Geological storage of CO2: A statistical approach to assessing performance and risk

Probabilistic and deterministic risk assessments were carried out by ECOMatters Inc. for the IEA Weyburn CO2 Monitoring and Storage Project (the “IEA Weyburn Project”). The objective of this work was to understand and evaluate the geological storage of CO2 from a risk assessment perspective within the context of a large enhanced oil recovery (EOR) project being carried out in the Weyburn field, which is part of the Williston Basin that straddles the Canadian (Saskatchewan) and USA (North Dakota) borders. Probabilistic Risk Assessment (PRA) is the preferred methodology for evaluating complex, long timeframe, process-driven problems such as the geological storage of CO2. A unique, computationally-efficient model CQUESTRA-1 (CQ-1) was developed to rapidly assess the 1000’s of cases required for a PRA, that statistically quantify the uncertainty associated with the many features, events and processes including their interactions over the long-term geological storage of CO2. This paper focuses on the application of the PRA methodology to the Weyburn field and illustrates its use with data from an operating EOR pattern. Sensitivity analysis is used to identify those parameters that most influence (positively or negatively) releases to the biosphere and storage of CO2 in the geological formations surrounding the reservoir. Deterministic calculations can also be carried out using CQ-1. Systematic calculations can be used in determining trends due to single parameter variation or best or worst case scenarios using combinations of parameters identified from the PRA sensitivity analysis. “What-if” or systematic studies of performance (e.g., leakage of CO2 to the biosphere, effectiveness of the aquifers, aquitards and well components, ...) are, therefore, easy and instructive to assess. These studies could be used to develop site selection criteria for CO2 storage in any geological media or evaluate mitigative strategies, with regard to both technique and timeframe, for well abandonment. The use of CQ-1 in a deterministic or RA mode is illustrated here with a base case study that uses the current best estimate of parameter values based on field or laboratory data and/or expert opinion.