Quantitative assessment of monitoring strategies for conformance verification of CO2 storage projects
Introduction
The main storage-related challenges for deployment of large-scale Carbon Capture and Storage (CCS) are capacity, confidence and cost. There must be high certainty that the site has the capacity to permanently hold the volumes that are planned to be injected, and that it can be operated safely and economically. These elements are normally addressed in plans that operators submit to the regulating authority before approval to operate a storage site is granted. In Europe, the basis for assessment of these plans is provided by the European Commission (EC) CCS Directive (EC, 2009). It states that a monitoring plan is required to enable comparison between actual and modelled behavior, detection of irregularities, detection of the migration of CO2, and the assessment of the effectiveness of corrective measures in case of leakages or significant irregularities. It furthermore states that reports need to be submitted, at a frequency determined by the competent authority, that contain all information relevant for assessment of compliance with storage permit conditions, and for ‘increasing the knowledge of CO2 behavior in the storage site’.
The specific requirements for safe storage operations are identified in a storage permit. A general requirement is that injected CO2 remains within the storage complex permanently. The storage permit will also specify maximum injection rates and pressures, as well as the maximum allowed reservoir pressure. Additional requirements may result from site-specific risks and could therefore differ from one project to another. Key concepts in the monitoring of these risks are captured by the notions of containment and conformance. Containment refers to the basic requirement that CO2 must remain permanently within the storage complex or within clearly identified boundaries inside that storage complex. According to the EU CCS Directive, conformance (or conformity) refers to consistency of the actual behavior of the injected CO2 with the modelled behavior. Conformance is a requirement for the transfer of responsibility after site closure, together with the absence of detectable leakage, and a demonstration that the site is evolving towards a situation of long-term stability. A slightly broader definition of conformance, that we will adopt here, also includes compliance with any additional requirements as specified in the storage permit. This definition is in agreement with that used by the government of Alberta, Canada, in its the regulatory framework for CCS, and combines the concepts of concordance (agreement between models and data) and performance (agreement with permit requirements) as discussed by Oldenburg (2018). We will define conformance verification to mean the activity aimed at establishing if a situation of conformance exists at any moment during operation of the site and will remain to exist in the future. Since a storage permit will not be granted if initial modelling suggests that conditions qualifying as non-conformance are likely to develop, it may be assumed that initial models prepared by the operator will suggest that the site can be operated in accordance with storage permit requirements.
The actual behavior of the CO2 after injection in the reservoir will be strongly controlled by conditions (fluid pressure, temperature and composition, stress field) and rock properties (porosity, permeability, connectivity, fault stability) inside the storage reservoir, which are site specific. While these properties may be approximately known at the locations of wells that have been drilled into the reservoir before the start of injection, they will generally be poorly known everywhere else. As a consequence, also the dynamic behavior of the CO2 will be uncertain. This will also be the case to some extent for storage in depleted gas reservoirs for which information obtained from monitoring data is generally limited to storage volume dimensions and pressure. Some of the uncertainties can be associated with risks of non-conformance situations, such as poor injectivity (preventing injection of the planned volumes), leakage through old wellbores or through the overburden, or migration of CO2 outside of the licensed storage area (all leading to non-containment), and high seismicity (violating safe operation standards). Some of these risks may be more relevant in some sites than in others.
Monitoring programs should facilitate the identification of irregularities that could point at non-conformance situations and trigger mitigating actions. Information extracted from monitoring data will generally be uncertain due to measurement or interpretation errors and the sparsity or limited resolution of data, requiring the use of (also uncertain) models to fill the gaps (Harp et al., 2019). Conformance verification activities involve the comparison between modelled behavior of the CO2 and its ‘actual’ behavior, which must be inferred from monitoring data, and they must therefore take this uncertainty into account.
No clear guidelines or frameworks currently exist for identifying monitoring tasks and setting monitoring performance requirements in a conformance verification context (Bourne et al., 2014). However, experience with monitoring of industrial-scale CO2 storage operations has been gathered in a number of projects that include Sleipner (e.g. Arts et al., 2008; Furre et al., 2017), Snøvhit (Maldal and Tappel, 2004), In-Salah (Mathieson et al., 2011), Otway (Jenkins et al., 2015) and Quest (Bourne et al., 2014). Decisions about which monitoring technologies to adopt to address identified risks in these projects have generally been based on site-specific monitoring technology feasibility studies that employ qualitative or scenario-based approaches for risk assessment involving collective expert judgement (Bourne et al., 2014). A method often used in support of such approaches is the bowtie method which aims to map and rank threats, consequences, and safeguards. Safeguards include monitoring that can detect irregularities, and a decision logic to interpret the monitoring data and suggest control measures. The possible monitoring technologies are ranked by experts, and an ultimate selection is made by balancing benefits against costs. Applications of such a risk-based framework for developing site-specific measurement, monitoring and verification plans were presented by Bourne et al. (2014) for Quest, by Dean and Tucker (2017) for the Peterhead CCS project (Spence et al., 2014) and by Metcalfe et al. (2017) for the White Rose project. This framework aims to produce a systematic risk assessment and to establish monitoring performance targets that will reduce storage risks to a desired level. As part of this approach extensive modelling based on variation of identified uncertain model aspects was performed to assess the risk of CO2 migration beyond spill points or storage site boundaries. Active seismic surveys were planned to establish a baseline for repeat surveys and detect lateral movement of the CO2, and to identify changes in the formations above the storage formation. Revision of the monitoring plans would be necessary in the case of unexpected plume shape or migration velocity, injection pressures, or other deviations from the modelled behavior. Mismatches between dynamic models and monitoring data could then lead to model updates, additional monitoring or corrective measures. It may not always be clear a priori what the character of mismatches would need to be in order to trigger one or more of these actions. The current practice appears to be to conduct a qualitative assessment of such mismatches, leading to an expert-based judgement of required actions.
Quantitative assessment of risks, and the design of monitoring strategies that address them, is often considered complex and computationally demanding since it requires extensive evaluation of mathematical models that capture the full range of possible behavior of the system, identification of the ranges of possible values of all parameters that influence system behavior, and a derivation of risk under uncertainty. However, there are important benefits to such an approach. Perhaps the most important are the possibility to properly account for all uncertainties and their interdependencies, both in models and data (without the need to rank them based on extensive sensitivity experiments), and the possibility to assess the contribution of different monitoring technologies in quantitative terms.
Several examples of quantitative workflows related to risk management in CO2 storage operations can be found in the literature (see Section 2). This includes applications of optimization approaches to above-zone monitoring well placement aimed at minimizing the time to leak detection (Sun et al., 2013; Cameron, 2013; Yonkofski et al., 2016, 2017; Jeong et al., 2018) and workflows for leakage location estimation and uncertainty reduction based on monitoring data (Sun and Nicot, 2012; Jung et al., 2013; Jung et al., 2015; Hu et al., 2015; Chen et al., 2018; Chen et al., 2020). Some of these methods were also applied to the FutureGen 2.0 candidate storage site (Vermeul et al., 2016; Bacon et al., 2019) illustrating the feasibility of the application of quantitative life-cycle management workflows to realistic settings.
Chadwick and Noy (2015) considered modeling-monitoring convergence as an indicator of the possibility to demonstrate conformance. This is based on their observation that model predictions tend to become more reliable when models are in sufficient agreement with historic data. Following this idea, Harp et al. (2017) proposed an approach for identifying robust pressure management strategies based on decision gap theory. Rather than representing the uncertainty probabilistically, the methodology aims to quantify the degree that one can be incorrect in the characterization of the system and still ensure that performance (conformance) criteria are met. This could be viewed as a metric for comparing alternative monitoring strategies. This approach was adopted by Harp et al. (2019) in a follow-up study that demonstrated an application based on pressure monitoring at a single well where a conformance criterion was defined in terms of the pressure at the well. The workflow uses derivatives of observables with respect to uncertain model parameters which will not be generally available for more complex simulators and measurements or for conformance criteria that are not directly observable. The approach appears to require that a critical value for the conformance criterion is not exceeded in any of the model realizations, which seems quite restrictive. Also, it is not entirely clear how the approach should be extended to applications with multiple conformance criteria and large numbers of uncertain parameters, such as heterogeneous permeability fields.
As an alternative to the workflow of Harp et al. (2019), we develop and demonstrate a generic conformance analysis approach that can be applied to cases with arbitrary numbers and types of uncertainties and arbitrary combinations of monitoring techniques. The approach is inspired by ensemble-based workflows for quantitative evaluation of monitoring strategies for oil field management (e.g. Le and Reynolds, 2013; 2014; He et al., 2020, Barros et al., 2016; Barros, 2018). These studies have suggested that such approaches are indeed capable of providing useful comparisons and rankings of alternative strategies under realistically complex uncertainty scenarios. Here we will therefore propose a quantitative ensemble-based workflow that can be applied to objectively assess the current and future conformance state of the storage system. The workflow can also be used on top of practical risk identification and mitigation frameworks as currently used, to produce objective assessments of the validity of simulation models, the effectiveness of measurement acquisition strategies, and ultimately support decisions about contingency monitoring or corrective measures. We demonstrate an application of the workflow prior to the start of injection with the aim of evaluating possible monitoring strategies in terms of their capability to correctly assess the conformance of the site at any time in the future.
The remainder of this paper is organized as follows. In the next section we review the literature on quantitative workflows for evaluation of monitoring designs in subsurface applications. In Section 3 we define terminology related to conformance that will be used in this document. In Section 4 we introduce our proposed quantitative workflow for conformance assessment. Section 5 provides a relatively simple example application of the workflow to illustrate the main steps as well as an application to a complex case based on a real potential storage site. In Section 6 we discuss possible extensions and alternative applications of the workflow that could be useful in support of CO2 storage operations. Finally, the main conclusions are provided in Section 7.
Section snippets
Quantitative evaluation of monitoring plans
Early work on quantitative workflows for the assessment of monitoring strategies goes back to at least the mid-1970s, fueled by legislation in the US aimed at protecting groundwater resources from contamination originating from e.g. landfills or agricultural pesticide use. Later applications have included rainfall gauge placement throughout a watershed, river monitoring systems, water distribution networks, and sewer systems. Work in this domain includes studies on pollution source
Conformance assessment
In this section we define the terms and concepts that will be used in the conformance verification workflow. A central concept in CO2 storage is containment, meaning the (permanent) retainment of all CO2 within safe boundaries after injection. These boundaries could be the bounding faults and storage seals of an entire storage complex, possibly consisting of multiple compartments or layers, or alternatively, internal boundaries, defined by internal faults or spill point contours. We define
Uncertainty quantification
Uncertainty in the expected behavior of CO2, and in the behavior of the site as a whole, during and after injection can be related to many factors. The main reservoir-related factors are associated with the properties of the reservoir rock such as permeability and porosity, the presence of conducting or sealing fractures, faults, and baffles, the distribution of saturation and temperature of fluids and gases already present in the reservoir, and the physical-chemical interaction of the CO2 with
Conceptual 2D model example
In a first illustrative example, we consider a 2D model of a storage aquifer with dimensions 1380 × 1380 × 10 m discretized into 69 × 69 × 1 = 4761 active grid blocks. The model has two vertical wells: an injection well operated with a prescribed rate target of 1 × 105 m³/day (at standard conditions) and maximum allowed injection pressure of 200 bar, and a brine discharge well which is opened 180 days after the start of injection and is operated at a fixed bottom-hole pressure of 80 bar. The
Discussion
The results described in the previous section show how our proposed procedure for quantitative conformance assessment can be used to evaluate and compare different monitoring strategies, providing the key elements to support decisions on the optimal design of monitoring plans for CO2 storage projects. In addition, this type of analysis can be used to determine possible technical limits of existing monitoring technologies in specific applications, in particular regarding the level of confidence
Summary and conclusions
We have presented a novel model-based framework for quantitative conformance verification. We have provided a formal notion of conformance for use in quantitative approaches along with the required ingredients to define useful quantifiable conformance criteria. The workflow builds on concepts used in state-of-the-art subsurface reservoir management practices, such as ensemble-based uncertainty quantification and history matching. We have demonstrated how the developed methodology can be used to
CRediT authorship contribution statement
E.G.D. Barros: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Supervision, Visualization, Writing – original draft, Writing – review & editing. O. Leeuwenburgh: Conceptualization, Methodology, Formal analysis, Investigation, Supervision, Writing – original draft, Writing – review & editing. S.P. Szklarz: Methodology, Software, Validation, Formal analysis, Visualization, Writing – review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work has been produced with support from the SINTEF-coordinated Pre-ACT project (Project no. 271497) funded by RCN (Norway), Gassnova (Norway), BEIS (UK), RVO (Netherlands), and BMWi (Germany) and co-funded by the European Commission under the Horizon 2020 programme, ACT Grant Agreement no. 691712. We also acknowledge the industry partners for their contributions: Total, Equinor, Shell, TAQA. The Smeaheia model was provided by the partners of the Northern Lights consortium (Equinor, Shell,
References (74)
- et al.
Risk-based post injection site care and monitoring for commercial-scale carbon storage: reevaluation of the FutureGen 2.0 site using NRAP-Open-IAM and DREAM
Int. J. Greenhouse Gas Control
(2019) - et al.
A risk-based framework for measurement, monitoring and verification of the Quest CCS Project, Alberta, Canada
Int. J. Greenhouse Gas Control
(2014) - et al.
Geologic CO2 sequestration monitoring design: A machine learning and uncertainty quantification based approach
Applied Energy
(2018) - et al.
Reducing uncertainty in geologic CO2 sequestration risk assessment by assimilating monitoring data
IJGGC
(2020) - et al.
A risk-based framework for Measurement, Monitoring and Verification (MMV) of the Goldeneye storage complex for the Peterhead CCS project, UK
Int. J. Greenhouse Gas Control
(2017) - et al.
20 years of monitoring CO2-injection at Sleipner
Energy Procedia
(2017) - et al.
Development of robust pressure management strategies for geologic CO2 sequestration
Int. J. Greenhouse Gas Control
(2017) - et al.
A metric for evaluating conformance robustness during geologic CO2 sequestration operations
Int. J. Greenhouse Gas Control
(2019) - et al.
The state of the art in monitoring and verification - ten years on
Int. J. Greenhouse Gas Control
(2015) - et al.
Cost-optimal design of pressure-based monitoring networks for carbon sequestration projects, with consideration of geological uncertainty
Int. J. Greenhouse Gas Control
(2018)
Early detection of brine and CO2 leakage through abandoned wells using pressure and surface-deformation monitoring data: Concept and demonstration
Advances in Water Resources
Conformity assessment of monitoring and simulation of CO2 storage: a case study from the Ketzin pilot site
Int. J. Greenhouse Gas Control
CO2 underground storage for Snøhvit gas field development
Energy
Improving parameter estimation with an efficient sequential probabilistic collocation-based optimal design method
J. Hydrol. (Amst)
Sub-surface risk assessment for the Endurance CO2 store of the White Rose project
UK. Energy Procedia
Salah CO2 Storage JIP: CO2 sequestration monitoring and verification technologies applied at Krechba, Algeria
Energy Procedia
Bayesian analysis of data-worth considering model and parameter uncertainties
Adv. Water Resour.
Value of information analysis for adequate monitoring of carbon dioxide storage in geological reservoirs under uncertainty
Int. J. Greenhouse Gas Control
The peterhead-goldeneye gas post-combustion CCS project
Energy Procedia
Inversion of pressure anomaly data for detecting leakage at geologic carbon sequestration sites
Advances in Water Resources
Optimal design of pressure-based, leakage detection monitoring networks for geologic carbon sequestration repositories
Int. J. Greenhouse Gas Control
Data-space approaches for uncertainty quantification of CO2 plume location in geological carbon storage
Adv. Water Resour.
An overview of the monitoring program design for the FutureGen 2.0 CO2 storage site
IJGGC
Sequential data-worth analysis coupled with ensemble Kalman filter for soil water flow: a real-world case study
J. Hydrol. (Amst)
Conditioning of deep-learning surrogate models to image data with application to reservoir characterization
Knowl. Based Syst.
An optimization approach to design monitoring schemes for CO2 leakage detection
Int. J. Greenhouse Gas Control
Optimized, Budget-constrained Monitoring Well Placement Using DREAM
Energy Procedia
Ten years' experience of monitoring CO2 injection in the Utsira Sand at Sleipner, offshore Norway
First Break
Modelling Tools For Integrating geological, Geophysical and Contamination Data For Characterization of Groundwater plumes. PhD Thesis
Value of information in closed-loop reservoir management
Comput. Geosci.
Value of Information in Closed-Loop Reservoir management. PhD Thesis
Quantifying the Efficiency of Surveillance Strategies for Subsurface CO2 Storage Applications
Quantifying Efficiency Of Field-Wide Geophysical Surveys For Verifying CO2 Plume Conformance During Storage Operations
Practical Quantitative Monitoring Strategy Assessment For Conformance Verification of CO2 Storage Projects
Quantitative Conformance Assessment in CO2 Storage Reservoirs Under Geological Uncertainties Using Convolutional Neural Network Classifiers
Early Detection of Non-Conformance in Monitoring of CO2 Storage Reservoirs Using Auto-Encoder Neural Networks
Quantitative Decision Analysis for CO2 Storage Conformance Management: a Synthetic Case Study at Smeaheia, North Sea
Cited by (11)
Evaluating geophysical monitoring strategies for a CO<inf>2</inf> storage project
2024, Computers and GeosciencesA review of risk and uncertainty assessment for geologic carbon storage
2024, Renewable and Sustainable Energy ReviewsAssessment and uncertainty quantification of onshore geological CO<inf>2</inf> storage capacity in China
2022, International Journal of Greenhouse Gas ControlThe Implementation of a Deep Neural Network (DNN) Approach in a Case Study Predicting the Distribution of Carbon Dioxide (CO<inf>2</inf>) Gas Saturation
2024, IOP Conference Series: Earth and Environmental ScienceSubsurface carbon dioxide and hydrogen storage for a sustainable energy future
2023, Nature Reviews Earth and EnvironmentAn integrated framework for optimal monitoring and history matching in CO <inf>2</inf> storage projects
2023, Computational Geosciences