6.2.3 Features, Events and Processes methodology as an approach to risk assessment for CO2 storage

Many of the ongoing risk efforts are now cooperating to identify, classify and screen all factors that may influence the safety of storage facilities, by using the Features, Events and Processes (FEP) methodology (IPCC, 2005). Risk identification (or qualification) of hazards includes estimation of the probability of specific features, events, and processes (FEPs) that could contribute to, or prevent, unplanned CO2 migration from the confining zone (NETL, 2011; DNV, 2010):

  • Features include the physical characteristics or properties of the system, such as lithology, porosity, permeability, caprock thickness, faults, wells, leaky wellbores and nearby communities.

  • Events include discrete occurrences that may occur in the future affecting one or more components of the system, such as earthquakes, subsidence, drilling, penetration of the storage site by new wells, injection pressure increases, borehole casing leak, pipe fracture and well blow-outs.

  • Processes include physico-chemical processes, often marked by gradual or continuous changes, that influence the evolution of the system; chemical reactions, precipitation of minerals, ground water flow, multiphase flow, CO2 phase behaviour, gravity-driven CO2 movement or residual saturation trapping, geomechanical stress changes and corrosion of borehole casing.

Fig. 6-6 illustrates the relationship between, features, events and processes (FEP) and potential risk impacts. For example, the storage reservoir may have insufficient capacity or injectivity, leading to the risk that CO2 injection cannot be sustained over the life of the project. The impact assessment would estimate the techno-economic and societal impacts of such a scenario (NETL, 2011).

E. Fig . 6-6

Fig. 6-6: Examples of relationships among Features, Events, Processes, and Potential Impacts (NETL, 2011).

The risk is based on simulations of different scenarios built up from FEPs. Main steps in the assessment are: (a) establishing risk criteria, (b) description of the geological system by investigation and screening of all features,events and processes (FEPs) that are relevant to the long-term safety, so called FEP analysis, (c) identification of the most important FEPs, (d) scenario selection and analysis based on the FEP analysis, (e) system model development using numerical reservoir simulation, and (f) qualitative and quantitative consequence analysis (NETL, 2011; Chadwick et al., 2008).The very first step of risk is the definition of the assessment basis, which consists of: (a) risk acceptance criteria, (b) containment concept and (c) setting of the storage site.

In some cases, for the assessment of a storage structure, a modified performance assessment (PA) methodology can be used. PA is a system analysis that predicts the behaviour or "performance" of an element of a geological storage project (specified system) relative to one or more performance standards (system performance indicators). If the indicator is a health, safety and environmental (HSE) effect, the PA is termed a safety assessment.

The FEP analysis is performed using databases developed in earlier CO2 safety assessment studies (Wildenborg et al., 2005; Maul and Savage, 2004). The databases are used as selection tools for early screening of relevant FEPs. The main steps in the FEP analysis are illustrated in Fig. 6-7. The main tools that support the process are the FEP database and the visual analyser (Chadwick et al., 2008). A distinction can be made between features as static factors, and events and processes (EPs) as dynamic factors. For each individual EP the following aspects can be evaluated: (i) specifications of how the EP is interpreted, e.g. its relation to safety, (ii) semi-quantitative probability that an EP will occur, and (iii) potential impact if the EP occurs. EP grouping can be carried out and criteria for EP groups can be based on the information that is available in the FEP database (Wildenborg et al., 2005), such as: (i) common parameters (distinct features such as permeability, rock strength, etc), (ii) process types (mechanical, chemical, thermal, hydraulic, biological), (iii) effect type (on matrix, fluid, stored CO2, indirect), (iv) timescale of EP occurrence (in 100 years, in 1000 years or in 10000 years), (v) duration scale of EP while occurring (hours, days, centuries and longer), and (vi) spatial scale (metres, km, tens of km, basin scale) (Chadwick et al., 2008).

E. Fig . 6-7

Fig. 6-7: Main steps in used FEP analysis methodology. Based on the analysis process in Wildenborg et al., 2005 and Chadwick et al., 2008.

The FEP database holds FEPs that may have a potential effect on the safety of the storage system (Chadwick et al., 2008). It can help the site-specific description of the system and identification of site-specific issues, allowing comprehensive evaluation of each site's unique characteristics (CSLF, 2009). FEP database also ties information on individual FEPs to relevant literature and allow classification with respect to likelihood, spatial scale, time scale and so on. However, there are alternative approaches (IPCC, 2005). All FEPs in the database have a complete set of identification and classification attributes (Fig. 6-8). These attributes have been assigned generically, and could be used to filter case-specific FEPs with respect to the assessment basis (Chadwick et al., 2008).

Detailed lists of FEPs for geologic systems have evolved for various environmental needs, and these have been adapted to a generic database for geologic storage of CO2 by Quintessa (Savage et al., 2004; Maul et al., 2005). The Quintessa database (http://www.quintessa-online.com/fep.php) currently includes around 200 FEPs in a hierarchical structure, with individual FEPs grouped into eight categories. Each FEP has a text description and an associated discussion of its relevance to long-term performance and safety. Key references from the published literature are included to enable retrieval of more detailed information for each FEP. The database incorporates hyperlinks to other relevant sources of information (reports, websites, maps, photographs, videos, etc.), and is searchable in a variety of ways. The generic FEP database is intended to be the first stage in developing a FEP-based auditing capability for more detailed project-specific FEP databases. At present there are no project specific FEP databases in the system, but the capability is present and it is hoped that some project-specific databases will be added in the near future.

E. Fig . 6-8

Fig. 6-8: FEP Example of generic FEP attributes in the FEP database (Chadwick et al., 2008).

Utilising the definition of the storage system set out in the assessment basis, the FEPs are ranked and screened in order to identify the FEPs that are likely or very likely to occur. These FEPs are grouped and assigned to specific zones within the geological storage system (compartments). Because the future evolution of a geologic system cannot be precisely determined, various possible Scenarios for possible evolutions of the system and situations of particular interest are developed (NETL, 2011). Most Risk assessments involve the use of scenarios that describe possible future states of the storage facility and events that result in leakage of CO2 or other risks. Each scenario may be considered as an assemblage of selected FEPs (IPCC, 2005). Some rRisk assessments define a reference scenario that represents the most probable evolution of the system. Variant scenarios are then constructed with alternative FEPs. Various methods are used to structure and rationalize the process of scenario definition in an attempt to reduce the role of subjective judgements in determining the outcomes (IPCC, 2005). For example, based on the FEP analysis and the scenario formation, some "what if?" scenarios can be identified for simulation: (i) reference scenario assuming that no failure of the containment zone occurs, (ii) leaking seal scenario assuming that the seal will fail by geochemical processes, whereby CO2 enhances the permeability of the caprock and migrates into the overburden. (iii) leaking well scenario assuming that the sealing capacity of an existing old well will fail, followed by transport of CO2 along the well trajectory, (iv) leaking fault scenario assuming that there is a fault through the caprock, and that the sealing capacity of the fault will fail, followed by CO2 escape from the containment zone along the fault (Chadwick et al., 2008). The data input to early stage risk will frequently be associated with significant uncertainty. Consequently, early stage risk may be qualitative, based on FEPs, to lead to site selection or data characterization. It should be refined over time to incorporate new data, and at least at the closure of a project a minimal confidence should be gained to proceed with transfer of responsibility. Such analyses are typically based on expert elicitation activities, implying that the results to some extent depend on the subjective views and opinions of the experts involved. A key challenge is therefore how to enhance repeatability and consistency of risk assessments to make the associated process and results verifiable and auditable (DNV, 2010).

The FEP assessment methodology is useful but still has gaps in knowledge. This concerns many aspects, e.g. safety and risk terminology, usage of FEP database, scenario evaluation, assessment criteria, modelling tools and so on (Chadwick et al., 2008). Figure 6.9 shows different stages in a FEP analysis, from identification to scenario formation. The FEP approach has been used in many of the initial CO2 storage efforts, such as Sleipner in Norway (Torp and Gale, 2003), Weyburn in Canada (Stenhouse et al., 2006a, Stenhouse et al., 2006b), In Salah in Algeria (Riddiford et al., 2005), and the Decatur Project in the Illinois basin of the United States (Hnottavange-Telleen et al., 2009).

E. Fig . 6-9

Fig. 6-9: Different stages in a FEP analysis, from identification to scenario formation (Savage et al., 2004; Condor et al., 2011).

Relational approaches with FEPs

There are a number of different ways in which the FEPs and their relationships can be developed to describe a site's behaviour. The retained FEPs are classified in both spatial and contextual terms (Savage et al., 2005; Korre and Durucan, 2009). Three approaches have been used (CSLF, 2009):

  1. A "top-down" approach. An example of this approach is the Master Directed Diagram (mDD) approach, which was developed by Nirex of the UK (Nirex, 1998). An MDD is a diagram like a tree-like structure that has some of the attributes of a network.

  2. The Process Influence Diagram (PID) approach, which identifies and represents all possible influences between all FEPs within a system.

  3. The interaction matrix approach. FEPs representing components of the system under consideration are placed on the leading diagonal elements (LDEs) of the matrix. Interactions between LDEs are then noted in the off-diagonal elements (ODEs).

Among these approaches, the PID has been used for the risk in the Weyburn CO2 storage project (Stenhouse et al., 2005) and matrix representations of FEP interactions was applied to a hypothetical CO2 storage project (Savage et al., 2005; Korre and Durucan, 2009).

Regarding FEP methodology, there is some discussion as to whether a 'bottom-up' or 'top-down' approach is the best. 'Bottom-up' involves identifying every conceivable FEP and then building scenarios from these. This approach is time-consuming and might miss key scenarios through 'participant exhaustion' and time limitations. The 'bottom-up' approach uses the database directly to develop the assessment tools. 'Top-down' involves identifying a limited number of key risk scenarios. This approach might miss important FEPs, and important potential scenarios. In the 'top-down' approach, a FEP database can be used as an audit tool to ensure all relevant FEPs are included in the models, to document the reasons why others have not been considered (Condor at al., 2011) and to check completeness of the scenarios built. Overall, the 'top-down approach is favoured, but irrespective of the approach, it is important that the link between FEPs and scenarios is fully documented. An important issue connected with FEP/scenario risk analysis is that worst-case processes tend to be emphasised irrespective of how (un)likely they are to actually occur. Thus, leakage scenarios tend to get highlighted and qualifying uncertainties and assumptions ignored. Overall, quantitative assessment of the probability of any particular scenario occurring is very difficult, particularly for scenarios involving geological FEPs (e.g. fault leakage, caprock failure etc). An alternative to quantitative risk analysis may be to set out a storage plan, based on robust site characterisation, identify site-specific containment risks (and uncertainties), and design an efficient monitoring and remediation strategy (Chadwick et al., 2008).

Steps after FEPs analysis

After FEPs analysis, detailed site characterization and simulation provide data to assess exposure due to the vulnerabilities in a qualitative or quantitative manner. Successful CO2 geological storage requires thorough site characterization, especially for storage in saline formations that have not previously been considered an economic resource, as well as a clear understanding of the processes and mechanisms by which CO2 is transported and trapped (NETL, 2011). The estimated exposure indicates the probability that a particular negative event would occur. In the subsequent step, the effects of the vulnerabilities (impacts) are assessed using qualitative or quantitative tools. The impacts and exposure data from the previous two stages are used to assess the risk in the final step of the risk process, namely, risk characterization (Figure6. 4), through which the probability of the occurrence of events and the magnitude of loss from them are determined (effects assessment). In risk characterization, exposure and effects data are integrated to produce qualitative, semi-quantitative, or quantitative measures of risk. Ultimately, the set of quantitative and qualitative risk factors and their potential impacts become the basis for developing practical risk management and mitigation plans (NETL, 2011).