Final Report Summary - CO2CARE (CO2 Site Closure Assessment Research)
Executive Summary:
The CO2CARE (CO2 Site Closure Assessment REsearch) project focused on site closure and preparation for transfer of liability of a CO2 storage project in order to assist regulatory authorities and stakeholders in implementing the EU Directive 2009/31/EC on CO2 Geological Storage. The project, which started in January 2011, was funded by the EU 7th Framework Programme and the industry and ran for a period of three years until December 2013. CO2CARE consisted of an international consortium of 23 partners from Europe, USA, Canada, Japan and Australia, represented by universities, research institutes, and energy companies. In order to incorporate up-to-date results and monitoring data 9 key injection sites in Europe and worldwide formed an integral part of the project: (1) Ketzin, Germany; (2) Sleipner, Norway; (3) K12-B, The Netherlands; (4) Rousse, France; (5) Montmiral, France; (6) Frio, USA; (7) Wallula, USA; (8) Nagaoka, Japan and (9) Otway, Australia.
The main objectives of the project were closely linked to the three high-level requirements of the EU Directive (Article 18) for the ultimate transfer of responsibility of a CO2 storage site to the Competent Authority typically 20 to 30 years after site closure. These high-level requirements are:
1. Absence of any detectable leakage.
2. Conformity of actual behaviour of the injected CO2 with the modelled behaviour.
3. The storage site is evolving towards a situation of long-term stability.
In particular, the project mainly focused on three key areas:
1. Well abandonment and long-term integrity.
2. Reservoir management and prediction from closure to the long-term.
3. Risk management methodologies for long-term safety.
These objectives were achieved via integrated laboratory research, field experiments and state-of-the-art numerical modelling, supported by literature reviews and data from the above mentioned portfolio of real storage sites, covering a wide range of geological and geographical settings. The current regulatory requirements and well and site abandonment technologies as well as the behaviour of abandoned CCS/CO2-exposed wells were reviewed and assessed. Detailed studies dealt with laboratory experiments on rocks and cements under different CO2 environments and in combination with numerical simulations of the well geomechanical integrity while also considering in parallel main geochemical aspects. Possible leakage pathways have been studied including the monitoring aspects and a specific remediation technique has been investigated on the basis of field representative cases and scenarios with the fruitful involvement of the project´s industrial partners.
Key research findings from the CO2CARE project, compliant with the EU storage regulation, were distilled in form of Best Practice Guidelines and disseminated to key stakeholders. Main results of so-called “dry runs” for the storage sites Ketzin, Sleipner and K12-B, which were discussed with European regulators, also contributed to the guidelines. These dry-run documents provide a template for site abandonment and transfer of responsibility, based on real storage sites and useful for both site operators and regulators. In the first instance the documents could be used at the site licencing stage to develop initial site closure plans and to inform early discussions and negotiations between the interested parties. Subsequently they could be used to inform the developing dialogue between site operator and regulator as the closure and post-closure phases are reached. The Best Practice Guidelines document will help establishing effective protocols for site abandonment. They provide technical guidance on how best to meet the three key requirements of the Storage Directive (viz. conformance, leakage and long-term stability) and also wider guidelines on post-closure wellbore and reservoir management and a robust protocol for risk management. In addition the Best Practice Guidelines contain recommendations directly relevant for the forthcoming review of the EU Storage Directive.
Summary description of project context and objectives
With the announcement of the European Strategic Energy Technology (SET) Plan, Carbon Capture and Storage (CCS) is expected to significantly contribute to the reduction of CO2 emissions in the atmosphere en route to a low-carbon economy. In anticipation of major CCS deployment by 2020, comprehensive research programmes should deliver the necessary operational and post-closure site management technologies for CCS. As far as closure and post-closure issues are concerned, apart from a few recently completed pilot scale field CO2 injection projects, and experience gained from enhanced oil recovery (EOR), the CCS industry has quite limited experience.
The fundamental regulatory principles surrounding site closure, transfer of responsibility (liability) to the competent authority, and post-closure obligations are set out in the EU Directive on CO2 Geological Storage (EU Directive). Pursuant to Article 18 of the EU Directive, the overall philosophy of the Directive regarding site closure is enshrined in the three high-level criteria for transfer of liability:
1. Absence of any detectable leakage.
2. Conformity of actual behaviour of the injected CO2 with the modelled behaviour.
3. The storage site is evolving towards a situation of long-term stability.
These conditions define satisfactory long-term site performance at a high-level. Significant gaps however remain in understanding how these high-level principles will be implemented at real sites, not only at the closure stage, but also at storage commencement, when future closure arrangements will need to be incorporated into a site storage license acceptable to both operator and regulator.
The Directive does not define the specific technical acceptance criteria, based on real site performance data, which can demonstrate that a site meets the three requirements. The identification of such criteria and the development of site abandonment procedures and technologies which guarantee the fulfilment of these criteria have been the main objectives of CO2CARE.
The experience and knowledge gained from field pilots suggest that fluid behaviour and its interaction with the containing rocks determine the overall performance of the CO2 storage complex during the injection and post-closure periods. The key issues in this respect are: injectivity; wellbore behaviour; and the response of fractures and faults in the reservoir, seal and geosphere. At present, the most challenging aspect of site abandonment is to predict, validate and demonstrate storage site behaviour during the post-closure period, as current experience of monitoring techniques and predictive simulations is restricted to ongoing injection projects and analogous research results from oil and gas production. The longest post-injection monitoring periods associated with field pilots involve Nagaoka (Japan) and Frio (USA) where injection has ceased four to five years ago. Both sites are included in the CO2CARE portfolio. However, monitoring and verification of medium to very long-term storage performance prediction is not available.
CO2CARE conducted integrated laboratory and field experimental research, supported by advanced numerical modelling, and developed methodologies and strategies that will be needed to meet the site abandonment, post-closure safety and transfer of responsibility requirements of the EU Directive. The meet its overall objectives, the CO2CARE project was structured in six research and development work packages to address the different aspects of site abandonment: 1) Current practices and regulatory framework, 2) Well abandonment, 3) Post-Closure Reservoir Management, 4) Risk Management, 5) Best Practice and Regulatory Compliance, and 6) Dissemination. A seventh work package addressed the Project Management. The main objectives of the project have been achieved through:
• review of tested site-abandonment procedures and site-performance criteria from oil and gas industry;
• critical review of the existing international CO2 geological storage regulations, comparison with current practice in related/relevant fields, identifying differences and knowledge gaps on site abandonment technologies;
• compilation of a database of first abandoned CCS/CO2-exposed wells with the benefits arising from data available to international partners of the project in particular;
• research into the factors critical to well and reservoir behaviour focusing on integrity assurance after site abandonment at all-time scales;
• laboratory and field research into well plugging techniques to improve well integrity and monitoring methodologies to monitor wellbores during the closure and abandonment phases;
• improvement and validation of the reliability of predictive numerical modelling of future storage site behaviour, studying the conformity between monitored and simulated site behaviour, and assessing short-term performance and long-term stability of storage;
• understanding of issues crucial to long-term storage integrity by studying both long-term potential deterioration and stabilisation processes at the sites of the CO2CARE portfolio;
• assessment of the applicability and effectiveness of current technologies used in the oil and gas industry for CO2 storage, wellbore and site remediation and, if necessary, development of new methodologies, with associated protocols for site operators and authorities;
• development and implementation of reliable risk management procedures and tools for site abandonment and assessment of post-closure storage safety, definition of appropriate long-term monitoring and remediation plans;
• development of fully documented ‘dry-run’ applications for transfer of responsibility at the three real storage sites Sleipner, Ketzin and K12B, incorporating interactive stakeholder input and subject to a project regulatory critique and review;
• development of guidelines for regulatory compliance and best practice guidelines on CO2 storage site abandonment, taking into account the site dependence of the research findings;
• dissemination of project´s results to stakeholders and public, contribute to public acceptance.
To allow for “real-site” data and research and to achieve the project objectives CO2CARE has incorporated up-to-date results and monitoring data from the main European and international CO2 injection sites through the industrial partners of the project and CSLF collaborators. The field site portfolio included:
• Ketzin (Germany) – First European on-shore CO2 storage in a saline aquifer with a reservoir depth of about 630 m. The site has injected slightly more than 67,000 tonnes of CO2 from June 2008 to August 2013. One injection and two observation wells were available since 2007; two further observation wells were available since 2012.
• Sleipner (Norway) – Off-shore saline aquifer, about 1 M tonnes per year of CO2 injected since 1996. Aquifer depth 800 to 1,000 m, 17 years of monitoring data available.
• K12-B (The Netherlands) – Offshore Enhanced Gas Recovery (EGR) over 60,000 tonnes of CO2 injected since 2004 in a reservoir at approximately 3,800 metres depth. The nearly depleted gas field comprises multiple wells, both producers and injectors.
• Rousse (France) – On-shore depleted gas reservoir. Injection of 120,000 tonnes of CO2 from oxy-fuel combustion into dolomitic reservoir (depth of about 4,200 m) started in January 2010, lasted for 2 years, and is followed by a 2-year observation period. The well and site abandonment will follow this observation period.
• Montmiral (France) – Natural CO2 (98.5%) accumulation at about 2,400 m depth. The accumulation has been exploited for industrial purposes since 1990, now in the process of closure. No closure procedure standards exist for CO2 natural accumulation sites.
• Frio (Texas, USA) – Onshore saline aquifer at a depth of 1,500 m. Total injected volume 1,600 tonnes. Site officially closed; post closure monitoring data available.
• Wallula (Washington State, USA) – Embedded in the activities of the Big Sky regional partnership, injection of 1,000 tonnes of CO2 in a 1,200 m deep basalt formation started in summer 2013. CO2 will be supplied from a paper and pulp mill. Two years after closure, a new well will be drilled to study mineralisation and compare the actual CO2 front with predictions.
• Nagaoka (Japan) – Onshore saline aquifer located at the Minami-Nagaoka gas field, 200 km north of Tokyo. Supercritical CO2 was injected at a rate of 20-40 tonnes per day into a 12m thick permeable zone, which lies about 1,100 m below the ground surface and more than 3,000 m above the gas reservoir. Injection started on 7 July 2003 and ended on 11 January 2005 with a total injected amount of 10,400 tonnes of CO2.
• Otway (Australia) – The CO2CRC Otway Project is a large-scale demonstration of storage in a depleted gas field. Injection was paused at this site in September 2009 after the injection of 65,445 tonnes of an 80:40 CO2:CH4 mixture. The project is intensively monitored by seismic, geochemical, atmospheric, groundwater and soil gas sampling methods. Future plans include an injection in a formation above the gas reservoir, to investigate residual trapping mechanisms.
Project Context and Objectives:
Over three years, CO2CARE focused on site closure and preparation for transfer of liability of a CO2 storage project in order to assist regulatory authorities and stakeholders in implementing the EU Directive 2009/31/EC on CO2 Geological Storage. The main objectives of the project were closely linked to the three high-level requirements of the EU Directive for the ultimate transfer of responsibility of a CO2 storage site to the Competent Authority typically 20 to 30 years after site closure. These high-level requirements are:
• Absence of any detectable leakage.
• Conformity of actual behaviour of the injected CO2 with the modelled behaviour.
• The storage site is evolving towards a situation of long-term stability.
In particular, the project mainly focused on three key areas:
• Well abandonment and long-term integrity.
• Reservoir management and prediction from closure to the long-term.
• Risk management methodologies for long-term safety.
CO2CARE achieved its project objectives via integrated laboratory research, field experiments and state-of-the-art numerical modelling, supported by literature reviews and data from a portfolio of real storage sites that covers a wide range of geological and geographical settings. This portfolio included 9 key injection sites in Europe and worldwide: (1) Ketzin, Germany; (2) Sleipner, Norway; (3) K12-B, The Netherlands; (4) Rousse, France; (5) Montmiral, France; (6) Frio, USA; (7) Wallula, USA; (8) Nagaoka, Japan and (9) Otway, Australia. Key research findings from the CO2CARE project, compliant with the EU storage regulation, were distilled in form of Best Practice Guidelines and disseminated to key stakeholders. Main results of so-called “dry runs” for the storage sites Ketzin, Sleipner and K12-B, which were discussed with European regulators, also contributed to the guidelines. These dry-run documents are a central outcome of the project and provide a template for site abandonment and transfer of responsibility, based on real storage sites and useful for both site operators and regulators. In the first instance the documents could be used at the site licencing stage to develop initial site closure plans and to inform early discussions and negotiations between the interested parties. Subsequently they could be used to inform the developing dialogue between site operator and regulator as the closure and post-closure phases are reached. The Best Practice Guidelines document will help establishing effective protocols for site abandonment. They provide technical guidance on how best to meet the three key requirements of the Storage Directive and also wider guidelines on post-closure wellbore and reservoir management and a robust protocol for risk management. In addition the Best Practice Guidelines contain recommendations directly relevant for the forthcoming review of the EU Storage Directive.
Current practices and regulatory framework
According to the European and international regulations, the liability for the storage site can be transferred to the licensing authority/government once the safety and conformity of monitoring with model predictions has been demonstrated. In the EU the CO2 storage Directive 2009/31/EC set out the regulatory regime and guidance for permitting CO2 storage. Around the world, relevant bills and regulations have been introduced in recent years too. In addition, regulations originating from the oil and gas sector concerning well abandonment are also relevant to CO2 storage well abandonment. To demonstrate the safety of the injected CO2 all regulations require a combination of monitoring, modelling and risk assessment tasks. Although there is large variation in the specific requirements it is standard to require approval for these tasks as part of a plan submitted to the authority in charge. To demonstrate the safety of the injected CO2, the results of monitoring, modelling and risk assessment, regulations require demonstration of no leakage, conformity with modelling predictions and that the site is evolving towards long term stability. Some regulations contain additional requirements including demonstrating no environmental problems, that the plume will not encounter any leakage pathways and demonstrating well integrity. There is a variation in the time period over which safety must be shown in different regulations, and an optimum time period is considered flexible.
Particularly in relation to well abandonment, it is recognised that there are different methods, materials and tests that could be used and most CO2 storage regulations do not specify techniques to be followed or standards to be met. Specific details on plugging are provided by regulations on the abandonment of hydrocarbon wells and sometimes other injection wells and these provide the best available guidance for CO2 storage well abandonment, although they may require updating to deal with CO2 injection specific issues.
Regulations typically contain a provision for transfer of liability once safety (CO2 and well plugging) has been demonstrated. The EU Directive 2009/31/EC requires further monitoring after liability transfer as a back-up measure, while other regulations (e.g. EPA UIC) do not. The IEA model regulatory framework contains a clause that the operator should also provide suggestions for the monitoring to be conducted after liability transfer.
Site abandonment of oil and gas fields is described and defined as the activity of the operator to close and leave a site according to safety and environmental requirements. It can generally be divided into two main activities, i) the abandonment of the wellbores drilled during operation, including plugging of wells, and ii) the removal of surface installations (e.g. well equipment, production tanks and associated installations) and surface remediation.
In site closure operations in relevant industries, mainly in the oil and gas industry, well abandonment and ensuring long-term integrity of wellbores are considered very important in terms of secure geological storage of reactive substances. These industries provide important information on main issues in current well abandonment procedures in CO2 environments and particularly in geological CO2 storage. Additionally, subsequent alternative utilisation of oil and gas fields is not generally considered at the time of abandonment and this might lead to important issues when a field is being considered for subsequent geological storage of CO2.
Well abandonment
The long term processes at the wellbore have been studied to evaluate the risk of leakage through and/or along the wellbore at different spatial scales. The elastic stress changes applied on the wellbore system during abandonment are generally minimal. Laboratory experiments were carried out on a custom designed CO2 wellbore test set up consisting of full scale wellbore casing, cement sheath (made of industry standard API RP 10B Grade G cement mix) and a stainless steel ring to represent host reservoir stiffness. A microannulus was generated between the well casing and cement sheath interface to represent a damaged well casing scenario. CO2 flow through experiments were carried out for three representative CO2 injection site scenarios (scenario 1. Sleipner type: temperature: 40 °C, salinity: 3.5 %, average reservoir pressure: 10 MPa; scenario 2. Ketzin type: temperature: 34 °C, salinity: 25 %, average reservoir pressure: 7.3 MPa; scenario 3. North Sea deep reservoir type: temperature: 92 °C, salinity: 12.5 %, average reservoir pressure: 35 MPa). It was observed that the rate of flow of CO2 (at constant inlet pressure) through the microannulus progressively reduces for shallow reservoir setting (scenarios 1 and 2 above) indicating self-sealing behaviour of the microannulus but remained unchanged for North Sea deep reservoir type setting. The former is attributed to carbonation reactions between cement and pore water resulting in leaching of species which in turn reduced the permeability. The latter is attributed to the lack of such reactions at high temperatures possibly due to the absence of pore water. The time dependent permeability change of the microannulus between the well casing and cement sheath under constant flux of CO2 derived was used as an input to reservoir simulations carried out to assess field scenarios on possible leakage of CO2 stored in a reservoir. It is inferred from the simulations that the self-healing behaviour of the microannulus enables containment of CO2 indicated by the constant volume of mobile CO2 and increasing volume of dissolved and trapped CO2.
Influence of impurities on geochemical reactions in the wellbore system
Batch experiments at near reservoir temperature and pressure (30°C, 8 MPa [80 bar]) investigated the potential geochemical reactions between CO2, synthetic pore-waters with low levels of either SO2, or NOx, or H2S impurities and samples of either Utsira caprock, 13Cr80 alloy borehole steel, or ‘Class G’ cement. The caprock experiments showed that the addition of the impurities caused some increased reaction in comparison with previous experiments using only pure CO2 and N2. In general, the N2-pressurised experiments showed little reaction, and the final fluid chemistry was similar to that of the starting fluid. The CO2-pressurised experiments appear to have attained approximate steady-state concentrations for most dissolved species after about 2 month. The main solid phase reaction observed upon addition of CO2 was carbonate mineral dissolution, in particular calcite.
In the cement samples although there was extensive reaction of the samples, the reactions did not result in the wholesale disintegration of the borehole cements used. In addition, the presence of impurities, other than the formation of gypsum (when S present), did not seem to significantly affect the nature of the reactions. Cl concentrations were found to vary in most of the cement samples, though the extent varied with the differing impurities. Qualitative X-ray map profiles were run across the samples to investigate the trends in the Cl concentration. The pattern of Cl enrichment and depletion illustrates the multi-reaction process that has occurred in the cement due to fluid infiltration. Initially Cl diffused into the cement ahead of the CO2 and reacted with the hydrated CSH and Ca(OH)2, to form a calcium chloride bearing phases. The slower diffusing CO2 then reacts, as HCO3-, with the newly precipitated calcium chloride phase to form secondary carbonate, which then liberates Cl. The liberated Cl then further diffuses deeper into the cement. Sodium follows an inverse pattern to that observed for Cl, with the area preceding the front being enriched in Na and depleted in Cl.
The addition, impurities in the experiments with borehole steel caused some increased reaction in comparison with previous experiments using only pure CO2 and N2. The steel in the CO2-pressurised experiments appears to have reacted relatively rapidly, allowing a concentration of dissolved reduced Fe to build up in solution, the presence of both CO2 and ‘S’ caused greatly enhanced steel dissolution in contact with the synthetic Utsira pore-water.
Overall, for all the materials studied here, the addition of impurities did cause some enhanced reaction but it was the presence of CO2 (with and without impurities) that had most impact on the reactions of the caprock and the borehole infrastructure.
Post-Closure Reservoir Management
A multi-disciplinary approach has been applied in order to provide a comprehensive understanding of the processes related to CO2 storage, and to identify and further develop monitoring methods for the post injection phase. The process understanding has been supported by designing dedicated laboratory experiments focussing on trapping mechanisms, such as dissolution and capillary trapping. Here, significant technical progress has been made, but still a large gap remains between a model and real geological settings. Numerical simulations of trapping mechanisms and of deterioration processes (such as, e.g. fracture growth due to chemical processes) have been performed. Long-term reservoir simulations for the Sleipner and Ketzin sites have provided insight into the contribution of different trapping mechanisms to the long term containment of CO2 in the reservoir. Reservoir simulations and monitoring data of the Sleipner and Ketzin site have been compared for an investigation of the conformity between simulations and monitoring results. A workflow applying performance criteria has been established showing that conformity between simulations and monitoring data enhances with increasing knowledge about the reservoir but discrepancies remain. Therefore, particular attention must be paid to establish realistic end-member model scenarios. The end-member scenarios must ensure a safe operation of a storage site and consider remaining uncertainties on the reservoir properties. At Ketzin, geoelectric and seismic surveys were performed with a special focus on long-term observation. Several types of innovative seismic sources were applied which can be installed permanently and which generate highly repeatable and high-frequency signals suitable to resolve small changes in the monitored rock volume.
Understanding of the long-term evolution of trapping mechanisms
Dissolution of dense-phase CO2 in formation water, followed by convective sinking is a key stabilization process in aquifers, acting on timescales from decades to tens of millennia. The onset of dissolution and sinking is poorly-understood and CO2CARE focussed on examining these crucial early processes by means of laboratory experiments and numerical models. Study of upward, vertical migration of CO2 gas through the porous medium revealed several interesting features:
• Initial entry of the gas into the porous medium formed several small ‘leaders’, only one of which eventually became dominant. The location of these was possibly controlled by minor differences in pore geometry and a slight reduction in capillary entry pressure. Thus even very small heterogeneities could play an important role in controlling gas flow into a porous medium.
• Once the gas entered the porous medium, it grew into an irregularly-shaped, elongate feature. Eventually this separated from the base of the cell and moved upwards through the porous medium as discrete, buoyant ‘packets’ of gas.
• The ‘packets’ of gas did not migrate perfectly vertically, but appeared to migrate randomly within a conical region (about ±12-15° from vertical) above the initial point of gas entry into the porous medium. This movement caused localised circulation of water around the ‘packets’ of gas as they rose.
• On reaching the upper surface of the cell the first part of a ‘packet’ of gas to make contact with the overlying free solution became the drainage point of all that particular volume of gas. This produced a focussed string of bubbles within the overlying free solution. As each packet was of finite size and took a different path, then these bubble releases were periodic, and were released at different locations. Gas movement through the entire cell went from continuous where it entered the cell, to episodic where it escaped the cell.
• Migrating larger ‘packets’ of gas through the porous medium resulted in smaller, irregularly-shaped bubbles of residual gas being left behind. The distribution of these changed with each passage of a larger ‘packet’ of gas; however their average distribution just within the migration cone appeared fairly constant, as did their average abundance (approximately 10% of the area of the migration cone).
Time-lapse movies of downward migrating plumes of (coloured) CO2-rich water through the porous medium revealed several important processes:
• Initial dissolution of CO2 into the water created a thin boundary layer where diffusion dominated transport.
• CO2 dissolution increased the density of the water, resulting in gravitational instabilities and the formation of small density plumes.
• As the plumes grew, they increased in length relative to width. They also increased in complexity, developing increasing numbers of lobes with increasing size.
• As the plumes grew they decreased in number through a process of localised lateral migration and assimilation.
• Larger plumes also migrated laterally across the cell. It is unclear however, if this was a consequence of cell-wide convection (i.e. an experimental artefact caused by the finite size of the cell), or a product of the system ‘flipping’ between different stable states of plume arrangement.
• As the plumes descended within the porous medium their outer edges became progressively more diffuse as they mixed with the CO2-poor water. They also appeared to slow down as they approached the base of the cell. This may have been controlled by limitations imposed by the size of the cell, but later in the experiment may also have been a consequence of increasing CO2 concentrations at depth, and hence decreasing density contrast between the plumes and the surrounding water.
Most probably, the instabilities are triggered by geological or physical heterogeneity. In real reservoirs, physical and chemical properties and overall timescales are significantly different from those in the laboratory, but there is no reason to suppose that our conclusions are not generally applicable and support the contention that the experiments support the validity of longer-term reservoir-scale flow simulations.
An integrated assessment of CO2 trapping mechanisms has been performed for the Stuttgart Formation at the Ketzin pilot site for CO2 storage by numerical simulations using different model coupling schemes. Long-term dynamic multi-phase flow simulations established the basis for all model couplings across all time scales. Pressure changes resulting in geomechanical effects are largest during the operational phase, whereas temperature related effects at the Ketzin pilot site are negligible due to injection of CO2 pre-heated to reservoir temperature. Geochemical reactions are governed by kinetics, and therefore long-term stabilization is determined by chemical processes. Chemical processes responsible for mineral trapping are expected to mainly occur during long-term stabilization at the Ketzin pilot site. Hence, assessment of long-term trapping mechanisms carried out for the Ketzin pilot site was extended integrating the geochemical effects generated by the migration of dissolved CO2. A cut-off value of 0.2 mol CO2/kg H2O results in a conservative mineral trapping scenario reaching 11.7 % after 16,000 years of simulation while a cut-off value of 0.1 mol CO2/kg H2O results in the optimistic mineral trapping scenario 30.9 % after 16,000 years. Dynamic multi-phase flow simulations indicate that only 0.2 % of the CO2 injected from June 2008 to August 2013 (about 67,270 t CO2 in total) is in a gaseous state, but residually trapped after 16,000 years of simulation. Depending on the cut-off value for the dissolved CO2 in our coupled geochemical simulations, CO2 dissolution is the dominating trapping mechanism that accounts for 68.9 % to 88.1 %, respectively. In both nonreactive and reactive simulations, no gas phase is present anymore after around 4500 simulation years, a value which is probably much underestimated due to the effect of a coarse grid used in the simulations; in case of non-reactive simulation the dissolved CO2 spreads significantly farther than in the reactive one after 10,000 years, however only in tiny concentrations.
Assessment of leakage detection thresholds at Ketzin and Sleipner
4D time-lapse seismic techniques have proved highly successful at monitoring CO2 in sandstone reservoirs, where the elastic properties of the rock are particularly suited to imaging fluid replacement (high porosity and relatively low rigidity). Leakage detection in shale dominated caprock sequences presents a greater imaging challenge. CO2 accumulations will potentially be highly localised in porous zones within a low porosity and permeability matrix. Synthetic seismic modelling has been used to investigate the potential detection threshold (in the caprock) of existing time-lapse seismic surveys acquired as part of the Sleipner and Ketzin CO2 storage monitoring programmes. A key factor in the ability of time-lapse data to detect small time-dependent changes is the degree to which successive datasets can be accurately repeated. Perfect repeatability would produce a noise-free difference dataset capable of detecting very small time-lapse changes. In practice, repeatability is far from perfect - difference datasets suffer from a variable amount of repeatability error or noise which acts to obscure real changes in signal. A four stage workflow has been applied to quantify detectability of both the Sleipner and Ketzin time-lapse seismic surveys in the presence of repeatability noise.
• The repeatability and magnitude of coherent and ambient background noise was calculated for the field seismic data using the repeatability metrics NRMS and predictability.
• A rock physics model based on all available petrophysical data was developed and used to calculate the effects of fluid substitution on the seismic response.
• A baseline synthetic seismic survey was built using the available petrophysical data.
• A series of synthetic seismic repeat surveys incorporating a body of (CO2) gas of known volume and saturation were computed using a finite difference code. An appropriate level of noise was added to the data based on the calculated repeatability metrics.
The repeatability metrics NRMS and predictability have been computed for the 2006 3D seismic monitor survey at Sleipner (2006 time-lapse processing) across a time window spanning 500 to 850 milliseconds. This window spans most of the Pleistocene Nordland Shale sequence, which forms the caprock to the storage reservoir at Sleipner. The mean NRMS is 54% and the mean predictability ~97%. The signal-to-noise ratio was also computed based on power spectra of the autocorrelation (signal + noise) and local cross-correlation (signal only).
Traditional Gassmann fluid substitution is not valid in shale because it assumes perfect communication of fluid in the pore space. The soft porosity model has been designed as an alternative to Gassmann theory in shale-rich sequences. The model requires a complete suite of petrophysical log data (VP, VS, Bulk density) together with laboratory analyses of matrix mineralogy and porosity. This information is only available for a small section of the lower Pliocene shale sequence in well NO15/9A-11. Consequently, the soft porosity model has been computed and compared to the results of traditional Gassmann fluid substitution using various fluid mixing schemes and rock physics parameters. The data suggest that traditional Gassmann fluid substitution with a patchy mixing scheme based on the Voigt average provides an acceptable approximation to the soft porosity model in the Nordland Shale. The synthetic seismic data have all been computed from petrophysical log data acquired for well NO15/09-18. The baseline survey thus comprises a series of planer reflectors corresponding to acoustic impedance contrasts observed at the well. Band-limited Gaussian noise was added to the baseline survey to give a comparable signal-to-noise ratio to the real data. A series of synthetic models were produced for specific wedge geometries and “image matrices” composed of seismic difference sections (repeat survey with a CO2 filled wedge minus the baseline survey) were calculated. The data suggests that very small accumulations of CO2 require gas saturations > 50% to be clearly imaged on seismic difference sections. At low saturations (<= 10%) it is very hard to identify amplitude changes induced by the CO2. At saturations in excess of 50% even small accumulations, occupying pore volumes of as little as 10 m3 could potentially be identified by careful seismic interpretation.
Time lapse seismic data acquired at the Ketzin site show a high level of both coherent and random repeatability noise. This could reflect a poorly designed time-lapse processing flow, but is likely due to low fold in regions where there is significant surface infrastructure. The detectability modelling carried out as part of this study focussed on a Jurassic Sandstone aquifer system, located above the main Triassic storage reservoir. This aquifer formed the main reservoir for a gas storage operation which was decommissioned in 2000. Any CO2 leaking from the Stuttgart formation will likely undergo secondary trapping in this layer. The repeatability metrics NRMS and predictability have been computed for the 2009 3D seismic monitor survey at Ketzin (2009 time-lapse processing) across a time window in the overburden spanning the Jurassic sandstone aquifer. The mean NRMS is ~30% and the mean predictability ~95%. A series of synthetic models were produced for specific wedge geometries and “image matrices” composed of seismic difference sections (repeat survey with a CO2 filled wedge minus the baseline survey) were calculated. The data suggests that small accumulations of CO2 (< 2 m thick) require gas saturations > 70% to be clearly imaged on seismic difference sections. At low saturations (<= 50%) it is very hard to identify amplitude changes induced by the CO2 above the repeatability noise. If the maximum thickness of the wedge is increased to 5 m, the CO2 is clearly imaged at saturations >= 30%. At patchy saturations below 20%, it is unlikely that detailed seismic interpretation will identify accumulations of CO2 above the background repeatability noise.
Pressure gradient reversal and injection of polymer-gel solutions as remediation measures
For the case of leakage occurring on a storage site, remediation techniques were investigated by laboratory and simulation studies. They consist of pressure gradient reversal (PGR) and the injection of polymer gels.
The effectiveness of the pressure gradient reversal (PGR) method as a potential remediation technique for CO2 leakage from deep saline aquifers was investigated using a generic 3D reservoir/caprock model. A hypothetical CO2 storage operation involving CO2 injection at 1 Mt/year for up to 30 years down-dip of a structure high in the model domain was considered. Two leakage locations in the transient source region, at a distance of 200m and 1200m, respectively, to the CO2 injector were selected for conducting above-zone brine injection simulations. The brine injection simulation results indicate that the performance of PRG is strongly affected by how early leakage is detected from the start of injection (time-to-detection), which in turns is controlled by the detection threshold, leakage pathway permeability and the distance to the injection well.
Crosslinked hydrolysed polymer-gel injection is used in petroleum industry to improve conformity of fluid flow in the reservoir, remediate leakage around wells and also used in conjunction with the prospect of enhanced oil recovery at various temperature and pressure conditions. To test whether this technology is also suitable for CO2 storage and to study the possibility of mitigating undesired flow of CO2 in host reservoir through the use of polymer gel solutions has been studied through systematic laboratory tests on characterisation of suitable polymer gel solution both as bulk material and its characteristics as fluid in porous medium. The inferences and results obtained from laboratory experiments were utilised to analyse possible remediation scenarios through reservoir simulations performed using a generic reservoir model. One of the important characteristics of polymer gel solution is the gelation time: which is the time lag between the addition of cross linker and the formation of a solid gel, in between which the polymer gel solution retains its rheological characteristics. This parameter in conjunction with reservoir permeability, temperature and pressure, dictates the area around polymer injection well that can be remediated.
In the laboratory, polymer was injected into sandstone core samples of known permeability and was allowed to gelify in-situ. Pure CO2 was injected into the core sample after polymer gelified and it was observed that the permeability of core sample to CO2 was reduced and polymer gel was stable (did not degrade over time). This reduction in permeability was implemented in a representative CO2 leakage reservoir scenario, wherein the host CO2 was simulated to leak into an overlying formation through a fracture in the caprock. The simulation illustrates the effectiveness of polymer gel remediation treatment, wherein the polymer gel was successful in mitigating the leak and the CO2 that leaked prior to remediation treatment has been dissolved into the formation brine and remained immobile.
Study of conformity between monitored and simulated site behaviour; assessment of short term performance and assessment of long-term stability of storage
Demonstrating conformance is technically challenging because a perfect match between observed and modelled behaviour, that is also demonstrably unique, is likely to be impossible to achieve. CO2CARE research therefore focussed on three related issues:
• To show that predictive modelling capability increases systematically with time as monitoring data is progressively acquired. This indicates that storage processes are well understood.
• To show that as more monitoring data is acquired through time, uncertainties progressively reduce, but focus must still be maintained on the less likely ‘end-member’ model scenarios to avoid the possibility of unexpected or divergent future outcomes.
• To show that, at site abandonment, predictive models calibrated by monitoring data can reduce the uncertainty envelope sufficiently for unexpected or divergent future outcomes to be ruled out.
In the case of the commercial-scale Sleipner project, multiple 2D (axisymmetric) and 3D reservoir simulations were compared to results of 3D time-lapse seismics. Research focussed on predicting and measuring the progressive development of the CO2 plume. Conformance testing was based on a number of performance criteria:
• Plume footprint area.
• Maximum lateral migration distance of CO2 from the injection point.
• Area of CO2 accumulation trapped at top reservoir.
• Volume of CO2 accumulation trapped at top reservoir.
• Area of all CO2 layers summed.
• Spreading co-efficient (storage efficiency).
Initial predictions in 1996, using only baseline observations, with no plume monitoring data, had a high degree of uncertainty arising mostly from uncertainty in how the CO2 would be trapped within the reservoir. Possibilities ranged from a single layer at the reservoir top or several layers at different levels within the reservoir. As a consequence, performance measures, such as plume area footprint, had high range of possibilities. Predictive scenarios for 2001 and 2006 used the monitoring data that was available at those times. This principally comprised a suite of repeat 3D seismic surveys, acquired in 1999, 2001, 2004 and 2006, which confirmed that CO2 was trapped at multiple levels within the reservoir, and provided additional constraints such as arrival time of the CO2 at the reservoir top, estimation of CO2 flux into the topmost layer and individual layer extents. Improved reservoir temperature data also became available as the injection progressed. The predictive scenarios for 2001 and 2006 included only multi-layer plumes. In addition the 2006 predictions had improved reservoir temperature constraints, so end-member possibilities progressively became much closer together. As a result, predicted ranges for the performance measures are much reduced and the likelihood of the end-members leading to unexpected or divergent future outcomes is much reduced. A key performance measure that cannot be accurately assessed by the axisymmetric modelling is the lateral migration of the topmost layer of the plume, directly beneath the topseal (whose topography is known from the baseline data). Initial 3D models based on 1996 data had a very wide spread, but uncertainty reduced markedly as the time-lapse datasets became available. It is clear however that even by 2006, a perfect prediction of the 2008 plume was not easily obtainable. This is due to continued (though much reduced) geological uncertainty and also to likely limitations in the predictive model itself. Nevertheless, it is also clear that the basic process of layer development (buoyancy-driven migration by fill-spill beneath the topseal topography) is well understood. As more monitoring data is acquired uncertainty will reduce still further and the likelihood of unexpected divergent future outcomes is very small. Thus, as more seismic data became available, predictive models could be matched more accurately to the observations and became more reliable predictors of future performance.
Risk Management
The assessment and mitigation of the risks related to the geological storage of CO2 (e.g. loss of containment, potential direct effects of CO2 leakage on the biosphere) and the containment of CO2 in the storage complex particularly during the stages of site closure and the post-closure have been addressed according to the prescriptions in the EU Directive.
Plan for risk management prior to and after transfer of responsibility
A risk management plan was established covering the requirements of the EC Directive on the geological storage of carbon dioxide (DIRECTIVE 2009/31/EC) and the OSPAR Guidelines. The timeline considered by this plan encompasses the final injection period within the operational phase as well as the post-closure/pre-transfer and post-transfer period of a storage project (Phase 4-6 according to EC GD3). In order to provide a well-structured procedure for risk management within these project phases, 17 Site Closure Milestones (SCM) are introduced, which are implemented into the different project phases The milestones are closely linked to the requirements of the EC Storage Directive and describe key actions or key moments in time during site closure and transfer. The proposed SCMs (not to be confused with the milestones mentioned in EC GD3) lie in operational Phase 4 and post closure/pre-transfer Phase 5, which are, for a better understanding, split into three sub-phases based on the terms used in the EC Guidance Document:
• final operational sub-phase (part of Phase 4, including site closure)
• post-closure sub-phase (first phase of Phase 5)
• pre-transfer sub-phase (second phase of Phase 5, including transfer of the site)
Development of criteria for transfer of responsibility
A report on criteria for transfer of responsibility from the site owner to the state has been provided. Decision making in the post-operational phase by the operator is targeted on minimizing risks demonstrating that the site behaviour is understood and that conditions required to transfer the responsibility for the site to a Competent Authority (CA) according to the EC Storage Directive (Directive 2009/31/EC, EC 2009) have been met. To be able to demonstrate the fulfilment of the three high-level criteria of the EC further more concrete low-level criteria and milestones applicable on an operational level are required. All of the three fundamental criteria mentioned above rely on monitoring data and models. Monitoring and/or model data are the key outcomes to demonstrate the understanding of site behaviour. By developing a risk management (RM) plan a milestone chart specifically relating to each of the three fundamental criteria has been shaped (called Site Closure Milestones “SCM”). The criteria derived from the risk management (RM) plan are termed “R-type” criteria. In order to assess whether the R-type criteria are met, a traffic light system was set up to support decision making during and after closure of the storage site (cessation of injection). The traffic light system establishes whether or not the monitoring and modelling data are in compliance. The criteria derived from this traffic light system are “T-type” criteria. The system consists of a decision workflow determining the condition (state) by which the abandoned site in question is characterized. Three different states are foreseen:
• Status green: MMO of all parameters are within tolerance, i.e. the site is in a regular and expected condition.
• Status orange: MMO of one or more parameters are off tolerance; the operator has to prove whether the models in question have to be recalibrated or irregular site behaviour is present; this involves a discussion between operator, experts and the CA.
• Status red: Irregular site behaviour is present; additional monitoring, counter measures, and mitigation measures have to be applied.
The presence and practical implementation of required (mandatory) site-specific monitoring parameters subject to modelling and monitoring (Criterion T1) and a prioritization of models - mandatory and optional – (Criterion T2) are two important criteria for decision making with respect to risk management and the transfer of responsibility.
The practicability of the proposed traffic light workflow has been evaluated on the K12-B site to test how the traffic light system can support the closure of a CO2 storage site. The benefit of the proposed traffic light workflow could be demonstrated and it is shown how the traffic light system can support the decision making in CO2 storage site responsibility transfer and abandonment by presenting an example of how to deal with an unpredicted behaviour or irregularity of the storage site (using bottom-hole pressure and monitoring data). The approach to define criteria leading to the transfer of responsibility for the site revealed that, although based upon a generic framework, the definition of such criteria is highly site dependent. In particular, the definition of tolerable model-monitoring deviations and required accuracies/precisions of models and monitoring techniques is ambiguous and requires thorough considerations by the operator of the site and the Competent Authority. Although the traffic light system was designed for treating irregularities in the final stages of the storage lifetime it is well applicable to the treatment of irregularities in all stages of a storage project.
Method for determining full system risk and uncertainty including optimisation
For safe storage of anthropogenic CO2 in the subsurface, four trapping mechanisms have been identified, acting on time-scales ranging from months to tens of thousands of years: 1) structural/stratigraphic trapping; 2) residual trapping; 3) dissolution trapping; and 4) mineral trapping. This study mainly focused on structural/stratigraphic trapping of injected CO2 in a storage reservoir, with special reference to the post-closure and post-transfer phases of a storage project. It is recognised that mobile CO2 phase in the storage reservoir (source for potential leakage) has a tendency to follow the general topography and migrate up dip with time, further beyond the area where the plume is located at the time of site closure. If one considers some of the CO2CARE project field sites, for example Ketzin, CO2 has been forecasted to migrate up an anticline towards the North, until reaching a vertically sealing fault, after CO2 injection is ceased in 2013. With respect to the potential leakage risk in the post-closure phase and beyond at the Ketzin site, the following observations may be made.
• The risk of leakage through the injection and monitoring wells is likely to be low, simply due to the diminishing source (mobile CO2) in the near-wellbore region over time.
• The caprock/faults along the CO2 migration path pose a potential risk over short to medium term. Monitoring tools other than plume imaging may need to be considered to ensure containment.
Above-zone pressure monitoring, which allows the probing for leakage of a wider area beyond the present CO2 plume, has been employed at Ketzin. No noticeable pressure communication between the storage reservoir (630 m depth) and a shallower aquifer (~420 m depth) has been observed.
Assuming that CO2 injection was ceased in 2006 at Sleipner, post-injection plume migration has been simulated by BGS until 2034, showing CO2 plume extending further Northwards, and becoming largely stationary by 2034. Unlike the Ketzin site, above-zone pressure monitoring cannot to be applied effectively at Sleipner due to the very low overpressure (~0.1 bar) generated in the storage reservoir by CO2 injection.
A risk assessment methodology, which considers the dynamic CO2 plume migration behaviour in the storage reservoir, with special reference to the post-closure period, was developed. The methodology involves:
• First step: CO2 plume evolution and migration, in the long-term over time, as a dynamic source for potential leakage out of the primary storage reservoir is evaluated by taking into account reservoir heterogeneity and data uncertainty.
• The second step considers the characteristics of specific and potential leakage pathways (caprock fractures, coarse grained lobes; wells and faults) to account for leakage likelihood and leakage rate.
• The last step evaluates, through simulations and uncertainty assessment, the potential CO2 leakage through the identified leakage pathways for different scenarios regarding the leakage path permeability and a detection threshold.
In order to demonstrate the implementation of the methodology described above, a generic reservoir/caprock model was set up and used to simulate CO2 injection of 1 Mt/year for up to 30 years in a saline aquifer and evaluate the potential leakage risk. A number of leakage scenarios were also set up and assessed in order to present and describe different elements of the methodology developed. Reservoir simulation of CO2 injection at a rate of 1 Mt/year for 30 years from 2012 was carried out to evaluate the plume migration behaviour during injection and after the termination of injection. A pore volume multiplier of 100 was used during simulations to represent the connected pore volume beyond the model domain. It was found that the plume largely stabilised at about 120 years from the start of injection under both sealing/non-sealing fault conditions. The influence of the state of the fault (sealing or non-sealing) along the anticline, after being approached by the advancing CO2 plume, on further migration of the plume is clearly observed. Based upon this plume migration behaviour, and the tendency to migrate up dip following the formation topography, the plume footprint may be broadly divided into:
• the transient region (where the free CO2 largely has a limited residence time), and
• the non-transient region (where the free CO2 residence is more or less stable).
This distinction has a direct bearing on the potential leakage profiles in the two regions. Time lapse 3D seismic surveys have been successfully applied to image CO2 plume and monitor its migration in the storage reservoir in both Sleipner and Ketzin sites. As well as demonstrating CO2 containment during the injection period, the seismic monitoring data are valuable for calibrating the reservoir model. Nevertheless, forecast of plume migration during the post-closure period using the calibrated reservoir model is still likely to be subject to varying degree of uncertainty due to lack of knowledge of the farfield porosity/permeability distribution. One way to address this uncertainty is to perform statistical analysis by generating a large number of realisations of (farfield) porosity/permeability distribution of equal probability, based upon our understanding of the storage site geology and available seismic, well and core data. To illustrate the impact of heterogeneous porosity/permeability distribution, in particular the presence of high porosity/permeability channels, on CO2 plume migration, stochastically generated channels were implemented in the generic model. To this end, the reservoir formation was divided into two major facies types, namely the floodplain and channel facies. Facies modelling was performed using the Petrel software package. A total of 10 realisations with different channel orientations were generated. For each realisation the two facies were then populated stochastically with porosity and permeability using the Sequential Gaussian Simulation method. For illustration, simulation of CO2 injection, at an injection rate of 1 Mt/year for 30 years, as in the base case in the absence of high permeability channels, was carried out for the 10 realisations generated as described above. Compared to the base case without channel structures, it is seen that CO2 tends to migrate preferentially along the least resistance paths, i.e. the channels. Nevertheless, the injected CO2 eventually reaches and accumulates at the top of the anticline after injection is ceased. Based on the analysis of the entire footprint of the CO2 plume for the 10 realisations, a plume footprint mean probability map, which gives a spatial overview of the likelihood of CO2 plume migration spanning both the injection and post-injection phases, may be obtained. For comparison, the extent of the plume for the base case (without the channels) is superimposed on the map. Dividing the CO2 plume footprint into a transient and a non-transient region also helps track the amount of free CO2 in each region – the source of potential leakage risk - over time. Free mobile CO2 in the transient region starts to decline when injection is terminated after 30 years, and thus represents a diminishing source for potential leakage within this region over the 120 years simulation period. On the other hand, free CO2 accumulated in the non-transient region is likely to remain a source for potential leakage over hundreds of years after injection is terminated. Using the generic reservoir/caprock model, an attempt was made to compute and map potential leakage risk profiles, i.e. the total amount of leaked CO2 through the caprock, and the leakage time periods, at various locations in both the transient and non-transient regions. A total of 30 potential leakage areas were considered, each being represented by a leaky grid block. The leakage scenario considers one leaky block at one time iteratively. To simulate CO2 leakage, a leakage pathway is intentionally created by assigning a permeability of between 1 and 10 mD to the column of grid blocks (200m x 200m) in the caprock between the storage reservoir and the permeable layer above. During simulations, the cumulative leakage from the storage reservoir is monitored and injection is terminated when a pre-set leakage detection threshold/limit is exceeded. A detection threshold between 1,000 to 10,000 tonnes of CO2 was considered here. The time (year) it takes for the leakage to be detected during the simulation, i.e. time-to-leakage detection, referred to for simplicity as the time-to-detection (TTD), maybe computed and it is expected to vary spatially within the CO2 plume footprint. At each leakage location, the TTD depends on the combined effect of a detection threshold applied and the leakage pathway permeability assigned. The TTD values computed for a grid block permeability of 10 mD and a detection threshold of 10,000 tonnes are mapped on the plume footprint. As would be expected, it would generally take longer to detect CO2 leakage if the leakage location is further away from the injection well. Although CO2 injection is terminated once leakage is detected, leakage is continuously being monitored to obtain potential leakage profile, namely the total leakage duration and the cumulative CO2 leakage, at each leakage point. Simulations are stopped at year 2132, some 120 years from the start of injection at year 2012. The computed leakage profiles display very different trends in the two regions. Whereas the leaked CO2 mass reaches a plateau in the transient region, leakage is still continuing in the non-transient region in 2132. It is further noted that leakage in the transient region would continue for a further period of time and the total leakage duration is significantly longer than and positively correlates with the time-to-detection at each grid block. So far, the methodology focused on the migration behaviour of the free CO2 in a storage reservoir during and post CO2 injection periods, which is a dynamic source for any potential CO2 leakage, and demonstrated how this dynamic behaviour can affect the (unmitigated) leakage profiles in the presence of a leakage pathway along the plume migration path. The risk of CO2 leakage out of the storage formation depends on the existence of a permeable pathway in the caprock, capillary entry pressures for the pathway to be exceeded and a pressure differential (including buoyancy) sufficient to transport CO2 along this pathway. For CO2 containment risk assessment, three broad types of leakage pathways have been identified:
• caprock;
• injection/abandoned well;
• faults and fault zones.
Of the three leakage pathways, leakage through wells will lead to relatively localised leakage, and thus it may be argued that it will be possible to remedy such leakage without any significant volumes escaping out of the storage complex. Diffuse leakage through fractures or pore-space in the caprock, on the other hand, will generally be more difficult to detect and remedy. A probabilistic methodology for the existence of a percolating fracture network in a caprock/reservoir system has been developed and the geomechanical failure risk assessed, taking into account the uncertainty in the measurements and data available. Implementation of these methodologies has been demonstrated using the Otway data as part of the full system risk and uncertainty assessment methodological development. The main fracture data at the Otway site comes from borehole image logs in CRC-1 and a second well, CRC-2 drilled very close by, which terminates in the shallow Paaratte formation. Analysis of the FMI logs was carried out to extract information on fracture attributes (orientation, line density) in the reservoir and the caprock (Belfast mudstone and the overlying Skull Creek mudstone). Examination of core samples revealed only one open fracture, with most being closed and filled with carbonate. There was no information on the length of fractures. The orientation distribution of the Belfast and Skull Creek layers were taken from the data CO2CRC provided. The length distribution is not known, therefore a power-law length distribution was assumed, using a range in order to represent the full range of possible length distributions. Four scenarios of the ratio of open to closed fractures were used in fracture modelling: 0.05 0.1 0.2 and 1, though data indicates that a scenario with a lower aperture fraction is most realistic. The uncertainty in length distribution is the main cause of uncertainty in percolation probability for a given aperture scenario. By using a large number of different sets of fracture parameter values and modelling a large number of fracture network realisations for each set of values, the probability of a percolating pathways existing in the Belfast and Skull Creek layers were obtained. The results of percolation assessment indicate that it is very unlikely in all regions with an aperture fraction of 0.05 and 0.1. The findings of and knowledge gained from the studies, in conjunction with an understanding of the free CO2 behaviour in the storage reservoir, can help better evaluate CO2 leakage risk and associated uncertainty for the closure and post-closure phases for a specific storage site.
Uncertainty and risk assessment of the Ketzin Site
A methodology was applied for detailed uncertainty and risk assessment of the Ketzin pilot site during the post-closure period. The objective of the assessment is to quantify the uncertainty in the behaviour of the CO2 plume for the purpose of risk evaluation. The uncertainty in the plume behaviour is caused by the geological uncertainties, namely the distribution of the fluvial channels in the far-field region of the reservoir inside the Stuttgart Formation. Hence, multiple stochastic realisations were generated by honouring the data available in the near-field region at the injection well Ktzi 201 in order to represent equally likely static geological models. These models were subsequently implemented in the Eclipse 300 reservoir simulation software. The results obtained for the CO2 saturation distribution were summarised as probability maps that give an indication of the most likely locations in the reservoir containing mobile CO2 (assuming a cut-off of 5% for residual saturation) at different times for 500 years during the post-closure period. Furthermore, the probability maps were also used to quantify uncertainties in the parameters that characterise the plume behaviour, namely the plume arrival times, the plume residence times, and the maximum amount of mobile CO2 for the ‘transient’ and the ‘non-transient’ regions of the reservoir.
Best Practice and Regulatory compliance
Guidelines for regulatory compliance and Best Practice for storage site abandonment has been established. Key research findings from CO2CARE were distilled and integrated into site closure and abandonment protocols compliant with the EU storage regulation (the Storage Directive, OSPAR and the EU ETS). A set of hypothetical ‘dry-run’ applications for site closure and abandonment for the Sleipner, K12-B and Ketzin injection projects has been prepared. The dry-runs are hypothetical in the sense that they have assumed arbitrary dates for cessation of injection, but are otherwise realistic in that a full risk analysis was carried out for each site and scientific arguments presented to address the technical criteria for transfer of responsibility as required by the EU Storage Directive and national regulations.
The dry-run documents included:
• Description of the storage site background with geological setting and injection history.
• Risk assessments and risk management plans
• Detailed assessment of each of the three key criteria for transfer of responsibility:
o Zero detected leakage
o Conformance of modelled and observed site performance
o Evolution towards long-term stability
• Post-abandonment monitoring plans
Regulatory compliance via interactive stakeholder involvement
A Regulators Workshop for Site Abandonment Dry-runs - Best Practice for CO2 Storage and Transfer of Liabilities (Milestone 10) - was organised and the regulatory responses to a number of hypothetical but detailed closure cases for real CO2 storage sites aquired. The CO2CARE project participants presented the 'dry-run' case-studies for the transfer of responsibility for Sleipner (Norway), K12-B (The Netherlands) and Ketzin (Germany). The transfer cases are hypothetical in the sense that they have assumed arbitrary dates for cessation of injection, but are otherwise realistic in that a full risk analysis has been carried out for each site and scientific arguments were presented at the workshop to address the technical criteria for transfer of liability as required by the EU Storage Directive and national regulations. The role of the regulators in the workshop was to review and critically asses each of the case-studies, so that the CO2CARE project participants could develop best practice guidelines for site transfer. The outcome of the workshop was a series of key recommendations from the regulators, listed below:
Risk assessment process
• Acceptance of the Transfer Report (and Transfer) marks the end of a process that began with site selection and a permit application many years previously. Communication between operator and Regulator should have been continuous during that period, so the transfer report should contain no surprises. The transfer report is part of a process which should build mutual confidence between regulator and operator.
• Competent Authorities may wish to undertake their own simulations based on static models developed by the operators, to make an independent evaluation and consider the effectiveness of monitoring. Static geological models and numerical reservoir simulations run on in-house software platforms rather than commercially available platforms such as Petrel and Eclipse may not be acceptable for transfer purposes because the Competent Authority would not be able to run the models.
• The Risk Assessment methodology used in the update of the post-closure plan should be standardised and use a standard template, though each site is likely to throw up different risks so the assessments would depart from that point onwards. The standardised template should contain all relevant facts and issues. Risk assessment starts at the beginning of the life cycle of a CO2 storage site, implying that the standardisation of the methodology should start there too.
• Risk matrices should be made for all sites.
• Interactions with other uses and users, both currently and in the future should be included in considerations of risks.
• Only CAs can assess the risks that might arise from operation and closure of multiple assets within a formation or region.
• Long-term well integrity and caprock sealing should be considered. This should include documentation of the configuration and history of the well using operational information such as cement job evaluation, plug testing, final abandonment, and other issues.
Conditions of transfer
• The Transfer report has the function of a contract – and it needs to contain key messages for the public. The State is likely to produce a counterpart document as well.
• From an operators perspective, it may be necessary to include a statement in the storage permit along the lines of: “if the storage site performs as predicted, no leakage is detected by the monitoring technologies deployed according to the monitoring plan, the site is evolving towards a state of long term stability, and all the terms and conditions in the storage permit are met, then the Competent Authority will accept transfer of the site”. This would provide comfort to the investors and operator that the State would accept the site back once the storage operation had been successfully completed.
• In a discussion of liability associated with transfer it was pointed out that society must take or share the risk of storage if it wants the benefits of CCS.
• Are regulators obliged to take the site back if the site is conforming and operated as expected?
• The definition of the storage complex is very important because this encompasses the entire rock volume affected by the storage operation. Views of other stakeholders involved or potentially involved in this volume and area should be addressed in the transfer report.
• Is there anything to be learnt from the sale of oil and gas fields that can inform the CO2 storage site transfer process?
Managing uncertainty
• As predictions of the future site performance are based purely on forward modelling, more numerical reservoir simulation runs in which more model parameters are varied may be necessary. It may also be necessary to conduct modelling of other aspects of the site performance, e.g. geomechanical stability. Predictions based in single lines of evidence are likely to be insufficient.
• A far greater level of detail on how models were constructed and how they evolved throughout the storage site characterisation, construction and operation should be included. Furthermore, information on how such models take uncertainty into account would likely be necessary.
Post-transfer
• In the section on conformity of models and monitoring, at least three reasons for monitoring should be addressed: monitoring to improve knowledge of the site and its performance, early warning of potential departures from modelled performance allowing time to intervene, monitoring of the efficacy of corrective measures, technical and social baseline needs.
• Financial arrangements and a plan for post-transfer monitoring need to be made – it is too easy and not sufficient to say monitoring can stop; it will be needed.
• Liabilities for certain risks will remain with the operators post-transfer.
• The transfer report should contain recommendations of ‘light’ monitoring during the post-transfer period for the CA. Whilst not implying significant potential risks of leakage, such monitoring might be undertaken in areas or multiple injection and as part of a CA’s duty of care.
• As CO2 storage is a multi-generational operation, consideration should be given to what tools and data should be transferred to the CA. Models may need to be used post-transfer, if a significant irregularity or leakage was detected.
Key research results and the outcomes of the Regulators workshop and Dry-run documents were compiled into Best Practice Guidelines for storage site abandonment and transfer of responsibility. The guidelines are a set of pragmatic and workable generic procedures, suggested best practices and other recommendations and observations for the safe and sustainable closure of geological CO2 storage sites. CO2CARE has developed key learnings in the risk management, well abandonment and post-closure (post-injection) reservoir management:
Risk Management: We propose a set of site closure milestones to be fulfilled in chronological order before site closure and, subsequently, transfer of the storage site to the Competent Authority. We have also developed a set of high-level, risk based and technical criteria that can be used to determine whether each of the site closure milestones has been reached. We recommend that the Competent Authority and the storage site operator agree a priori on the specific conditions for deviations from the predicted site behaviour that will trigger corrective measures. These might be based on quantitative threshold values such as the measured difference, or offset, between predicted and measured performance measures. We also recommend that criteria for long term safety and security of the site should be based purely on technical considerations and should not be linked to prescriptive time spans.
Well abandonment: A key requirement for the long term safety and security of geological CO2 storage is that wells contacted by the stored carbon dioxide should not leak; it is clear that the quality of the original wellbore construction is the main factor in ensuring long-term wellbore integrity. Subsequent to this, it is necessary to assess whether wellbores will have been susceptible to additional damage or degradation during the CO2 injection operation. Longer-term prediction of wellbore performance remains challenging and depends on various types of predictive modelling and experimental or analogue information. Laboratory experiments conducted by CO2CARE exposed typical well casing, cement and rock materials to CO2 for periods up to several years. We found that for typical storage conditions, cement carbonation and steel corrosion reactions can cause porosity plugging in the rocks and wellbore annuli, tending to retard or prevent CO2 migration up, or alongside, the wellbore.
Post-closure reservoir management: Following cessation of CO2 injection the main aim is to demonstrate that the key regulatory requirements for transfer of storage site liability to the Competent Authority have been met. These are based around demonstrating understanding of reservoir processes and the ability to make robust predictions of future behaviour, and providing assurance against leakage. Migration of the CO2 plume and reservoir pressures are two key determinants of reservoir performance during the injection and immediate post-injection phases, and these have proved effective in verifying storage performance at the Sleipner, Ketzin and Rousse storage sites. 3D seismics and down-hole pressure measurements are proven technologies and have been key monitoring tools for reservoir management at these sites.
With regard to the three high-level criteria set by the EU Directive, CO2CARE developed few central findings and recommendations that should to be taken into consideration by the regulators when implementing and applying the Directive:
All leakage monitoring systems have a finite, site-specific, CO2 detection capability, so it is recommended that regulators use the term “no detectable leakage” in the context of whether a site is performing effectively in terms of both its greenhouse gas emissions mitigation function and health and safety requirements at the surface.
Demonstrating conformity between predictive models of reservoir performance and monitoring observations is technically challenging because a unique and perfect match is near-impossible to achieve. We recommend that regulators should set conformance criteria at realistic levels, focussing on progressive reduction of uncertainty with time and demonstration that the fundamental site-specific storage processes are understood.
Demonstrating that a site is moving towards a state of long-term stability is difficult due to the lack of long-term observational evidence from available storage projects. Predictive modelling is subject to significant uncertainty, so full use of additional analogue information is important to develop a logical case for site stabilization. Use should be made of monitoring data from sites already in the post-closure period (e.g. Nagaoka), experimental data and relevant geological analogues which demonstrate stabilization processes in similar circumstances and the time-scales on which they operate.
Project Results:
The research outcomes of the CO2CARE project will have a direct impact on European and overseas companies and institutions dealing with CO2 storage. Especially the close collaboration between scientific researchers, industry partners, and national authorities is promising. The distilled results of the project in the form of “Best Practice Guidelines” will play an important role in the CCS community worldwide. The Guidelines will not only support companies who are operating CO2 storage sites but also provide information for the regulators in the different countries in terms of the creation of specific requirements for the long term stability of CO2 storage complexes.
Progress beyond state-of-the-art
Geological storage of CO2 is a long-term process, the efficacy and safety of which needs to be assured in order for it to be accepted as a viable climate change mitigation option. The three main potential leakage paths for both the injection and post-closure periods are (a) abandoned wells and operational wells (b) the caprock through pre-existing or induced fractures or faults, and (c) undetected faults and fractures. For the purposes of post-closure risk assessment of CO2 storage, the EU Directive requires the site operators to investigate:
• the risk of fracturing the storage formation(s) and caprock,
• the risk of CO2 entry into the caprock,
• the risk of leakage from the storage site (through inadequately sealed wells),
• the rate of migration (in open-ended reservoirs),
• fracture sealing rates,
and establish remediation procedures for use in case of leakages or significant irregularities being observed. Pursuant to Articles 17 and 18, and Annex II of the Directive, the post closure criteria require the operators to utilise long-term monitoring data to revise the predictive models used and update the monitoring plans if necessary. This may also lead to an updated and a more advanced risk assessment at the time of site abandonment. The results by CO2CARE in this field notably enlarged the knowledge and progressed beyond state-of-the-art. By this, the results of CO2CARE will help to implement the European CO2 storage technology.
Wellbore integrity
Wellbores are recognised as one of the potential leakage pathways in geological storage of CO2. Important lessons regarding wellbore cement seal integrity can be learned from the oil industry, which has been conducting EOR for many years. However, while the oil industry is primarily concerned with the cement sheath integrity over the lifetime of a well (decades), geological storage of CO2 requires the consideration of a much longer time frame (hundreds to thousands of years). The primary concerns are that the Portland cements used to seal wellbore react readily with CO2, and the geomechanical/geochemical response of the wellbore to injection, abandonment and post-abandonment processes may compromise the integrity of the wellbore. Studies have shown that changes in downhole conditions can cause mechanical damage to the cemented annulus that may lead to a loss of zonal isolation. It is therefore important to have a firm understanding of the long-term behaviour of the complete mechanical system formed by the steel casing, the cement sheath, and the formation rocks. In terms of site abandonment, the crucial question is: under what downhole conditions may such an interface become progressively more transmissive? The initial width/quality and CO2-brine flux have been identified as the two key parameters controlling the long-term interface integrity. These important issues have been addressed through an integrated research programme encompassing laboratory testing, field installations and monitoring, as well as numerical simulations. Consistent with the above, research focus has progressively shifted to the complete system represented by the steel casing, the cement sheath, and the formation rocks. To take into account the long-term effect of CO2 on the cement sheath, the modelling approach has to consider the well on all its length anticipating the future behaviour of the well during the closure period and during the long-term post-closure one. However, previous studies suffer significant shortcomings in that the interface is synthetically created/manufactured and not necessarily representative of changes in real downhole conditions, particularly with respect to controlling the width and permeability of the interfaces created. CO2CARE performed newly designed laboratory experiments to overcome the disadvantages of current practice in this field and innovative well abandonment methodologies are introduced in the research programme. By these efforts, the results by CO2CARE will provide the operator with profound knowledge on wellbore integrity.
Reservoir and caprock integrity
The storage of CO2 in geological formations requires a caprock which forms a capillary seal to the injected CO2. Pressure depletion in a producing oil or gas reservoir, or CO2 injection into the reservoir or its subsequent cessation (at closure), modifies stress states such that newly created or existing fractures could open up, creating a connecting fracture network. In the context of post-injection (closure stage) risk assessment, the conductivity of induced fracture networks in the caprock around the wellbore have not been studied with uncertainty analysis so far. Furthermore, migration of CO2-rich fluid in porous rocks can cause mineral dissolution resulting in reduction of the rock strength. These processes are more intense in the rock joints and specially concentrated at joint extremities. The high dissolution ratio at these points will affect the rock toughness and can activate fracture propagation. The results by CO2CARE on the sub-critical crack propagation due to the geomechanical and geochemical processes and on the behaviour of storage integrity during the post-closure period will better facilitate long-term predictions on caprock and reservoir integrity.
Trapping mechanisms
Concerns and uncertainties at the post-closure and abandonment stage, particularly in terms of modelling, are mainly related to the longer-term storage processes which progressively immobilise the CO2 and lead to site stabilisation. These are primarily:
• Residual trapping by capillary forces.
• Solubility of CO2 in the brine leading to gravitational stabilization.
• Enhancement of dissolution by convection caused by brine density contrast.
• Geochemical fluid/mineral reactions leading to mineral fixing.
These mechanisms have been investigated mainly in isolation so far, but are recognized to interact. The dissolution process in a saline aquifer and the development of convection that enhances dissolution has been studied theoretically, but calibration data from pilot sites (e.g. Nagaoka) will be incorporated to support the numerical modelling. The contributions from each of these mechanisms are very much dependant on site specific geology and injection strategies. The results by CO2CARE integrating exhaustive experimental and modelling studies provide site-specific investigation of these mechanisms and their related impacts and thus will allow address these issues in the post-closure transfer process.
Monitoring methodologies
Seismic measurements (active methods and, to an increasing degree, passive monitoring) are a basic and important component in reservoir management. Repeated 2D or 3D seismic surveys serve for large scale monitoring of a reservoir in oil and gas production, town gas storage, and other applications such as CO2. While these surveys image structures and allow to infer petrophysical parameters from seismic attributes, the passive seismic monitoring aims at directly observing process related phenomena such as micro-seismicity or subtle but continuously occurring changes in the behaviour of seismic noise. Repeated 2D or 3D surveys are cost intensive and impose a strong impact on the local population and the landscape, especially onshore in populated areas, (e.g. by causing crop damage). Thus, this method is not an option for continuous monitoring with a high temporal resolution at a storage site in the abandonment phase. The monitoring concept studied within CO2CARE provides the framework for a cost efficient, (almost) non-invasive set of observations which offers a high temporal resolution. By this, one of the hurdles for implementation of CO2 storage will be overcome. Several systems exist which can be used as autonomous permanently deployed seismic sources for high resolution monitoring of a potentially problematic area of the reservoir but they have not yet been applied to CO2 storage, nor for comparison which is a prerequisite to assess their applicability to storage reservoir monitoring. At Ketzin, TNO has already deployed an array of buried continuously recording seismic receivers which forms an integral part of the seismic monitoring system tested.
An efficient application of monitoring methods during and after site abandonment is highly dependent on the best possible characterisation of the storage reservoir, based on data collected during the storage phase. These seismic and geoelectrical data allow a set-up up of a detailed structural model and to quantify the CO2 distribution within the reservoir.
While seismic measurements alone, even under very good conditions, are not able to provide a detailed quantification of the actual mass of CO2 the integrated interpretation of seismic and electromagnetic data is expected to enhance the potential for a quantitative interpretation due to the complementary sensitivity of both methods and the possibility to use cross property relations between electrical conductivity and elastic properties in an integrated tomographic approach.
Well established logging technologies from the oil and gas industry, such as Pulsed Neutron Gamma or Induction-Logs, are being verified in the field for CO2-exposed wells. These methods offer an instantaneous image of saturation at the time of the measurements. However, suitable technologies must be able to identify the saturation of CO2 in the reservoir and its change with time. In these respects, temperature measurements with fibre-optic cables combined with heat-pulse experiments have been recently proven to be successful in hydrology to detect water leakage at dams and offer a promising direction of development. Their employment in the detection of heat fluxes in wells relies, so far, on few numerical simulations and lab experiments. CO2CARE will assess this integrated system, which has been implemented at Ketzin. The temperature distribution along the wells has been monitored with high temporal and spatial resolution using fibre-optic sensor cables for distributed temperature sensing (DTS) in the “smart closed” well. These measurements enable the determination of local changes in fluid saturation within the reservoir in the immediate neighbourhood of the well by measuring the effective thermal conductivity of the formation and by this provide an additional monitoring tool to assess near-wellbore phenomena and well integrity problems.
Remediation
The hydrocarbons industry has many years of experience in applying gel treatments to reduce channelling in high-pressure gas floods and to reduce water production from gas wells. Referred to as relative permeability modification (RPM) or disproportionate permeability reduction (DPR) and water shut off (WSO) treatments, there are many examples of production performance modelling data for gel treated wells. Inspired by the successful field performance of polymer-gel solutions in modifying the permeability of both the rock matrix and fractures in oil and gas applications, CO2CARE has investigated the application of this technology and development of polymer-gel based remediation techniques to prevent leakage by creating a low permeability or an impermeable barrier above the caprock or fractured formations. Injection of saline water above the caprock, at a higher pressure than the CO2 pressure in the reservoir, would create an inverse pressure gradient to reverse the flow direction and increase the solubility of CO2 in the saline water barrier formed, and prevent or limit leakage. Such a procedure could enable fast and relatively low cost mitigation action once a leakage is detected. Furthermore, coupled with fluid management procedures during aquifer storage (saline water extraction and re-injection above the caprock), this methodology can also be used to minimise displacement and migration of native brine, and avoid pressure build up in closed or semi-closed structures. CO2CARE designed and modelled realistic CO2 storage reservoir fluid management scenarios for saline aquifers and depleted gas reservoirs and tested the effectiveness of pressure gradient reversal methods as a leakage mitigation/remediation technique. By these results, CO2CARE provide the operator with specific mitigation measures in case of leakage and by this fulfils one the regulatory obligations.
Risk management in CO2 storage site abandonment
The post-closure risk management procedures, the technology required to enable this and the use of corrective (remediation) measures that has been researched and developed in CO2CARE are pursuant to Articles 17 and 18 and Annexes I and II of the EU Directive and the OSPAR Framework for Risk Assessment and Management (FRAM). Considering the requirements of site closure and post-closure stages, the specific work carried out in CO2CARE has focussed on Tier 3 Risk Assessment. Additional work on Tier 2, particularly in relation to the development of appropriate leakage mitigation plan, will also be necessary depending on the status of the risk assessment work and the stage in the storage lifetime of the various sites. Prediction, assessment and optimisation of CO2 geological storage require robust performance assessment tools that can predict the fate of the CO2 in the subsurface in the long-term. The main objectives of such tools some of which have been developed across different disciplines and are being further developed currently consider the dominant physical, chemical and mechanical processes in different spatial scales and predict the fate of injected CO2 in the subsurface on the basis of such limited spatial and sometimes also temporal boundaries. In order to facilitate the purposes of risk management, namely to quantify the risks of individual scenarios for health, safety and the environment and identify appropriate risk management measures, it is essential to integrate such models over the different scales and, in particular, evaluate the significance of data, parametric and modelling uncertainties on risk estimates. Building on the idea of a Total System Approach for performance assessment of geological repositories robust models that investigate the full range of uncertainties and boundary conditions have been explored and developed in CO2CARE specifically for site abandonment.
Dissemination activities
The dissemination activities in CO2CARE focussed on different aspects which are described in the following:
Website: The Website was online in February 2011. New sections ( Multimedia, FAQ, Newsletters, Press Releases) were added during the second period to the project website. In the „Multimedia“-section, information films for almost all CO2 storage sites were added. Some information films were provided with English subtitles afterwards. A new section containing 10 FAQ was set up.
Scientific conferences: After the Kick-off meeting in January/February 2011 at GFZ Potsdam, three CO2CARE Annual Scientific Conferences took place: March 2012 - Imperial College, London; March 2013 - TNO, Utrecht; November 2013 - GFZ, Potsdam – final conference.
The so-called “Brainstorming Day on the long-term fate of geologically stored CO2” took place in June 2013 in Trondheim. The meeting was initiated and hosted by Statoil. The organising committee consisted of members of the projects “Mustang, Ultimate CO2, CO2CARE, CarbFix, and PANACEA”. Purpose of the conference was to debate key issues, innovations and best practices. Furthermore, CO2CARE issues were presented at several international conferences and meetings. (see A2 – List of Dissemination Activities).
Targeted Workshops: The first Targeted Workshop "Modelling and monitoring inputs for risk management and dry-runs elaboration" was organised by IFPEN and NERC in October 2012 at IFP Energies Nouvelles in Rueil-Malmaison - France. In addition to the workshop, IFPEN organised a public questionnaire on CCS, involving up to 36 students from IFP- school, who created some videos describing the meaning and the utility of CCS related to a social and economic context. These videos produced by different student-teams are accessible via the project website (“Multimedia”) and via YouTube. The second Targeted Workshop "Best Practice Guidelines" was organised by GFZ and NERC-BGS. In a meeting after the final CO2CARE conference, all WP leaders and consortium members discussed a draft of the document containing all the material collected by NERC-BGS.
Project brochures and the Best Practice Guidelines Summary: The first project brochure was created by BRGM with support by GFZ. This 4-page leaflet in English language mainly contains a short description of the project, some selected results, and a brief summary of what has been achieved so far and forthcoming outcomes. The second 8-page image brochure (M36) contains the summary of the three high-level site abandonment criteria (No Leakage, Predicted and Observed Performance and Long-term Stability) together with Risk Management. It describes the three main sites considered in the project and provides some samples of results from our research into well abandonment and reservoir management issues. Two schemes complete the brochure: a site-closure milestone chart leading to the transfer of responsibility according to the EU Storage Directive and a flow diagram of the traffic light system for risk-related decision. The English version of the brochure was generated with 1000 copies printed and distributed to the project partners and the European Commission. In addition, the brochure was translated into German, French and Italian by NERC, IFPEN, BRGM, OGS and GFZ and published in limited editions. All 4 versions are also available on the website in form of pdf in the section “Multimedia”.
The Best Practice Guidelines Summary was issued by NERC-BGS in English language. It was distilled and synthesised from the full Best Practice Guidelines document (Deliverable D5.4) into a coherent document. A final hi-resolution printer’s version was generated with 1000 copies printed and distributed to the project partners and the European Commission. The document is also available as pdf via the CO2CARE website.
Press Releases: In connection with the final CO2CARE conference, press releases were created by GFZ and translated by the partners into English, French, and Italian language. All versions were also uploaded to the website.
Contacts with media and outreach: Media training courses, mainly for press contacts, were attended by staff members of GFZ. In various interviews, Michael Kühn has given general information on CCS technology and in particular on the research work and progress at the pilot site Ketzin. Moreover, in Ketzin, an information centre, funded by the Federal Ministry of Education and Research (BMBF), was established for the public. CO2CARE results are demonstrated in the centre as well.
CO2CARE Newsletters: Six newsletters were issued between 2011 and 2013 on the CO2CARE website. These newsletters are freely available via the project`s website.
Scientific publications: Within the project`s lifetime, 23 full paper publications in scientific journals were issued (see section A). Three new publications are in preparation for 2014. Moreover, 14 abstract for international conferences are available, and 9 new abstracts are in preparation. During the final CO2CARE conference, 41 abstracts were issued in form of a conference volume. For 2014 a special issue on “Site Closure Assessment” is planned in collaboration with Greenhouse Gases: Science and Technology (Wiley-Blackwell). This new journal is covering the methods of carbon capture and storage, as well as utilisation of carbon dioxide as a feedstock for fuels and chemicals.
Potential Impact:
Potential impacts and use
Most of the results obtained in CO2CARE can be easily used by the community with some adaptation from case to case. They mainly consist in pragmatic recommendations based on laboratory and modelling results. In some cases, the key point could be to assess the field data that are required to run the proposed methodologies. At least most of the programme results has been published and are therefore available to the general public and stakeholders. The results will benefit to reducing risks by a better prediction of possible phenomena and by improving the definition of monitoring objectives. They support the transfer of responsibility of the sites and help to ensure a safe long term containment of (future) CO2 storage sites. In principle, the workflow can be also applied throughout the entire life-cycle of a storage project, whenever there is a mismatch between modelled and observed behaviour of sites and an irregularity in the storage operations occurs. The dry-run documents are expected to provide a widely used template for site abandonment and transfer of responsibility, based on real storage sites under a range of conditions and useful for both site operators and regulators. In the first instance the documents could be used at the site licencing stage to develop initial site closure plans and to inform early discussions and negotiations between the interested parties. Subsequently they could be used to inform the developing dialogue between site operator and regulator as the closure and post-closure phases are reached. The Best Practice Guidelines document will be of use in establishing effective protocols for site abandonment, useful for both site operators and regulators. Technical guidance is provided on how best to meet the three key requirements of the Storage Directive (viz. conformance, leakage and long-term stability) and also wider guidelines on post-closure wellbore and reservoir management and a robust protocol for risk management. In addition the Best Practice Guidelines contains recommendations directly relevant for the forthcoming review of the EU Storage Directive.
List of Websites:
http://www.co2care.org
List of beneficiaries with contact names
Beneficiary No. Beneficiaries name / website Contact persons
1 GFZ - Helmholtz Centre Potsdam, GFZ German Research Centre for Geosciences /Deutsches GeoForschungsZentrum, Potsdam, Germany,
Website: http://www.gfz-potsdam.de
Project coordinator: Axel Liebscher
Project manager: Mario Wipki
2 Imperial - Imperial College of Science, Technology and Medicine, London, Great Britain
Website: http://www3.imperial.ac.uk/earthscienceandengineering
Sevket Durucan
3 IFPEN-IFP Energies Nouvelles, Rueil Malmaison, France
Website: http://www.ifpenergiesnouvelles.fr/
Jean-Pierre Deflandre
4 TNO - Nederlandse Organisatie voor Toegepast Natuurwetenschappelijk Onderzoek, Utrecht, The Netherlands
Website: https://www.tno.nl/index
Jens Wollenweber
5 BGS- British Geological Survey (NERC), Great Britain
Website: http://www.bgs.ac.uk
Andy Chadwick
6 Istituto Nazionale di Oceanografia e di Geofisica Sperimentale (OGS), Sgonico ( TS ) – Italy
Website: www.ogs.trieste.it
Gualtiero Böhm
Stefano Picotti
7 Bureau de Recherches Géologiques et Minières (BRGM), Orléans Cedex 2 - France
Website:http://www.brgm.fr/brgm/CO2-en/default.htm
Hubert Fabriol
8 GEUS - Geological Survey of Denmark and Greenland, Denmark
Website: http://www.geus.dk/geuspage-uk.htm
Peter Frykman
9 Uppsala - Uppsala Universitet, Sweden
Website: http://www.uu.se/en/
Christopher Juhlin
10 AirLiquide - L'AIR LIQUIDE, Société Anonyme pour l'Etude et l'Exploitation des Procédés Georges Claude, France
Website: http://www.airliquide.com/
Claire Bourhy-Weber
11 RWE - RWE Power AG, Cologne, Germany
Website: www.rwe.com
Thomas Thielemann
12 Statoil Petroleum, Trondheim, Norway
Website: http://www.statoil.com/
Sveinung Hagen
13 Shell - Shell International BV, The Netherlands
Website: http://www.shell.com/
Claus Otto
14 Vattenfall - Vattenfall Research and Development AB, Germany, Sweden
Website: http://www.vattenfall.com/
Thomas Schulte
15 Veolia - Veolia Environnement Recherche & Innovation SNC, France
Website: http://www.veolia.com/en/
Natalia Quisel
16 AITF - Alberta Innovates – Technology Futures, Canada
Website: http://www.ucalgary.ca/research/funding/aitc
Stefan Bachu
17 UofA - University of Alberta, Canada
Website: http://www.ualberta.ca/
Rick Chalaturnyk
18 RITE - Research Institute of Innovative Technology for the Earth, Japan
Website: http://www.rite.or.jp/index_e.html
Ziqiu Xue
19 GCCC-BEG - The University of Texas System, USA
Website: http://www.beg.utexas.edu/gccc/
Susan Hovorka
20 LBNL - The Regents of the University of California, USA
Website: http://esd.lbl.gov/home/
Jens Birkholzer
21 PNNL - Battelle Memorial Institute Corporation, USA
Website: http://www.pnl.gov/
Alain Bonneville
22 CO2CRC - CO2CRC Management Pty Ltd, Australia
Website: http://www.co2crc.com.au/
Charles Jenkins
23 TOTAL S.A. - Total Exploration Production in Pau, France
Website: http://www.total.com
Cristophe Urbanczyk
The CO2CARE (CO2 Site Closure Assessment REsearch) project focused on site closure and preparation for transfer of liability of a CO2 storage project in order to assist regulatory authorities and stakeholders in implementing the EU Directive 2009/31/EC on CO2 Geological Storage. The project, which started in January 2011, was funded by the EU 7th Framework Programme and the industry and ran for a period of three years until December 2013. CO2CARE consisted of an international consortium of 23 partners from Europe, USA, Canada, Japan and Australia, represented by universities, research institutes, and energy companies. In order to incorporate up-to-date results and monitoring data 9 key injection sites in Europe and worldwide formed an integral part of the project: (1) Ketzin, Germany; (2) Sleipner, Norway; (3) K12-B, The Netherlands; (4) Rousse, France; (5) Montmiral, France; (6) Frio, USA; (7) Wallula, USA; (8) Nagaoka, Japan and (9) Otway, Australia.
The main objectives of the project were closely linked to the three high-level requirements of the EU Directive (Article 18) for the ultimate transfer of responsibility of a CO2 storage site to the Competent Authority typically 20 to 30 years after site closure. These high-level requirements are:
1. Absence of any detectable leakage.
2. Conformity of actual behaviour of the injected CO2 with the modelled behaviour.
3. The storage site is evolving towards a situation of long-term stability.
In particular, the project mainly focused on three key areas:
1. Well abandonment and long-term integrity.
2. Reservoir management and prediction from closure to the long-term.
3. Risk management methodologies for long-term safety.
These objectives were achieved via integrated laboratory research, field experiments and state-of-the-art numerical modelling, supported by literature reviews and data from the above mentioned portfolio of real storage sites, covering a wide range of geological and geographical settings. The current regulatory requirements and well and site abandonment technologies as well as the behaviour of abandoned CCS/CO2-exposed wells were reviewed and assessed. Detailed studies dealt with laboratory experiments on rocks and cements under different CO2 environments and in combination with numerical simulations of the well geomechanical integrity while also considering in parallel main geochemical aspects. Possible leakage pathways have been studied including the monitoring aspects and a specific remediation technique has been investigated on the basis of field representative cases and scenarios with the fruitful involvement of the project´s industrial partners.
Key research findings from the CO2CARE project, compliant with the EU storage regulation, were distilled in form of Best Practice Guidelines and disseminated to key stakeholders. Main results of so-called “dry runs” for the storage sites Ketzin, Sleipner and K12-B, which were discussed with European regulators, also contributed to the guidelines. These dry-run documents provide a template for site abandonment and transfer of responsibility, based on real storage sites and useful for both site operators and regulators. In the first instance the documents could be used at the site licencing stage to develop initial site closure plans and to inform early discussions and negotiations between the interested parties. Subsequently they could be used to inform the developing dialogue between site operator and regulator as the closure and post-closure phases are reached. The Best Practice Guidelines document will help establishing effective protocols for site abandonment. They provide technical guidance on how best to meet the three key requirements of the Storage Directive (viz. conformance, leakage and long-term stability) and also wider guidelines on post-closure wellbore and reservoir management and a robust protocol for risk management. In addition the Best Practice Guidelines contain recommendations directly relevant for the forthcoming review of the EU Storage Directive.
Summary description of project context and objectives
With the announcement of the European Strategic Energy Technology (SET) Plan, Carbon Capture and Storage (CCS) is expected to significantly contribute to the reduction of CO2 emissions in the atmosphere en route to a low-carbon economy. In anticipation of major CCS deployment by 2020, comprehensive research programmes should deliver the necessary operational and post-closure site management technologies for CCS. As far as closure and post-closure issues are concerned, apart from a few recently completed pilot scale field CO2 injection projects, and experience gained from enhanced oil recovery (EOR), the CCS industry has quite limited experience.
The fundamental regulatory principles surrounding site closure, transfer of responsibility (liability) to the competent authority, and post-closure obligations are set out in the EU Directive on CO2 Geological Storage (EU Directive). Pursuant to Article 18 of the EU Directive, the overall philosophy of the Directive regarding site closure is enshrined in the three high-level criteria for transfer of liability:
1. Absence of any detectable leakage.
2. Conformity of actual behaviour of the injected CO2 with the modelled behaviour.
3. The storage site is evolving towards a situation of long-term stability.
These conditions define satisfactory long-term site performance at a high-level. Significant gaps however remain in understanding how these high-level principles will be implemented at real sites, not only at the closure stage, but also at storage commencement, when future closure arrangements will need to be incorporated into a site storage license acceptable to both operator and regulator.
The Directive does not define the specific technical acceptance criteria, based on real site performance data, which can demonstrate that a site meets the three requirements. The identification of such criteria and the development of site abandonment procedures and technologies which guarantee the fulfilment of these criteria have been the main objectives of CO2CARE.
The experience and knowledge gained from field pilots suggest that fluid behaviour and its interaction with the containing rocks determine the overall performance of the CO2 storage complex during the injection and post-closure periods. The key issues in this respect are: injectivity; wellbore behaviour; and the response of fractures and faults in the reservoir, seal and geosphere. At present, the most challenging aspect of site abandonment is to predict, validate and demonstrate storage site behaviour during the post-closure period, as current experience of monitoring techniques and predictive simulations is restricted to ongoing injection projects and analogous research results from oil and gas production. The longest post-injection monitoring periods associated with field pilots involve Nagaoka (Japan) and Frio (USA) where injection has ceased four to five years ago. Both sites are included in the CO2CARE portfolio. However, monitoring and verification of medium to very long-term storage performance prediction is not available.
CO2CARE conducted integrated laboratory and field experimental research, supported by advanced numerical modelling, and developed methodologies and strategies that will be needed to meet the site abandonment, post-closure safety and transfer of responsibility requirements of the EU Directive. The meet its overall objectives, the CO2CARE project was structured in six research and development work packages to address the different aspects of site abandonment: 1) Current practices and regulatory framework, 2) Well abandonment, 3) Post-Closure Reservoir Management, 4) Risk Management, 5) Best Practice and Regulatory Compliance, and 6) Dissemination. A seventh work package addressed the Project Management. The main objectives of the project have been achieved through:
• review of tested site-abandonment procedures and site-performance criteria from oil and gas industry;
• critical review of the existing international CO2 geological storage regulations, comparison with current practice in related/relevant fields, identifying differences and knowledge gaps on site abandonment technologies;
• compilation of a database of first abandoned CCS/CO2-exposed wells with the benefits arising from data available to international partners of the project in particular;
• research into the factors critical to well and reservoir behaviour focusing on integrity assurance after site abandonment at all-time scales;
• laboratory and field research into well plugging techniques to improve well integrity and monitoring methodologies to monitor wellbores during the closure and abandonment phases;
• improvement and validation of the reliability of predictive numerical modelling of future storage site behaviour, studying the conformity between monitored and simulated site behaviour, and assessing short-term performance and long-term stability of storage;
• understanding of issues crucial to long-term storage integrity by studying both long-term potential deterioration and stabilisation processes at the sites of the CO2CARE portfolio;
• assessment of the applicability and effectiveness of current technologies used in the oil and gas industry for CO2 storage, wellbore and site remediation and, if necessary, development of new methodologies, with associated protocols for site operators and authorities;
• development and implementation of reliable risk management procedures and tools for site abandonment and assessment of post-closure storage safety, definition of appropriate long-term monitoring and remediation plans;
• development of fully documented ‘dry-run’ applications for transfer of responsibility at the three real storage sites Sleipner, Ketzin and K12B, incorporating interactive stakeholder input and subject to a project regulatory critique and review;
• development of guidelines for regulatory compliance and best practice guidelines on CO2 storage site abandonment, taking into account the site dependence of the research findings;
• dissemination of project´s results to stakeholders and public, contribute to public acceptance.
To allow for “real-site” data and research and to achieve the project objectives CO2CARE has incorporated up-to-date results and monitoring data from the main European and international CO2 injection sites through the industrial partners of the project and CSLF collaborators. The field site portfolio included:
• Ketzin (Germany) – First European on-shore CO2 storage in a saline aquifer with a reservoir depth of about 630 m. The site has injected slightly more than 67,000 tonnes of CO2 from June 2008 to August 2013. One injection and two observation wells were available since 2007; two further observation wells were available since 2012.
• Sleipner (Norway) – Off-shore saline aquifer, about 1 M tonnes per year of CO2 injected since 1996. Aquifer depth 800 to 1,000 m, 17 years of monitoring data available.
• K12-B (The Netherlands) – Offshore Enhanced Gas Recovery (EGR) over 60,000 tonnes of CO2 injected since 2004 in a reservoir at approximately 3,800 metres depth. The nearly depleted gas field comprises multiple wells, both producers and injectors.
• Rousse (France) – On-shore depleted gas reservoir. Injection of 120,000 tonnes of CO2 from oxy-fuel combustion into dolomitic reservoir (depth of about 4,200 m) started in January 2010, lasted for 2 years, and is followed by a 2-year observation period. The well and site abandonment will follow this observation period.
• Montmiral (France) – Natural CO2 (98.5%) accumulation at about 2,400 m depth. The accumulation has been exploited for industrial purposes since 1990, now in the process of closure. No closure procedure standards exist for CO2 natural accumulation sites.
• Frio (Texas, USA) – Onshore saline aquifer at a depth of 1,500 m. Total injected volume 1,600 tonnes. Site officially closed; post closure monitoring data available.
• Wallula (Washington State, USA) – Embedded in the activities of the Big Sky regional partnership, injection of 1,000 tonnes of CO2 in a 1,200 m deep basalt formation started in summer 2013. CO2 will be supplied from a paper and pulp mill. Two years after closure, a new well will be drilled to study mineralisation and compare the actual CO2 front with predictions.
• Nagaoka (Japan) – Onshore saline aquifer located at the Minami-Nagaoka gas field, 200 km north of Tokyo. Supercritical CO2 was injected at a rate of 20-40 tonnes per day into a 12m thick permeable zone, which lies about 1,100 m below the ground surface and more than 3,000 m above the gas reservoir. Injection started on 7 July 2003 and ended on 11 January 2005 with a total injected amount of 10,400 tonnes of CO2.
• Otway (Australia) – The CO2CRC Otway Project is a large-scale demonstration of storage in a depleted gas field. Injection was paused at this site in September 2009 after the injection of 65,445 tonnes of an 80:40 CO2:CH4 mixture. The project is intensively monitored by seismic, geochemical, atmospheric, groundwater and soil gas sampling methods. Future plans include an injection in a formation above the gas reservoir, to investigate residual trapping mechanisms.
Project Context and Objectives:
Over three years, CO2CARE focused on site closure and preparation for transfer of liability of a CO2 storage project in order to assist regulatory authorities and stakeholders in implementing the EU Directive 2009/31/EC on CO2 Geological Storage. The main objectives of the project were closely linked to the three high-level requirements of the EU Directive for the ultimate transfer of responsibility of a CO2 storage site to the Competent Authority typically 20 to 30 years after site closure. These high-level requirements are:
• Absence of any detectable leakage.
• Conformity of actual behaviour of the injected CO2 with the modelled behaviour.
• The storage site is evolving towards a situation of long-term stability.
In particular, the project mainly focused on three key areas:
• Well abandonment and long-term integrity.
• Reservoir management and prediction from closure to the long-term.
• Risk management methodologies for long-term safety.
CO2CARE achieved its project objectives via integrated laboratory research, field experiments and state-of-the-art numerical modelling, supported by literature reviews and data from a portfolio of real storage sites that covers a wide range of geological and geographical settings. This portfolio included 9 key injection sites in Europe and worldwide: (1) Ketzin, Germany; (2) Sleipner, Norway; (3) K12-B, The Netherlands; (4) Rousse, France; (5) Montmiral, France; (6) Frio, USA; (7) Wallula, USA; (8) Nagaoka, Japan and (9) Otway, Australia. Key research findings from the CO2CARE project, compliant with the EU storage regulation, were distilled in form of Best Practice Guidelines and disseminated to key stakeholders. Main results of so-called “dry runs” for the storage sites Ketzin, Sleipner and K12-B, which were discussed with European regulators, also contributed to the guidelines. These dry-run documents are a central outcome of the project and provide a template for site abandonment and transfer of responsibility, based on real storage sites and useful for both site operators and regulators. In the first instance the documents could be used at the site licencing stage to develop initial site closure plans and to inform early discussions and negotiations between the interested parties. Subsequently they could be used to inform the developing dialogue between site operator and regulator as the closure and post-closure phases are reached. The Best Practice Guidelines document will help establishing effective protocols for site abandonment. They provide technical guidance on how best to meet the three key requirements of the Storage Directive and also wider guidelines on post-closure wellbore and reservoir management and a robust protocol for risk management. In addition the Best Practice Guidelines contain recommendations directly relevant for the forthcoming review of the EU Storage Directive.
Current practices and regulatory framework
According to the European and international regulations, the liability for the storage site can be transferred to the licensing authority/government once the safety and conformity of monitoring with model predictions has been demonstrated. In the EU the CO2 storage Directive 2009/31/EC set out the regulatory regime and guidance for permitting CO2 storage. Around the world, relevant bills and regulations have been introduced in recent years too. In addition, regulations originating from the oil and gas sector concerning well abandonment are also relevant to CO2 storage well abandonment. To demonstrate the safety of the injected CO2 all regulations require a combination of monitoring, modelling and risk assessment tasks. Although there is large variation in the specific requirements it is standard to require approval for these tasks as part of a plan submitted to the authority in charge. To demonstrate the safety of the injected CO2, the results of monitoring, modelling and risk assessment, regulations require demonstration of no leakage, conformity with modelling predictions and that the site is evolving towards long term stability. Some regulations contain additional requirements including demonstrating no environmental problems, that the plume will not encounter any leakage pathways and demonstrating well integrity. There is a variation in the time period over which safety must be shown in different regulations, and an optimum time period is considered flexible.
Particularly in relation to well abandonment, it is recognised that there are different methods, materials and tests that could be used and most CO2 storage regulations do not specify techniques to be followed or standards to be met. Specific details on plugging are provided by regulations on the abandonment of hydrocarbon wells and sometimes other injection wells and these provide the best available guidance for CO2 storage well abandonment, although they may require updating to deal with CO2 injection specific issues.
Regulations typically contain a provision for transfer of liability once safety (CO2 and well plugging) has been demonstrated. The EU Directive 2009/31/EC requires further monitoring after liability transfer as a back-up measure, while other regulations (e.g. EPA UIC) do not. The IEA model regulatory framework contains a clause that the operator should also provide suggestions for the monitoring to be conducted after liability transfer.
Site abandonment of oil and gas fields is described and defined as the activity of the operator to close and leave a site according to safety and environmental requirements. It can generally be divided into two main activities, i) the abandonment of the wellbores drilled during operation, including plugging of wells, and ii) the removal of surface installations (e.g. well equipment, production tanks and associated installations) and surface remediation.
In site closure operations in relevant industries, mainly in the oil and gas industry, well abandonment and ensuring long-term integrity of wellbores are considered very important in terms of secure geological storage of reactive substances. These industries provide important information on main issues in current well abandonment procedures in CO2 environments and particularly in geological CO2 storage. Additionally, subsequent alternative utilisation of oil and gas fields is not generally considered at the time of abandonment and this might lead to important issues when a field is being considered for subsequent geological storage of CO2.
Well abandonment
The long term processes at the wellbore have been studied to evaluate the risk of leakage through and/or along the wellbore at different spatial scales. The elastic stress changes applied on the wellbore system during abandonment are generally minimal. Laboratory experiments were carried out on a custom designed CO2 wellbore test set up consisting of full scale wellbore casing, cement sheath (made of industry standard API RP 10B Grade G cement mix) and a stainless steel ring to represent host reservoir stiffness. A microannulus was generated between the well casing and cement sheath interface to represent a damaged well casing scenario. CO2 flow through experiments were carried out for three representative CO2 injection site scenarios (scenario 1. Sleipner type: temperature: 40 °C, salinity: 3.5 %, average reservoir pressure: 10 MPa; scenario 2. Ketzin type: temperature: 34 °C, salinity: 25 %, average reservoir pressure: 7.3 MPa; scenario 3. North Sea deep reservoir type: temperature: 92 °C, salinity: 12.5 %, average reservoir pressure: 35 MPa). It was observed that the rate of flow of CO2 (at constant inlet pressure) through the microannulus progressively reduces for shallow reservoir setting (scenarios 1 and 2 above) indicating self-sealing behaviour of the microannulus but remained unchanged for North Sea deep reservoir type setting. The former is attributed to carbonation reactions between cement and pore water resulting in leaching of species which in turn reduced the permeability. The latter is attributed to the lack of such reactions at high temperatures possibly due to the absence of pore water. The time dependent permeability change of the microannulus between the well casing and cement sheath under constant flux of CO2 derived was used as an input to reservoir simulations carried out to assess field scenarios on possible leakage of CO2 stored in a reservoir. It is inferred from the simulations that the self-healing behaviour of the microannulus enables containment of CO2 indicated by the constant volume of mobile CO2 and increasing volume of dissolved and trapped CO2.
Influence of impurities on geochemical reactions in the wellbore system
Batch experiments at near reservoir temperature and pressure (30°C, 8 MPa [80 bar]) investigated the potential geochemical reactions between CO2, synthetic pore-waters with low levels of either SO2, or NOx, or H2S impurities and samples of either Utsira caprock, 13Cr80 alloy borehole steel, or ‘Class G’ cement. The caprock experiments showed that the addition of the impurities caused some increased reaction in comparison with previous experiments using only pure CO2 and N2. In general, the N2-pressurised experiments showed little reaction, and the final fluid chemistry was similar to that of the starting fluid. The CO2-pressurised experiments appear to have attained approximate steady-state concentrations for most dissolved species after about 2 month. The main solid phase reaction observed upon addition of CO2 was carbonate mineral dissolution, in particular calcite.
In the cement samples although there was extensive reaction of the samples, the reactions did not result in the wholesale disintegration of the borehole cements used. In addition, the presence of impurities, other than the formation of gypsum (when S present), did not seem to significantly affect the nature of the reactions. Cl concentrations were found to vary in most of the cement samples, though the extent varied with the differing impurities. Qualitative X-ray map profiles were run across the samples to investigate the trends in the Cl concentration. The pattern of Cl enrichment and depletion illustrates the multi-reaction process that has occurred in the cement due to fluid infiltration. Initially Cl diffused into the cement ahead of the CO2 and reacted with the hydrated CSH and Ca(OH)2, to form a calcium chloride bearing phases. The slower diffusing CO2 then reacts, as HCO3-, with the newly precipitated calcium chloride phase to form secondary carbonate, which then liberates Cl. The liberated Cl then further diffuses deeper into the cement. Sodium follows an inverse pattern to that observed for Cl, with the area preceding the front being enriched in Na and depleted in Cl.
The addition, impurities in the experiments with borehole steel caused some increased reaction in comparison with previous experiments using only pure CO2 and N2. The steel in the CO2-pressurised experiments appears to have reacted relatively rapidly, allowing a concentration of dissolved reduced Fe to build up in solution, the presence of both CO2 and ‘S’ caused greatly enhanced steel dissolution in contact with the synthetic Utsira pore-water.
Overall, for all the materials studied here, the addition of impurities did cause some enhanced reaction but it was the presence of CO2 (with and without impurities) that had most impact on the reactions of the caprock and the borehole infrastructure.
Post-Closure Reservoir Management
A multi-disciplinary approach has been applied in order to provide a comprehensive understanding of the processes related to CO2 storage, and to identify and further develop monitoring methods for the post injection phase. The process understanding has been supported by designing dedicated laboratory experiments focussing on trapping mechanisms, such as dissolution and capillary trapping. Here, significant technical progress has been made, but still a large gap remains between a model and real geological settings. Numerical simulations of trapping mechanisms and of deterioration processes (such as, e.g. fracture growth due to chemical processes) have been performed. Long-term reservoir simulations for the Sleipner and Ketzin sites have provided insight into the contribution of different trapping mechanisms to the long term containment of CO2 in the reservoir. Reservoir simulations and monitoring data of the Sleipner and Ketzin site have been compared for an investigation of the conformity between simulations and monitoring results. A workflow applying performance criteria has been established showing that conformity between simulations and monitoring data enhances with increasing knowledge about the reservoir but discrepancies remain. Therefore, particular attention must be paid to establish realistic end-member model scenarios. The end-member scenarios must ensure a safe operation of a storage site and consider remaining uncertainties on the reservoir properties. At Ketzin, geoelectric and seismic surveys were performed with a special focus on long-term observation. Several types of innovative seismic sources were applied which can be installed permanently and which generate highly repeatable and high-frequency signals suitable to resolve small changes in the monitored rock volume.
Understanding of the long-term evolution of trapping mechanisms
Dissolution of dense-phase CO2 in formation water, followed by convective sinking is a key stabilization process in aquifers, acting on timescales from decades to tens of millennia. The onset of dissolution and sinking is poorly-understood and CO2CARE focussed on examining these crucial early processes by means of laboratory experiments and numerical models. Study of upward, vertical migration of CO2 gas through the porous medium revealed several interesting features:
• Initial entry of the gas into the porous medium formed several small ‘leaders’, only one of which eventually became dominant. The location of these was possibly controlled by minor differences in pore geometry and a slight reduction in capillary entry pressure. Thus even very small heterogeneities could play an important role in controlling gas flow into a porous medium.
• Once the gas entered the porous medium, it grew into an irregularly-shaped, elongate feature. Eventually this separated from the base of the cell and moved upwards through the porous medium as discrete, buoyant ‘packets’ of gas.
• The ‘packets’ of gas did not migrate perfectly vertically, but appeared to migrate randomly within a conical region (about ±12-15° from vertical) above the initial point of gas entry into the porous medium. This movement caused localised circulation of water around the ‘packets’ of gas as they rose.
• On reaching the upper surface of the cell the first part of a ‘packet’ of gas to make contact with the overlying free solution became the drainage point of all that particular volume of gas. This produced a focussed string of bubbles within the overlying free solution. As each packet was of finite size and took a different path, then these bubble releases were periodic, and were released at different locations. Gas movement through the entire cell went from continuous where it entered the cell, to episodic where it escaped the cell.
• Migrating larger ‘packets’ of gas through the porous medium resulted in smaller, irregularly-shaped bubbles of residual gas being left behind. The distribution of these changed with each passage of a larger ‘packet’ of gas; however their average distribution just within the migration cone appeared fairly constant, as did their average abundance (approximately 10% of the area of the migration cone).
Time-lapse movies of downward migrating plumes of (coloured) CO2-rich water through the porous medium revealed several important processes:
• Initial dissolution of CO2 into the water created a thin boundary layer where diffusion dominated transport.
• CO2 dissolution increased the density of the water, resulting in gravitational instabilities and the formation of small density plumes.
• As the plumes grew, they increased in length relative to width. They also increased in complexity, developing increasing numbers of lobes with increasing size.
• As the plumes grew they decreased in number through a process of localised lateral migration and assimilation.
• Larger plumes also migrated laterally across the cell. It is unclear however, if this was a consequence of cell-wide convection (i.e. an experimental artefact caused by the finite size of the cell), or a product of the system ‘flipping’ between different stable states of plume arrangement.
• As the plumes descended within the porous medium their outer edges became progressively more diffuse as they mixed with the CO2-poor water. They also appeared to slow down as they approached the base of the cell. This may have been controlled by limitations imposed by the size of the cell, but later in the experiment may also have been a consequence of increasing CO2 concentrations at depth, and hence decreasing density contrast between the plumes and the surrounding water.
Most probably, the instabilities are triggered by geological or physical heterogeneity. In real reservoirs, physical and chemical properties and overall timescales are significantly different from those in the laboratory, but there is no reason to suppose that our conclusions are not generally applicable and support the contention that the experiments support the validity of longer-term reservoir-scale flow simulations.
An integrated assessment of CO2 trapping mechanisms has been performed for the Stuttgart Formation at the Ketzin pilot site for CO2 storage by numerical simulations using different model coupling schemes. Long-term dynamic multi-phase flow simulations established the basis for all model couplings across all time scales. Pressure changes resulting in geomechanical effects are largest during the operational phase, whereas temperature related effects at the Ketzin pilot site are negligible due to injection of CO2 pre-heated to reservoir temperature. Geochemical reactions are governed by kinetics, and therefore long-term stabilization is determined by chemical processes. Chemical processes responsible for mineral trapping are expected to mainly occur during long-term stabilization at the Ketzin pilot site. Hence, assessment of long-term trapping mechanisms carried out for the Ketzin pilot site was extended integrating the geochemical effects generated by the migration of dissolved CO2. A cut-off value of 0.2 mol CO2/kg H2O results in a conservative mineral trapping scenario reaching 11.7 % after 16,000 years of simulation while a cut-off value of 0.1 mol CO2/kg H2O results in the optimistic mineral trapping scenario 30.9 % after 16,000 years. Dynamic multi-phase flow simulations indicate that only 0.2 % of the CO2 injected from June 2008 to August 2013 (about 67,270 t CO2 in total) is in a gaseous state, but residually trapped after 16,000 years of simulation. Depending on the cut-off value for the dissolved CO2 in our coupled geochemical simulations, CO2 dissolution is the dominating trapping mechanism that accounts for 68.9 % to 88.1 %, respectively. In both nonreactive and reactive simulations, no gas phase is present anymore after around 4500 simulation years, a value which is probably much underestimated due to the effect of a coarse grid used in the simulations; in case of non-reactive simulation the dissolved CO2 spreads significantly farther than in the reactive one after 10,000 years, however only in tiny concentrations.
Assessment of leakage detection thresholds at Ketzin and Sleipner
4D time-lapse seismic techniques have proved highly successful at monitoring CO2 in sandstone reservoirs, where the elastic properties of the rock are particularly suited to imaging fluid replacement (high porosity and relatively low rigidity). Leakage detection in shale dominated caprock sequences presents a greater imaging challenge. CO2 accumulations will potentially be highly localised in porous zones within a low porosity and permeability matrix. Synthetic seismic modelling has been used to investigate the potential detection threshold (in the caprock) of existing time-lapse seismic surveys acquired as part of the Sleipner and Ketzin CO2 storage monitoring programmes. A key factor in the ability of time-lapse data to detect small time-dependent changes is the degree to which successive datasets can be accurately repeated. Perfect repeatability would produce a noise-free difference dataset capable of detecting very small time-lapse changes. In practice, repeatability is far from perfect - difference datasets suffer from a variable amount of repeatability error or noise which acts to obscure real changes in signal. A four stage workflow has been applied to quantify detectability of both the Sleipner and Ketzin time-lapse seismic surveys in the presence of repeatability noise.
• The repeatability and magnitude of coherent and ambient background noise was calculated for the field seismic data using the repeatability metrics NRMS and predictability.
• A rock physics model based on all available petrophysical data was developed and used to calculate the effects of fluid substitution on the seismic response.
• A baseline synthetic seismic survey was built using the available petrophysical data.
• A series of synthetic seismic repeat surveys incorporating a body of (CO2) gas of known volume and saturation were computed using a finite difference code. An appropriate level of noise was added to the data based on the calculated repeatability metrics.
The repeatability metrics NRMS and predictability have been computed for the 2006 3D seismic monitor survey at Sleipner (2006 time-lapse processing) across a time window spanning 500 to 850 milliseconds. This window spans most of the Pleistocene Nordland Shale sequence, which forms the caprock to the storage reservoir at Sleipner. The mean NRMS is 54% and the mean predictability ~97%. The signal-to-noise ratio was also computed based on power spectra of the autocorrelation (signal + noise) and local cross-correlation (signal only).
Traditional Gassmann fluid substitution is not valid in shale because it assumes perfect communication of fluid in the pore space. The soft porosity model has been designed as an alternative to Gassmann theory in shale-rich sequences. The model requires a complete suite of petrophysical log data (VP, VS, Bulk density) together with laboratory analyses of matrix mineralogy and porosity. This information is only available for a small section of the lower Pliocene shale sequence in well NO15/9A-11. Consequently, the soft porosity model has been computed and compared to the results of traditional Gassmann fluid substitution using various fluid mixing schemes and rock physics parameters. The data suggest that traditional Gassmann fluid substitution with a patchy mixing scheme based on the Voigt average provides an acceptable approximation to the soft porosity model in the Nordland Shale. The synthetic seismic data have all been computed from petrophysical log data acquired for well NO15/09-18. The baseline survey thus comprises a series of planer reflectors corresponding to acoustic impedance contrasts observed at the well. Band-limited Gaussian noise was added to the baseline survey to give a comparable signal-to-noise ratio to the real data. A series of synthetic models were produced for specific wedge geometries and “image matrices” composed of seismic difference sections (repeat survey with a CO2 filled wedge minus the baseline survey) were calculated. The data suggests that very small accumulations of CO2 require gas saturations > 50% to be clearly imaged on seismic difference sections. At low saturations (<= 10%) it is very hard to identify amplitude changes induced by the CO2. At saturations in excess of 50% even small accumulations, occupying pore volumes of as little as 10 m3 could potentially be identified by careful seismic interpretation.
Time lapse seismic data acquired at the Ketzin site show a high level of both coherent and random repeatability noise. This could reflect a poorly designed time-lapse processing flow, but is likely due to low fold in regions where there is significant surface infrastructure. The detectability modelling carried out as part of this study focussed on a Jurassic Sandstone aquifer system, located above the main Triassic storage reservoir. This aquifer formed the main reservoir for a gas storage operation which was decommissioned in 2000. Any CO2 leaking from the Stuttgart formation will likely undergo secondary trapping in this layer. The repeatability metrics NRMS and predictability have been computed for the 2009 3D seismic monitor survey at Ketzin (2009 time-lapse processing) across a time window in the overburden spanning the Jurassic sandstone aquifer. The mean NRMS is ~30% and the mean predictability ~95%. A series of synthetic models were produced for specific wedge geometries and “image matrices” composed of seismic difference sections (repeat survey with a CO2 filled wedge minus the baseline survey) were calculated. The data suggests that small accumulations of CO2 (< 2 m thick) require gas saturations > 70% to be clearly imaged on seismic difference sections. At low saturations (<= 50%) it is very hard to identify amplitude changes induced by the CO2 above the repeatability noise. If the maximum thickness of the wedge is increased to 5 m, the CO2 is clearly imaged at saturations >= 30%. At patchy saturations below 20%, it is unlikely that detailed seismic interpretation will identify accumulations of CO2 above the background repeatability noise.
Pressure gradient reversal and injection of polymer-gel solutions as remediation measures
For the case of leakage occurring on a storage site, remediation techniques were investigated by laboratory and simulation studies. They consist of pressure gradient reversal (PGR) and the injection of polymer gels.
The effectiveness of the pressure gradient reversal (PGR) method as a potential remediation technique for CO2 leakage from deep saline aquifers was investigated using a generic 3D reservoir/caprock model. A hypothetical CO2 storage operation involving CO2 injection at 1 Mt/year for up to 30 years down-dip of a structure high in the model domain was considered. Two leakage locations in the transient source region, at a distance of 200m and 1200m, respectively, to the CO2 injector were selected for conducting above-zone brine injection simulations. The brine injection simulation results indicate that the performance of PRG is strongly affected by how early leakage is detected from the start of injection (time-to-detection), which in turns is controlled by the detection threshold, leakage pathway permeability and the distance to the injection well.
Crosslinked hydrolysed polymer-gel injection is used in petroleum industry to improve conformity of fluid flow in the reservoir, remediate leakage around wells and also used in conjunction with the prospect of enhanced oil recovery at various temperature and pressure conditions. To test whether this technology is also suitable for CO2 storage and to study the possibility of mitigating undesired flow of CO2 in host reservoir through the use of polymer gel solutions has been studied through systematic laboratory tests on characterisation of suitable polymer gel solution both as bulk material and its characteristics as fluid in porous medium. The inferences and results obtained from laboratory experiments were utilised to analyse possible remediation scenarios through reservoir simulations performed using a generic reservoir model. One of the important characteristics of polymer gel solution is the gelation time: which is the time lag between the addition of cross linker and the formation of a solid gel, in between which the polymer gel solution retains its rheological characteristics. This parameter in conjunction with reservoir permeability, temperature and pressure, dictates the area around polymer injection well that can be remediated.
In the laboratory, polymer was injected into sandstone core samples of known permeability and was allowed to gelify in-situ. Pure CO2 was injected into the core sample after polymer gelified and it was observed that the permeability of core sample to CO2 was reduced and polymer gel was stable (did not degrade over time). This reduction in permeability was implemented in a representative CO2 leakage reservoir scenario, wherein the host CO2 was simulated to leak into an overlying formation through a fracture in the caprock. The simulation illustrates the effectiveness of polymer gel remediation treatment, wherein the polymer gel was successful in mitigating the leak and the CO2 that leaked prior to remediation treatment has been dissolved into the formation brine and remained immobile.
Study of conformity between monitored and simulated site behaviour; assessment of short term performance and assessment of long-term stability of storage
Demonstrating conformance is technically challenging because a perfect match between observed and modelled behaviour, that is also demonstrably unique, is likely to be impossible to achieve. CO2CARE research therefore focussed on three related issues:
• To show that predictive modelling capability increases systematically with time as monitoring data is progressively acquired. This indicates that storage processes are well understood.
• To show that as more monitoring data is acquired through time, uncertainties progressively reduce, but focus must still be maintained on the less likely ‘end-member’ model scenarios to avoid the possibility of unexpected or divergent future outcomes.
• To show that, at site abandonment, predictive models calibrated by monitoring data can reduce the uncertainty envelope sufficiently for unexpected or divergent future outcomes to be ruled out.
In the case of the commercial-scale Sleipner project, multiple 2D (axisymmetric) and 3D reservoir simulations were compared to results of 3D time-lapse seismics. Research focussed on predicting and measuring the progressive development of the CO2 plume. Conformance testing was based on a number of performance criteria:
• Plume footprint area.
• Maximum lateral migration distance of CO2 from the injection point.
• Area of CO2 accumulation trapped at top reservoir.
• Volume of CO2 accumulation trapped at top reservoir.
• Area of all CO2 layers summed.
• Spreading co-efficient (storage efficiency).
Initial predictions in 1996, using only baseline observations, with no plume monitoring data, had a high degree of uncertainty arising mostly from uncertainty in how the CO2 would be trapped within the reservoir. Possibilities ranged from a single layer at the reservoir top or several layers at different levels within the reservoir. As a consequence, performance measures, such as plume area footprint, had high range of possibilities. Predictive scenarios for 2001 and 2006 used the monitoring data that was available at those times. This principally comprised a suite of repeat 3D seismic surveys, acquired in 1999, 2001, 2004 and 2006, which confirmed that CO2 was trapped at multiple levels within the reservoir, and provided additional constraints such as arrival time of the CO2 at the reservoir top, estimation of CO2 flux into the topmost layer and individual layer extents. Improved reservoir temperature data also became available as the injection progressed. The predictive scenarios for 2001 and 2006 included only multi-layer plumes. In addition the 2006 predictions had improved reservoir temperature constraints, so end-member possibilities progressively became much closer together. As a result, predicted ranges for the performance measures are much reduced and the likelihood of the end-members leading to unexpected or divergent future outcomes is much reduced. A key performance measure that cannot be accurately assessed by the axisymmetric modelling is the lateral migration of the topmost layer of the plume, directly beneath the topseal (whose topography is known from the baseline data). Initial 3D models based on 1996 data had a very wide spread, but uncertainty reduced markedly as the time-lapse datasets became available. It is clear however that even by 2006, a perfect prediction of the 2008 plume was not easily obtainable. This is due to continued (though much reduced) geological uncertainty and also to likely limitations in the predictive model itself. Nevertheless, it is also clear that the basic process of layer development (buoyancy-driven migration by fill-spill beneath the topseal topography) is well understood. As more monitoring data is acquired uncertainty will reduce still further and the likelihood of unexpected divergent future outcomes is very small. Thus, as more seismic data became available, predictive models could be matched more accurately to the observations and became more reliable predictors of future performance.
Risk Management
The assessment and mitigation of the risks related to the geological storage of CO2 (e.g. loss of containment, potential direct effects of CO2 leakage on the biosphere) and the containment of CO2 in the storage complex particularly during the stages of site closure and the post-closure have been addressed according to the prescriptions in the EU Directive.
Plan for risk management prior to and after transfer of responsibility
A risk management plan was established covering the requirements of the EC Directive on the geological storage of carbon dioxide (DIRECTIVE 2009/31/EC) and the OSPAR Guidelines. The timeline considered by this plan encompasses the final injection period within the operational phase as well as the post-closure/pre-transfer and post-transfer period of a storage project (Phase 4-6 according to EC GD3). In order to provide a well-structured procedure for risk management within these project phases, 17 Site Closure Milestones (SCM) are introduced, which are implemented into the different project phases The milestones are closely linked to the requirements of the EC Storage Directive and describe key actions or key moments in time during site closure and transfer. The proposed SCMs (not to be confused with the milestones mentioned in EC GD3) lie in operational Phase 4 and post closure/pre-transfer Phase 5, which are, for a better understanding, split into three sub-phases based on the terms used in the EC Guidance Document:
• final operational sub-phase (part of Phase 4, including site closure)
• post-closure sub-phase (first phase of Phase 5)
• pre-transfer sub-phase (second phase of Phase 5, including transfer of the site)
Development of criteria for transfer of responsibility
A report on criteria for transfer of responsibility from the site owner to the state has been provided. Decision making in the post-operational phase by the operator is targeted on minimizing risks demonstrating that the site behaviour is understood and that conditions required to transfer the responsibility for the site to a Competent Authority (CA) according to the EC Storage Directive (Directive 2009/31/EC, EC 2009) have been met. To be able to demonstrate the fulfilment of the three high-level criteria of the EC further more concrete low-level criteria and milestones applicable on an operational level are required. All of the three fundamental criteria mentioned above rely on monitoring data and models. Monitoring and/or model data are the key outcomes to demonstrate the understanding of site behaviour. By developing a risk management (RM) plan a milestone chart specifically relating to each of the three fundamental criteria has been shaped (called Site Closure Milestones “SCM”). The criteria derived from the risk management (RM) plan are termed “R-type” criteria. In order to assess whether the R-type criteria are met, a traffic light system was set up to support decision making during and after closure of the storage site (cessation of injection). The traffic light system establishes whether or not the monitoring and modelling data are in compliance. The criteria derived from this traffic light system are “T-type” criteria. The system consists of a decision workflow determining the condition (state) by which the abandoned site in question is characterized. Three different states are foreseen:
• Status green: MMO of all parameters are within tolerance, i.e. the site is in a regular and expected condition.
• Status orange: MMO of one or more parameters are off tolerance; the operator has to prove whether the models in question have to be recalibrated or irregular site behaviour is present; this involves a discussion between operator, experts and the CA.
• Status red: Irregular site behaviour is present; additional monitoring, counter measures, and mitigation measures have to be applied.
The presence and practical implementation of required (mandatory) site-specific monitoring parameters subject to modelling and monitoring (Criterion T1) and a prioritization of models - mandatory and optional – (Criterion T2) are two important criteria for decision making with respect to risk management and the transfer of responsibility.
The practicability of the proposed traffic light workflow has been evaluated on the K12-B site to test how the traffic light system can support the closure of a CO2 storage site. The benefit of the proposed traffic light workflow could be demonstrated and it is shown how the traffic light system can support the decision making in CO2 storage site responsibility transfer and abandonment by presenting an example of how to deal with an unpredicted behaviour or irregularity of the storage site (using bottom-hole pressure and monitoring data). The approach to define criteria leading to the transfer of responsibility for the site revealed that, although based upon a generic framework, the definition of such criteria is highly site dependent. In particular, the definition of tolerable model-monitoring deviations and required accuracies/precisions of models and monitoring techniques is ambiguous and requires thorough considerations by the operator of the site and the Competent Authority. Although the traffic light system was designed for treating irregularities in the final stages of the storage lifetime it is well applicable to the treatment of irregularities in all stages of a storage project.
Method for determining full system risk and uncertainty including optimisation
For safe storage of anthropogenic CO2 in the subsurface, four trapping mechanisms have been identified, acting on time-scales ranging from months to tens of thousands of years: 1) structural/stratigraphic trapping; 2) residual trapping; 3) dissolution trapping; and 4) mineral trapping. This study mainly focused on structural/stratigraphic trapping of injected CO2 in a storage reservoir, with special reference to the post-closure and post-transfer phases of a storage project. It is recognised that mobile CO2 phase in the storage reservoir (source for potential leakage) has a tendency to follow the general topography and migrate up dip with time, further beyond the area where the plume is located at the time of site closure. If one considers some of the CO2CARE project field sites, for example Ketzin, CO2 has been forecasted to migrate up an anticline towards the North, until reaching a vertically sealing fault, after CO2 injection is ceased in 2013. With respect to the potential leakage risk in the post-closure phase and beyond at the Ketzin site, the following observations may be made.
• The risk of leakage through the injection and monitoring wells is likely to be low, simply due to the diminishing source (mobile CO2) in the near-wellbore region over time.
• The caprock/faults along the CO2 migration path pose a potential risk over short to medium term. Monitoring tools other than plume imaging may need to be considered to ensure containment.
Above-zone pressure monitoring, which allows the probing for leakage of a wider area beyond the present CO2 plume, has been employed at Ketzin. No noticeable pressure communication between the storage reservoir (630 m depth) and a shallower aquifer (~420 m depth) has been observed.
Assuming that CO2 injection was ceased in 2006 at Sleipner, post-injection plume migration has been simulated by BGS until 2034, showing CO2 plume extending further Northwards, and becoming largely stationary by 2034. Unlike the Ketzin site, above-zone pressure monitoring cannot to be applied effectively at Sleipner due to the very low overpressure (~0.1 bar) generated in the storage reservoir by CO2 injection.
A risk assessment methodology, which considers the dynamic CO2 plume migration behaviour in the storage reservoir, with special reference to the post-closure period, was developed. The methodology involves:
• First step: CO2 plume evolution and migration, in the long-term over time, as a dynamic source for potential leakage out of the primary storage reservoir is evaluated by taking into account reservoir heterogeneity and data uncertainty.
• The second step considers the characteristics of specific and potential leakage pathways (caprock fractures, coarse grained lobes; wells and faults) to account for leakage likelihood and leakage rate.
• The last step evaluates, through simulations and uncertainty assessment, the potential CO2 leakage through the identified leakage pathways for different scenarios regarding the leakage path permeability and a detection threshold.
In order to demonstrate the implementation of the methodology described above, a generic reservoir/caprock model was set up and used to simulate CO2 injection of 1 Mt/year for up to 30 years in a saline aquifer and evaluate the potential leakage risk. A number of leakage scenarios were also set up and assessed in order to present and describe different elements of the methodology developed. Reservoir simulation of CO2 injection at a rate of 1 Mt/year for 30 years from 2012 was carried out to evaluate the plume migration behaviour during injection and after the termination of injection. A pore volume multiplier of 100 was used during simulations to represent the connected pore volume beyond the model domain. It was found that the plume largely stabilised at about 120 years from the start of injection under both sealing/non-sealing fault conditions. The influence of the state of the fault (sealing or non-sealing) along the anticline, after being approached by the advancing CO2 plume, on further migration of the plume is clearly observed. Based upon this plume migration behaviour, and the tendency to migrate up dip following the formation topography, the plume footprint may be broadly divided into:
• the transient region (where the free CO2 largely has a limited residence time), and
• the non-transient region (where the free CO2 residence is more or less stable).
This distinction has a direct bearing on the potential leakage profiles in the two regions. Time lapse 3D seismic surveys have been successfully applied to image CO2 plume and monitor its migration in the storage reservoir in both Sleipner and Ketzin sites. As well as demonstrating CO2 containment during the injection period, the seismic monitoring data are valuable for calibrating the reservoir model. Nevertheless, forecast of plume migration during the post-closure period using the calibrated reservoir model is still likely to be subject to varying degree of uncertainty due to lack of knowledge of the farfield porosity/permeability distribution. One way to address this uncertainty is to perform statistical analysis by generating a large number of realisations of (farfield) porosity/permeability distribution of equal probability, based upon our understanding of the storage site geology and available seismic, well and core data. To illustrate the impact of heterogeneous porosity/permeability distribution, in particular the presence of high porosity/permeability channels, on CO2 plume migration, stochastically generated channels were implemented in the generic model. To this end, the reservoir formation was divided into two major facies types, namely the floodplain and channel facies. Facies modelling was performed using the Petrel software package. A total of 10 realisations with different channel orientations were generated. For each realisation the two facies were then populated stochastically with porosity and permeability using the Sequential Gaussian Simulation method. For illustration, simulation of CO2 injection, at an injection rate of 1 Mt/year for 30 years, as in the base case in the absence of high permeability channels, was carried out for the 10 realisations generated as described above. Compared to the base case without channel structures, it is seen that CO2 tends to migrate preferentially along the least resistance paths, i.e. the channels. Nevertheless, the injected CO2 eventually reaches and accumulates at the top of the anticline after injection is ceased. Based on the analysis of the entire footprint of the CO2 plume for the 10 realisations, a plume footprint mean probability map, which gives a spatial overview of the likelihood of CO2 plume migration spanning both the injection and post-injection phases, may be obtained. For comparison, the extent of the plume for the base case (without the channels) is superimposed on the map. Dividing the CO2 plume footprint into a transient and a non-transient region also helps track the amount of free CO2 in each region – the source of potential leakage risk - over time. Free mobile CO2 in the transient region starts to decline when injection is terminated after 30 years, and thus represents a diminishing source for potential leakage within this region over the 120 years simulation period. On the other hand, free CO2 accumulated in the non-transient region is likely to remain a source for potential leakage over hundreds of years after injection is terminated. Using the generic reservoir/caprock model, an attempt was made to compute and map potential leakage risk profiles, i.e. the total amount of leaked CO2 through the caprock, and the leakage time periods, at various locations in both the transient and non-transient regions. A total of 30 potential leakage areas were considered, each being represented by a leaky grid block. The leakage scenario considers one leaky block at one time iteratively. To simulate CO2 leakage, a leakage pathway is intentionally created by assigning a permeability of between 1 and 10 mD to the column of grid blocks (200m x 200m) in the caprock between the storage reservoir and the permeable layer above. During simulations, the cumulative leakage from the storage reservoir is monitored and injection is terminated when a pre-set leakage detection threshold/limit is exceeded. A detection threshold between 1,000 to 10,000 tonnes of CO2 was considered here. The time (year) it takes for the leakage to be detected during the simulation, i.e. time-to-leakage detection, referred to for simplicity as the time-to-detection (TTD), maybe computed and it is expected to vary spatially within the CO2 plume footprint. At each leakage location, the TTD depends on the combined effect of a detection threshold applied and the leakage pathway permeability assigned. The TTD values computed for a grid block permeability of 10 mD and a detection threshold of 10,000 tonnes are mapped on the plume footprint. As would be expected, it would generally take longer to detect CO2 leakage if the leakage location is further away from the injection well. Although CO2 injection is terminated once leakage is detected, leakage is continuously being monitored to obtain potential leakage profile, namely the total leakage duration and the cumulative CO2 leakage, at each leakage point. Simulations are stopped at year 2132, some 120 years from the start of injection at year 2012. The computed leakage profiles display very different trends in the two regions. Whereas the leaked CO2 mass reaches a plateau in the transient region, leakage is still continuing in the non-transient region in 2132. It is further noted that leakage in the transient region would continue for a further period of time and the total leakage duration is significantly longer than and positively correlates with the time-to-detection at each grid block. So far, the methodology focused on the migration behaviour of the free CO2 in a storage reservoir during and post CO2 injection periods, which is a dynamic source for any potential CO2 leakage, and demonstrated how this dynamic behaviour can affect the (unmitigated) leakage profiles in the presence of a leakage pathway along the plume migration path. The risk of CO2 leakage out of the storage formation depends on the existence of a permeable pathway in the caprock, capillary entry pressures for the pathway to be exceeded and a pressure differential (including buoyancy) sufficient to transport CO2 along this pathway. For CO2 containment risk assessment, three broad types of leakage pathways have been identified:
• caprock;
• injection/abandoned well;
• faults and fault zones.
Of the three leakage pathways, leakage through wells will lead to relatively localised leakage, and thus it may be argued that it will be possible to remedy such leakage without any significant volumes escaping out of the storage complex. Diffuse leakage through fractures or pore-space in the caprock, on the other hand, will generally be more difficult to detect and remedy. A probabilistic methodology for the existence of a percolating fracture network in a caprock/reservoir system has been developed and the geomechanical failure risk assessed, taking into account the uncertainty in the measurements and data available. Implementation of these methodologies has been demonstrated using the Otway data as part of the full system risk and uncertainty assessment methodological development. The main fracture data at the Otway site comes from borehole image logs in CRC-1 and a second well, CRC-2 drilled very close by, which terminates in the shallow Paaratte formation. Analysis of the FMI logs was carried out to extract information on fracture attributes (orientation, line density) in the reservoir and the caprock (Belfast mudstone and the overlying Skull Creek mudstone). Examination of core samples revealed only one open fracture, with most being closed and filled with carbonate. There was no information on the length of fractures. The orientation distribution of the Belfast and Skull Creek layers were taken from the data CO2CRC provided. The length distribution is not known, therefore a power-law length distribution was assumed, using a range in order to represent the full range of possible length distributions. Four scenarios of the ratio of open to closed fractures were used in fracture modelling: 0.05 0.1 0.2 and 1, though data indicates that a scenario with a lower aperture fraction is most realistic. The uncertainty in length distribution is the main cause of uncertainty in percolation probability for a given aperture scenario. By using a large number of different sets of fracture parameter values and modelling a large number of fracture network realisations for each set of values, the probability of a percolating pathways existing in the Belfast and Skull Creek layers were obtained. The results of percolation assessment indicate that it is very unlikely in all regions with an aperture fraction of 0.05 and 0.1. The findings of and knowledge gained from the studies, in conjunction with an understanding of the free CO2 behaviour in the storage reservoir, can help better evaluate CO2 leakage risk and associated uncertainty for the closure and post-closure phases for a specific storage site.
Uncertainty and risk assessment of the Ketzin Site
A methodology was applied for detailed uncertainty and risk assessment of the Ketzin pilot site during the post-closure period. The objective of the assessment is to quantify the uncertainty in the behaviour of the CO2 plume for the purpose of risk evaluation. The uncertainty in the plume behaviour is caused by the geological uncertainties, namely the distribution of the fluvial channels in the far-field region of the reservoir inside the Stuttgart Formation. Hence, multiple stochastic realisations were generated by honouring the data available in the near-field region at the injection well Ktzi 201 in order to represent equally likely static geological models. These models were subsequently implemented in the Eclipse 300 reservoir simulation software. The results obtained for the CO2 saturation distribution were summarised as probability maps that give an indication of the most likely locations in the reservoir containing mobile CO2 (assuming a cut-off of 5% for residual saturation) at different times for 500 years during the post-closure period. Furthermore, the probability maps were also used to quantify uncertainties in the parameters that characterise the plume behaviour, namely the plume arrival times, the plume residence times, and the maximum amount of mobile CO2 for the ‘transient’ and the ‘non-transient’ regions of the reservoir.
Best Practice and Regulatory compliance
Guidelines for regulatory compliance and Best Practice for storage site abandonment has been established. Key research findings from CO2CARE were distilled and integrated into site closure and abandonment protocols compliant with the EU storage regulation (the Storage Directive, OSPAR and the EU ETS). A set of hypothetical ‘dry-run’ applications for site closure and abandonment for the Sleipner, K12-B and Ketzin injection projects has been prepared. The dry-runs are hypothetical in the sense that they have assumed arbitrary dates for cessation of injection, but are otherwise realistic in that a full risk analysis was carried out for each site and scientific arguments presented to address the technical criteria for transfer of responsibility as required by the EU Storage Directive and national regulations.
The dry-run documents included:
• Description of the storage site background with geological setting and injection history.
• Risk assessments and risk management plans
• Detailed assessment of each of the three key criteria for transfer of responsibility:
o Zero detected leakage
o Conformance of modelled and observed site performance
o Evolution towards long-term stability
• Post-abandonment monitoring plans
Regulatory compliance via interactive stakeholder involvement
A Regulators Workshop for Site Abandonment Dry-runs - Best Practice for CO2 Storage and Transfer of Liabilities (Milestone 10) - was organised and the regulatory responses to a number of hypothetical but detailed closure cases for real CO2 storage sites aquired. The CO2CARE project participants presented the 'dry-run' case-studies for the transfer of responsibility for Sleipner (Norway), K12-B (The Netherlands) and Ketzin (Germany). The transfer cases are hypothetical in the sense that they have assumed arbitrary dates for cessation of injection, but are otherwise realistic in that a full risk analysis has been carried out for each site and scientific arguments were presented at the workshop to address the technical criteria for transfer of liability as required by the EU Storage Directive and national regulations. The role of the regulators in the workshop was to review and critically asses each of the case-studies, so that the CO2CARE project participants could develop best practice guidelines for site transfer. The outcome of the workshop was a series of key recommendations from the regulators, listed below:
Risk assessment process
• Acceptance of the Transfer Report (and Transfer) marks the end of a process that began with site selection and a permit application many years previously. Communication between operator and Regulator should have been continuous during that period, so the transfer report should contain no surprises. The transfer report is part of a process which should build mutual confidence between regulator and operator.
• Competent Authorities may wish to undertake their own simulations based on static models developed by the operators, to make an independent evaluation and consider the effectiveness of monitoring. Static geological models and numerical reservoir simulations run on in-house software platforms rather than commercially available platforms such as Petrel and Eclipse may not be acceptable for transfer purposes because the Competent Authority would not be able to run the models.
• The Risk Assessment methodology used in the update of the post-closure plan should be standardised and use a standard template, though each site is likely to throw up different risks so the assessments would depart from that point onwards. The standardised template should contain all relevant facts and issues. Risk assessment starts at the beginning of the life cycle of a CO2 storage site, implying that the standardisation of the methodology should start there too.
• Risk matrices should be made for all sites.
• Interactions with other uses and users, both currently and in the future should be included in considerations of risks.
• Only CAs can assess the risks that might arise from operation and closure of multiple assets within a formation or region.
• Long-term well integrity and caprock sealing should be considered. This should include documentation of the configuration and history of the well using operational information such as cement job evaluation, plug testing, final abandonment, and other issues.
Conditions of transfer
• The Transfer report has the function of a contract – and it needs to contain key messages for the public. The State is likely to produce a counterpart document as well.
• From an operators perspective, it may be necessary to include a statement in the storage permit along the lines of: “if the storage site performs as predicted, no leakage is detected by the monitoring technologies deployed according to the monitoring plan, the site is evolving towards a state of long term stability, and all the terms and conditions in the storage permit are met, then the Competent Authority will accept transfer of the site”. This would provide comfort to the investors and operator that the State would accept the site back once the storage operation had been successfully completed.
• In a discussion of liability associated with transfer it was pointed out that society must take or share the risk of storage if it wants the benefits of CCS.
• Are regulators obliged to take the site back if the site is conforming and operated as expected?
• The definition of the storage complex is very important because this encompasses the entire rock volume affected by the storage operation. Views of other stakeholders involved or potentially involved in this volume and area should be addressed in the transfer report.
• Is there anything to be learnt from the sale of oil and gas fields that can inform the CO2 storage site transfer process?
Managing uncertainty
• As predictions of the future site performance are based purely on forward modelling, more numerical reservoir simulation runs in which more model parameters are varied may be necessary. It may also be necessary to conduct modelling of other aspects of the site performance, e.g. geomechanical stability. Predictions based in single lines of evidence are likely to be insufficient.
• A far greater level of detail on how models were constructed and how they evolved throughout the storage site characterisation, construction and operation should be included. Furthermore, information on how such models take uncertainty into account would likely be necessary.
Post-transfer
• In the section on conformity of models and monitoring, at least three reasons for monitoring should be addressed: monitoring to improve knowledge of the site and its performance, early warning of potential departures from modelled performance allowing time to intervene, monitoring of the efficacy of corrective measures, technical and social baseline needs.
• Financial arrangements and a plan for post-transfer monitoring need to be made – it is too easy and not sufficient to say monitoring can stop; it will be needed.
• Liabilities for certain risks will remain with the operators post-transfer.
• The transfer report should contain recommendations of ‘light’ monitoring during the post-transfer period for the CA. Whilst not implying significant potential risks of leakage, such monitoring might be undertaken in areas or multiple injection and as part of a CA’s duty of care.
• As CO2 storage is a multi-generational operation, consideration should be given to what tools and data should be transferred to the CA. Models may need to be used post-transfer, if a significant irregularity or leakage was detected.
Key research results and the outcomes of the Regulators workshop and Dry-run documents were compiled into Best Practice Guidelines for storage site abandonment and transfer of responsibility. The guidelines are a set of pragmatic and workable generic procedures, suggested best practices and other recommendations and observations for the safe and sustainable closure of geological CO2 storage sites. CO2CARE has developed key learnings in the risk management, well abandonment and post-closure (post-injection) reservoir management:
Risk Management: We propose a set of site closure milestones to be fulfilled in chronological order before site closure and, subsequently, transfer of the storage site to the Competent Authority. We have also developed a set of high-level, risk based and technical criteria that can be used to determine whether each of the site closure milestones has been reached. We recommend that the Competent Authority and the storage site operator agree a priori on the specific conditions for deviations from the predicted site behaviour that will trigger corrective measures. These might be based on quantitative threshold values such as the measured difference, or offset, between predicted and measured performance measures. We also recommend that criteria for long term safety and security of the site should be based purely on technical considerations and should not be linked to prescriptive time spans.
Well abandonment: A key requirement for the long term safety and security of geological CO2 storage is that wells contacted by the stored carbon dioxide should not leak; it is clear that the quality of the original wellbore construction is the main factor in ensuring long-term wellbore integrity. Subsequent to this, it is necessary to assess whether wellbores will have been susceptible to additional damage or degradation during the CO2 injection operation. Longer-term prediction of wellbore performance remains challenging and depends on various types of predictive modelling and experimental or analogue information. Laboratory experiments conducted by CO2CARE exposed typical well casing, cement and rock materials to CO2 for periods up to several years. We found that for typical storage conditions, cement carbonation and steel corrosion reactions can cause porosity plugging in the rocks and wellbore annuli, tending to retard or prevent CO2 migration up, or alongside, the wellbore.
Post-closure reservoir management: Following cessation of CO2 injection the main aim is to demonstrate that the key regulatory requirements for transfer of storage site liability to the Competent Authority have been met. These are based around demonstrating understanding of reservoir processes and the ability to make robust predictions of future behaviour, and providing assurance against leakage. Migration of the CO2 plume and reservoir pressures are two key determinants of reservoir performance during the injection and immediate post-injection phases, and these have proved effective in verifying storage performance at the Sleipner, Ketzin and Rousse storage sites. 3D seismics and down-hole pressure measurements are proven technologies and have been key monitoring tools for reservoir management at these sites.
With regard to the three high-level criteria set by the EU Directive, CO2CARE developed few central findings and recommendations that should to be taken into consideration by the regulators when implementing and applying the Directive:
All leakage monitoring systems have a finite, site-specific, CO2 detection capability, so it is recommended that regulators use the term “no detectable leakage” in the context of whether a site is performing effectively in terms of both its greenhouse gas emissions mitigation function and health and safety requirements at the surface.
Demonstrating conformity between predictive models of reservoir performance and monitoring observations is technically challenging because a unique and perfect match is near-impossible to achieve. We recommend that regulators should set conformance criteria at realistic levels, focussing on progressive reduction of uncertainty with time and demonstration that the fundamental site-specific storage processes are understood.
Demonstrating that a site is moving towards a state of long-term stability is difficult due to the lack of long-term observational evidence from available storage projects. Predictive modelling is subject to significant uncertainty, so full use of additional analogue information is important to develop a logical case for site stabilization. Use should be made of monitoring data from sites already in the post-closure period (e.g. Nagaoka), experimental data and relevant geological analogues which demonstrate stabilization processes in similar circumstances and the time-scales on which they operate.
Project Results:
The research outcomes of the CO2CARE project will have a direct impact on European and overseas companies and institutions dealing with CO2 storage. Especially the close collaboration between scientific researchers, industry partners, and national authorities is promising. The distilled results of the project in the form of “Best Practice Guidelines” will play an important role in the CCS community worldwide. The Guidelines will not only support companies who are operating CO2 storage sites but also provide information for the regulators in the different countries in terms of the creation of specific requirements for the long term stability of CO2 storage complexes.
Progress beyond state-of-the-art
Geological storage of CO2 is a long-term process, the efficacy and safety of which needs to be assured in order for it to be accepted as a viable climate change mitigation option. The three main potential leakage paths for both the injection and post-closure periods are (a) abandoned wells and operational wells (b) the caprock through pre-existing or induced fractures or faults, and (c) undetected faults and fractures. For the purposes of post-closure risk assessment of CO2 storage, the EU Directive requires the site operators to investigate:
• the risk of fracturing the storage formation(s) and caprock,
• the risk of CO2 entry into the caprock,
• the risk of leakage from the storage site (through inadequately sealed wells),
• the rate of migration (in open-ended reservoirs),
• fracture sealing rates,
and establish remediation procedures for use in case of leakages or significant irregularities being observed. Pursuant to Articles 17 and 18, and Annex II of the Directive, the post closure criteria require the operators to utilise long-term monitoring data to revise the predictive models used and update the monitoring plans if necessary. This may also lead to an updated and a more advanced risk assessment at the time of site abandonment. The results by CO2CARE in this field notably enlarged the knowledge and progressed beyond state-of-the-art. By this, the results of CO2CARE will help to implement the European CO2 storage technology.
Wellbore integrity
Wellbores are recognised as one of the potential leakage pathways in geological storage of CO2. Important lessons regarding wellbore cement seal integrity can be learned from the oil industry, which has been conducting EOR for many years. However, while the oil industry is primarily concerned with the cement sheath integrity over the lifetime of a well (decades), geological storage of CO2 requires the consideration of a much longer time frame (hundreds to thousands of years). The primary concerns are that the Portland cements used to seal wellbore react readily with CO2, and the geomechanical/geochemical response of the wellbore to injection, abandonment and post-abandonment processes may compromise the integrity of the wellbore. Studies have shown that changes in downhole conditions can cause mechanical damage to the cemented annulus that may lead to a loss of zonal isolation. It is therefore important to have a firm understanding of the long-term behaviour of the complete mechanical system formed by the steel casing, the cement sheath, and the formation rocks. In terms of site abandonment, the crucial question is: under what downhole conditions may such an interface become progressively more transmissive? The initial width/quality and CO2-brine flux have been identified as the two key parameters controlling the long-term interface integrity. These important issues have been addressed through an integrated research programme encompassing laboratory testing, field installations and monitoring, as well as numerical simulations. Consistent with the above, research focus has progressively shifted to the complete system represented by the steel casing, the cement sheath, and the formation rocks. To take into account the long-term effect of CO2 on the cement sheath, the modelling approach has to consider the well on all its length anticipating the future behaviour of the well during the closure period and during the long-term post-closure one. However, previous studies suffer significant shortcomings in that the interface is synthetically created/manufactured and not necessarily representative of changes in real downhole conditions, particularly with respect to controlling the width and permeability of the interfaces created. CO2CARE performed newly designed laboratory experiments to overcome the disadvantages of current practice in this field and innovative well abandonment methodologies are introduced in the research programme. By these efforts, the results by CO2CARE will provide the operator with profound knowledge on wellbore integrity.
Reservoir and caprock integrity
The storage of CO2 in geological formations requires a caprock which forms a capillary seal to the injected CO2. Pressure depletion in a producing oil or gas reservoir, or CO2 injection into the reservoir or its subsequent cessation (at closure), modifies stress states such that newly created or existing fractures could open up, creating a connecting fracture network. In the context of post-injection (closure stage) risk assessment, the conductivity of induced fracture networks in the caprock around the wellbore have not been studied with uncertainty analysis so far. Furthermore, migration of CO2-rich fluid in porous rocks can cause mineral dissolution resulting in reduction of the rock strength. These processes are more intense in the rock joints and specially concentrated at joint extremities. The high dissolution ratio at these points will affect the rock toughness and can activate fracture propagation. The results by CO2CARE on the sub-critical crack propagation due to the geomechanical and geochemical processes and on the behaviour of storage integrity during the post-closure period will better facilitate long-term predictions on caprock and reservoir integrity.
Trapping mechanisms
Concerns and uncertainties at the post-closure and abandonment stage, particularly in terms of modelling, are mainly related to the longer-term storage processes which progressively immobilise the CO2 and lead to site stabilisation. These are primarily:
• Residual trapping by capillary forces.
• Solubility of CO2 in the brine leading to gravitational stabilization.
• Enhancement of dissolution by convection caused by brine density contrast.
• Geochemical fluid/mineral reactions leading to mineral fixing.
These mechanisms have been investigated mainly in isolation so far, but are recognized to interact. The dissolution process in a saline aquifer and the development of convection that enhances dissolution has been studied theoretically, but calibration data from pilot sites (e.g. Nagaoka) will be incorporated to support the numerical modelling. The contributions from each of these mechanisms are very much dependant on site specific geology and injection strategies. The results by CO2CARE integrating exhaustive experimental and modelling studies provide site-specific investigation of these mechanisms and their related impacts and thus will allow address these issues in the post-closure transfer process.
Monitoring methodologies
Seismic measurements (active methods and, to an increasing degree, passive monitoring) are a basic and important component in reservoir management. Repeated 2D or 3D seismic surveys serve for large scale monitoring of a reservoir in oil and gas production, town gas storage, and other applications such as CO2. While these surveys image structures and allow to infer petrophysical parameters from seismic attributes, the passive seismic monitoring aims at directly observing process related phenomena such as micro-seismicity or subtle but continuously occurring changes in the behaviour of seismic noise. Repeated 2D or 3D surveys are cost intensive and impose a strong impact on the local population and the landscape, especially onshore in populated areas, (e.g. by causing crop damage). Thus, this method is not an option for continuous monitoring with a high temporal resolution at a storage site in the abandonment phase. The monitoring concept studied within CO2CARE provides the framework for a cost efficient, (almost) non-invasive set of observations which offers a high temporal resolution. By this, one of the hurdles for implementation of CO2 storage will be overcome. Several systems exist which can be used as autonomous permanently deployed seismic sources for high resolution monitoring of a potentially problematic area of the reservoir but they have not yet been applied to CO2 storage, nor for comparison which is a prerequisite to assess their applicability to storage reservoir monitoring. At Ketzin, TNO has already deployed an array of buried continuously recording seismic receivers which forms an integral part of the seismic monitoring system tested.
An efficient application of monitoring methods during and after site abandonment is highly dependent on the best possible characterisation of the storage reservoir, based on data collected during the storage phase. These seismic and geoelectrical data allow a set-up up of a detailed structural model and to quantify the CO2 distribution within the reservoir.
While seismic measurements alone, even under very good conditions, are not able to provide a detailed quantification of the actual mass of CO2 the integrated interpretation of seismic and electromagnetic data is expected to enhance the potential for a quantitative interpretation due to the complementary sensitivity of both methods and the possibility to use cross property relations between electrical conductivity and elastic properties in an integrated tomographic approach.
Well established logging technologies from the oil and gas industry, such as Pulsed Neutron Gamma or Induction-Logs, are being verified in the field for CO2-exposed wells. These methods offer an instantaneous image of saturation at the time of the measurements. However, suitable technologies must be able to identify the saturation of CO2 in the reservoir and its change with time. In these respects, temperature measurements with fibre-optic cables combined with heat-pulse experiments have been recently proven to be successful in hydrology to detect water leakage at dams and offer a promising direction of development. Their employment in the detection of heat fluxes in wells relies, so far, on few numerical simulations and lab experiments. CO2CARE will assess this integrated system, which has been implemented at Ketzin. The temperature distribution along the wells has been monitored with high temporal and spatial resolution using fibre-optic sensor cables for distributed temperature sensing (DTS) in the “smart closed” well. These measurements enable the determination of local changes in fluid saturation within the reservoir in the immediate neighbourhood of the well by measuring the effective thermal conductivity of the formation and by this provide an additional monitoring tool to assess near-wellbore phenomena and well integrity problems.
Remediation
The hydrocarbons industry has many years of experience in applying gel treatments to reduce channelling in high-pressure gas floods and to reduce water production from gas wells. Referred to as relative permeability modification (RPM) or disproportionate permeability reduction (DPR) and water shut off (WSO) treatments, there are many examples of production performance modelling data for gel treated wells. Inspired by the successful field performance of polymer-gel solutions in modifying the permeability of both the rock matrix and fractures in oil and gas applications, CO2CARE has investigated the application of this technology and development of polymer-gel based remediation techniques to prevent leakage by creating a low permeability or an impermeable barrier above the caprock or fractured formations. Injection of saline water above the caprock, at a higher pressure than the CO2 pressure in the reservoir, would create an inverse pressure gradient to reverse the flow direction and increase the solubility of CO2 in the saline water barrier formed, and prevent or limit leakage. Such a procedure could enable fast and relatively low cost mitigation action once a leakage is detected. Furthermore, coupled with fluid management procedures during aquifer storage (saline water extraction and re-injection above the caprock), this methodology can also be used to minimise displacement and migration of native brine, and avoid pressure build up in closed or semi-closed structures. CO2CARE designed and modelled realistic CO2 storage reservoir fluid management scenarios for saline aquifers and depleted gas reservoirs and tested the effectiveness of pressure gradient reversal methods as a leakage mitigation/remediation technique. By these results, CO2CARE provide the operator with specific mitigation measures in case of leakage and by this fulfils one the regulatory obligations.
Risk management in CO2 storage site abandonment
The post-closure risk management procedures, the technology required to enable this and the use of corrective (remediation) measures that has been researched and developed in CO2CARE are pursuant to Articles 17 and 18 and Annexes I and II of the EU Directive and the OSPAR Framework for Risk Assessment and Management (FRAM). Considering the requirements of site closure and post-closure stages, the specific work carried out in CO2CARE has focussed on Tier 3 Risk Assessment. Additional work on Tier 2, particularly in relation to the development of appropriate leakage mitigation plan, will also be necessary depending on the status of the risk assessment work and the stage in the storage lifetime of the various sites. Prediction, assessment and optimisation of CO2 geological storage require robust performance assessment tools that can predict the fate of the CO2 in the subsurface in the long-term. The main objectives of such tools some of which have been developed across different disciplines and are being further developed currently consider the dominant physical, chemical and mechanical processes in different spatial scales and predict the fate of injected CO2 in the subsurface on the basis of such limited spatial and sometimes also temporal boundaries. In order to facilitate the purposes of risk management, namely to quantify the risks of individual scenarios for health, safety and the environment and identify appropriate risk management measures, it is essential to integrate such models over the different scales and, in particular, evaluate the significance of data, parametric and modelling uncertainties on risk estimates. Building on the idea of a Total System Approach for performance assessment of geological repositories robust models that investigate the full range of uncertainties and boundary conditions have been explored and developed in CO2CARE specifically for site abandonment.
Dissemination activities
The dissemination activities in CO2CARE focussed on different aspects which are described in the following:
Website: The Website was online in February 2011. New sections ( Multimedia, FAQ, Newsletters, Press Releases) were added during the second period to the project website. In the „Multimedia“-section, information films for almost all CO2 storage sites were added. Some information films were provided with English subtitles afterwards. A new section containing 10 FAQ was set up.
Scientific conferences: After the Kick-off meeting in January/February 2011 at GFZ Potsdam, three CO2CARE Annual Scientific Conferences took place: March 2012 - Imperial College, London; March 2013 - TNO, Utrecht; November 2013 - GFZ, Potsdam – final conference.
The so-called “Brainstorming Day on the long-term fate of geologically stored CO2” took place in June 2013 in Trondheim. The meeting was initiated and hosted by Statoil. The organising committee consisted of members of the projects “Mustang, Ultimate CO2, CO2CARE, CarbFix, and PANACEA”. Purpose of the conference was to debate key issues, innovations and best practices. Furthermore, CO2CARE issues were presented at several international conferences and meetings. (see A2 – List of Dissemination Activities).
Targeted Workshops: The first Targeted Workshop "Modelling and monitoring inputs for risk management and dry-runs elaboration" was organised by IFPEN and NERC in October 2012 at IFP Energies Nouvelles in Rueil-Malmaison - France. In addition to the workshop, IFPEN organised a public questionnaire on CCS, involving up to 36 students from IFP- school, who created some videos describing the meaning and the utility of CCS related to a social and economic context. These videos produced by different student-teams are accessible via the project website (“Multimedia”) and via YouTube. The second Targeted Workshop "Best Practice Guidelines" was organised by GFZ and NERC-BGS. In a meeting after the final CO2CARE conference, all WP leaders and consortium members discussed a draft of the document containing all the material collected by NERC-BGS.
Project brochures and the Best Practice Guidelines Summary: The first project brochure was created by BRGM with support by GFZ. This 4-page leaflet in English language mainly contains a short description of the project, some selected results, and a brief summary of what has been achieved so far and forthcoming outcomes. The second 8-page image brochure (M36) contains the summary of the three high-level site abandonment criteria (No Leakage, Predicted and Observed Performance and Long-term Stability) together with Risk Management. It describes the three main sites considered in the project and provides some samples of results from our research into well abandonment and reservoir management issues. Two schemes complete the brochure: a site-closure milestone chart leading to the transfer of responsibility according to the EU Storage Directive and a flow diagram of the traffic light system for risk-related decision. The English version of the brochure was generated with 1000 copies printed and distributed to the project partners and the European Commission. In addition, the brochure was translated into German, French and Italian by NERC, IFPEN, BRGM, OGS and GFZ and published in limited editions. All 4 versions are also available on the website in form of pdf in the section “Multimedia”.
The Best Practice Guidelines Summary was issued by NERC-BGS in English language. It was distilled and synthesised from the full Best Practice Guidelines document (Deliverable D5.4) into a coherent document. A final hi-resolution printer’s version was generated with 1000 copies printed and distributed to the project partners and the European Commission. The document is also available as pdf via the CO2CARE website.
Press Releases: In connection with the final CO2CARE conference, press releases were created by GFZ and translated by the partners into English, French, and Italian language. All versions were also uploaded to the website.
Contacts with media and outreach: Media training courses, mainly for press contacts, were attended by staff members of GFZ. In various interviews, Michael Kühn has given general information on CCS technology and in particular on the research work and progress at the pilot site Ketzin. Moreover, in Ketzin, an information centre, funded by the Federal Ministry of Education and Research (BMBF), was established for the public. CO2CARE results are demonstrated in the centre as well.
CO2CARE Newsletters: Six newsletters were issued between 2011 and 2013 on the CO2CARE website. These newsletters are freely available via the project`s website.
Scientific publications: Within the project`s lifetime, 23 full paper publications in scientific journals were issued (see section A). Three new publications are in preparation for 2014. Moreover, 14 abstract for international conferences are available, and 9 new abstracts are in preparation. During the final CO2CARE conference, 41 abstracts were issued in form of a conference volume. For 2014 a special issue on “Site Closure Assessment” is planned in collaboration with Greenhouse Gases: Science and Technology (Wiley-Blackwell). This new journal is covering the methods of carbon capture and storage, as well as utilisation of carbon dioxide as a feedstock for fuels and chemicals.
Potential Impact:
Potential impacts and use
Most of the results obtained in CO2CARE can be easily used by the community with some adaptation from case to case. They mainly consist in pragmatic recommendations based on laboratory and modelling results. In some cases, the key point could be to assess the field data that are required to run the proposed methodologies. At least most of the programme results has been published and are therefore available to the general public and stakeholders. The results will benefit to reducing risks by a better prediction of possible phenomena and by improving the definition of monitoring objectives. They support the transfer of responsibility of the sites and help to ensure a safe long term containment of (future) CO2 storage sites. In principle, the workflow can be also applied throughout the entire life-cycle of a storage project, whenever there is a mismatch between modelled and observed behaviour of sites and an irregularity in the storage operations occurs. The dry-run documents are expected to provide a widely used template for site abandonment and transfer of responsibility, based on real storage sites under a range of conditions and useful for both site operators and regulators. In the first instance the documents could be used at the site licencing stage to develop initial site closure plans and to inform early discussions and negotiations between the interested parties. Subsequently they could be used to inform the developing dialogue between site operator and regulator as the closure and post-closure phases are reached. The Best Practice Guidelines document will be of use in establishing effective protocols for site abandonment, useful for both site operators and regulators. Technical guidance is provided on how best to meet the three key requirements of the Storage Directive (viz. conformance, leakage and long-term stability) and also wider guidelines on post-closure wellbore and reservoir management and a robust protocol for risk management. In addition the Best Practice Guidelines contains recommendations directly relevant for the forthcoming review of the EU Storage Directive.
List of Websites:
http://www.co2care.org
List of beneficiaries with contact names
Beneficiary No. Beneficiaries name / website Contact persons
1 GFZ - Helmholtz Centre Potsdam, GFZ German Research Centre for Geosciences /Deutsches GeoForschungsZentrum, Potsdam, Germany,
Website: http://www.gfz-potsdam.de
Project coordinator: Axel Liebscher
Project manager: Mario Wipki
2 Imperial - Imperial College of Science, Technology and Medicine, London, Great Britain
Website: http://www3.imperial.ac.uk/earthscienceandengineering
Sevket Durucan
3 IFPEN-IFP Energies Nouvelles, Rueil Malmaison, France
Website: http://www.ifpenergiesnouvelles.fr/
Jean-Pierre Deflandre
4 TNO - Nederlandse Organisatie voor Toegepast Natuurwetenschappelijk Onderzoek, Utrecht, The Netherlands
Website: https://www.tno.nl/index
Jens Wollenweber
5 BGS- British Geological Survey (NERC), Great Britain
Website: http://www.bgs.ac.uk
Andy Chadwick
6 Istituto Nazionale di Oceanografia e di Geofisica Sperimentale (OGS), Sgonico ( TS ) – Italy
Website: www.ogs.trieste.it
Gualtiero Böhm
Stefano Picotti
7 Bureau de Recherches Géologiques et Minières (BRGM), Orléans Cedex 2 - France
Website:http://www.brgm.fr/brgm/CO2-en/default.htm
Hubert Fabriol
8 GEUS - Geological Survey of Denmark and Greenland, Denmark
Website: http://www.geus.dk/geuspage-uk.htm
Peter Frykman
9 Uppsala - Uppsala Universitet, Sweden
Website: http://www.uu.se/en/
Christopher Juhlin
10 AirLiquide - L'AIR LIQUIDE, Société Anonyme pour l'Etude et l'Exploitation des Procédés Georges Claude, France
Website: http://www.airliquide.com/
Claire Bourhy-Weber
11 RWE - RWE Power AG, Cologne, Germany
Website: www.rwe.com
Thomas Thielemann
12 Statoil Petroleum, Trondheim, Norway
Website: http://www.statoil.com/
Sveinung Hagen
13 Shell - Shell International BV, The Netherlands
Website: http://www.shell.com/
Claus Otto
14 Vattenfall - Vattenfall Research and Development AB, Germany, Sweden
Website: http://www.vattenfall.com/
Thomas Schulte
15 Veolia - Veolia Environnement Recherche & Innovation SNC, France
Website: http://www.veolia.com/en/
Natalia Quisel
16 AITF - Alberta Innovates – Technology Futures, Canada
Website: http://www.ucalgary.ca/research/funding/aitc
Stefan Bachu
17 UofA - University of Alberta, Canada
Website: http://www.ualberta.ca/
Rick Chalaturnyk
18 RITE - Research Institute of Innovative Technology for the Earth, Japan
Website: http://www.rite.or.jp/index_e.html
Ziqiu Xue
19 GCCC-BEG - The University of Texas System, USA
Website: http://www.beg.utexas.edu/gccc/
Susan Hovorka
20 LBNL - The Regents of the University of California, USA
Website: http://esd.lbl.gov/home/
Jens Birkholzer
21 PNNL - Battelle Memorial Institute Corporation, USA
Website: http://www.pnl.gov/
Alain Bonneville
22 CO2CRC - CO2CRC Management Pty Ltd, Australia
Website: http://www.co2crc.com.au/
Charles Jenkins
23 TOTAL S.A. - Total Exploration Production in Pau, France
Website: http://www.total.com
Cristophe Urbanczyk