Evolve: A Linear Workflow for Quantifying Reservoir Uncertainty

Abstract

Abstract The EVOLVE workflow addresses a longstanding and difficult problem in reservoir modeling: quantifying the uncertainty in NPV forecasts ensuing from uncertainties related to geological and simulation parameters, forecast scenarios, and economic variables. The EVOLVE workflow is linear in nature composed of four key stages: (1) screening and selection of geological models; (2) calibration to field-scale measured data; (3) local calibration using well-level measured data; and (4) applying differing forecast scenarios on the resulting ensemble of models. NPV calculations are an extension of stage (4) as forecast responses are translated into NPV numbers using economic parameters such oil price, operating expenses (OPEX), and capital expenses (CAPEX). The novelty of the EVOLVE workflow is the repeated evolution and reduction of model ensembles in a cascading fashion as new parameter distributions are injected in the various stages (1) through (4) and existing parameter distributions modified, frozen, or removed. A particular characteristic of the EVOLVE workflow lies in its judicious use of geomodeling and flow simulation, the use of distances, distance-based generalized sensitivity analysis (dGSA), efficient calibration to measured field and well responses, and multidimensional scaling (MDS). These are powerful concepts and are united to form a unique workflow. The separation into four linear stages is deliberate and takes its cue from the favored approach geoscientists and engineers have used to address the uncertainty-problem in the past.