Model selection for CO2 sequestration using surface deformation and injection data

摘要: Abstract Incorporating time-lapse observations from sources such as remote sensing and seismic is important for monitoring the migration of CO 2 plumes in carbon capture storage projects. Significant computational costs make it impractical to run coupled flow-geomechanical simulations all plausible geologic scenarios a prior ensemble. This study presents model selection framework that integrates fast approximator proxy evaluate flow geomechanical responses models. The utilizes particle tracking algorithm mimic paths injected within Pressure changes rock deformation resulting injection are estimated using finite-element algorithm. reliability proposed tested by comparing simulation results with regards shapes extents plumes, reservoir pressure, vertical displacement at top layer given reservoir. Models showing similar grouped into clusters invoking multi-dimensional scaling followed k -means clustering. A representative each cluster selected, its dynamic evaluated running simulations. posterior ensemble consists models whose conforms observation data. approach applied history matching realistic channelized fractured inspired In Salah, Algeria. two case studies demonstrate incorporation surface data contributes reduction uncertainty improving quality well responses.