemented legacy wells to surface or near-surface
Induced Seismicity : The potential increase in pore pressure due to injection resulting in reactivation of faults and fractures.
Consequence - An unwanted top event has a consequence or a risk with certain impact (ex: release of CO2 to atmosphere), they are of generally longer-term. The consequence and its impact must be prevented by barriers.
Are the measures called ‘controls’ When taken to prevent before top event or hazard; or called ‘recovery’ when measure is taken to mitigate the ‘Consequence’ after the top event.