6.4.4 In Salah, Algeria

In the In Salah CO2 storage project, different Quantitative Risk assessments (QRA) methodologies have been conducted. Pre-injection risk highlighted the key risks and informed the baseline data acquisition programme and early monitoring (Mathiesson, 2012). Four methods were applied to this project (Dodds et al., 2011; Paulley et al., 2011): the RISQUE QRA process developed for CO2CRC (Bowden and Rigg, 2004); the Certification Framework (CF) (Oldenburg et al., 2009); the Quantitative Risk Through Time Analysis (QRTT), an approach developed within BP, and the FEPs approach.

  1. Based on the URS RISQUE method, the risk quotient for 'migration direction' was determined where a likelihood of the event was assessed a value of a 'possible' (0.01) to 'highly probable' (0.1) with a leakage rate of 200,000 to 250,000 T/yr. The leakage rate was based on future injection rates and modeled plume migration. It was assumed that, if this risk was to eventuate, there would be a delay in the detection of up to 5 years due to data acquisition, interpretation and implementation of a response strategy. Three response actions were evaluated to demonstrate the effect of differing responses on the risk quotient. The expert group evaluated the relative reduction in likelihoods and/or consequences that could reasonably be achieved through implementing three possible risk response actions. The dominant containment risk event seems to be migration direction, which would be considered to be an unacceptable risk since it exceeds the target risk for a single event by around one order of magnitude. All of the other containment risk events show risk levels that are more than one order of magnitude less than the target for individual events and are therefore COnsidered to pose an acceptable risk. Migration direction poses around two orders of magnitude more risk than the second highest event, well leakage. There are many potential processes that could allow this loss of containment by migration direction (e.g. uncertainty regarding the location and depth of the structural spill-point, possibility of a fractured reservoir) (Dodds et al., 2011).

  2. The Certification Framework (CF) was applied at two different stages in the state of knowledge of the project: (a) at the pre-injection stage, using data available just prior to injection around mid-2004; and (2) after four years of injection (September 2008) to be comparable to the other risk assessments. The main risk drivers for the project are CO2 leakage into potable groundwater and into the natural gas cap. The CF approach takes great care in defining boundaries of the storage region. Both well leakage and fault/fracture leakage are likely under some conditions, but overall the risk is low due to ongoing mitigation and monitoring activities. Results of the application of the CF during these different state-of-knowledge periods show that the assessment of likelihood of various leakage scenarios increased as more information became available, while assessment of impact stayed the same (Oldenburg et al., 2008). The overall CO2 Leakage Risk (CLR) as determined by the CF method is estimated as low for the In Salah Storage project at Krechba. The largest risk is to USDW by CO2 leakage into wells via poorly cemented annuli and a subsurface blowout via casing defects and available research indicates such an event has less than a 1% probability over the project life. However given the known poor seal integrity at several suspended legacy appraisal wells within the lease area, this probability is likely higher at Krechba (Dodds et al., 2011).

  3. The QRTT (Quantitative Risk Through Time) technique is an internal BP methodology that evaluates the relationship between the risk mechanisms for CO2 loss and the stochastically forecasted, changing dynamics of the storage system (i.e., formation pressure, fluid chemistry). The In Salah QRTT analysis was carried out over three pathways to represent the risk mechanics from the three injectors. The URS 2008 RISQUE risk outputs were used to populate the QRTT tool. To assign pressure dependency on the various risks, it was assumed that the likelihoods for relevant risk were judged at the maximum likely pressure that the risk mechanism would experience. The temporal risk analysis of the In Salah CO2 storage project is displayed as a series of risk curves for cumulative risk, overburden integrity, well integrity and lateral leakage. The temporal risk output shows that the heightened leakage risk for the project occurs during the operational (injection phase). The majority of risk is a consequence of the high injection pressure relative to the low permeability and small pressure window of operation for the In Salah Project. The key risk controlling this is migration direction. Well leakage risk is moderate through the 1,000 year risk period (Dodds et al., 2011).

  4. A Structured qualitative approach needed to support assessment has been applied to this industrial scale project at Krechba. A qualitative Performance Assessment (PA) framework was devised and implemented. The approach included identification of the FEPs that describe the Krechba system and its likely evolution. An 'expected evolution' scenario was then identified by systematically evaluating existing knowledge. Scenarios describing potential situations that could involve alternative evolution mechanisms were also identified; these included consideration of mechanisms that could in principal lead to containment failure. These scenarios can be analysed to show that they are either unlikely to occur and/or will be limited impact and so do not represent threats to adequate performance. After audit against Quintessa's freely available generic online CO2 FEP database to ensure demonstrate comprehensiveness, the site-specific scenarios identified and the associated list of remaining uncertainties, were used to prioritise future (e.g. systems modelling) work. The process was systematic, transparent and in line with guidance from documents concerning legislation and regulation. The outcomes have been used to identify uncertainties, prioritise ongoing work, including systems modelling approaches, and update the FEP and scenario descriptions (Paulley et al., 2011).