Fault reactivation case study for probabilistic assessment of carbon dioxide sequestration

This preliminary investigation into carbon dioxide (CO2) sequestration in a Brazilian offshore oil field has the objective of assessing the potential risk of caprock failure for CO2 geological storage in deep aquifers. The technical evaluation of a carbon storage site requires the sustainable pore pressure to be defined. CO2 injection into a hydrocarbon reservoir or aquifer alters the stress locally and also in the surrounding rocks. This brings with it the risk of geological fault reactivation, which would create a likely path for the escape of CO2 to other rock layers, or even to the surface. Several methodologies have been developed to estimate fault reactivation pressures; however, even the most sophisticated solutions depend on data reliability. The aim of this study was to investigate the loss of caprock integrity due to fault reactivation by geological carbon sequestration in a real aquifer using a probabilistic approach that considers the variation of the material properties and the initial stress state. The probabilistic response of the model was evaluated using the integrated nonlinear finite element simulator AEEPEC2D (R) and NESSUS (R) probabilistic analysis software. The random variables in the probabilistic model are the average effective vertical stress gradient, elastic modulus, Poisson's ratio, the lateral earth pressure coefficient, and the cohesion, friction angle and thickness of the fault. The uncertainties in the properties both rock layers and faults were incorporated into the analysis using the mean value and advanced mean value plus methods. Geomechanical 20 finite element models of the formation were developed to evaluate the stress changes associated with pore pressure variation in cases where an induced fault slippage would be a threat to the integrity of the seal. The proposed methodology was applied to three realistic geological sections with several faults cutting the aquifer. The cumulative distribution curve of maximum admissible pore pressure variation was derived. In addition, the most influential variables for each fault were identified. (C) 2014 Elsevier Ltd. All rights reserved.