Reservoir Forecast Quality – Impact of Reservoir Modeling, Uncertainty Assessment of Sparse Data, and Decision Bias
The reservoir performance forecasts provided at the time of project approval and financial sanction are significantly optimistic. Published summary lookbacks (e.g. Nandurdikar and Wallace, 2011 and Rajvanshi et al., 2012) suggest that the industry as a whole produce only 75% of that forecast. These lookbacks suggest that projects with subsurface or reservoir “issues” produce only 55% of that forecast. Analysis of potential sources for the forecast optimism indicate that the largest contributors are due to sparse data and decision bias. Based primarily on several large Permian Basin and Middle East reservoirs and synthetic case histories 30-40% of the reservoir-associated forecast optimism can be attributed to the impact of sparse data and bias often associated with sparse data collection. A roughly equal amount of the optimism can be attributed to decision bias driven largely by a technical team consciously or unconsciously motivated by management desire to move a project forward. Other sources of optimism include reservoir modeling workflows and associated model parameters such as grid size and semivariogram parameters. With the exception of model grid size and perhaps some well optimization workflows, reservoir model related impacts including vertical upscaling appear to be small by comparison with the exception of gross modeling errors related model geometry/layering or permeability heterogeneity, for example. Forecast optimism can be reduced by appropriate statistical analysis, model parameter sensitivity studies, and truly independent project decision peer reviews particularly involving experienced technical staff who have detailed knowledge about the outcome of past projects in similar geological settings. Preservation of such knowledge maybe overlooked or minimized during industry down cycles.
AAPG Datapages/Search and Discovery Article #90291 ©2017 AAPG Annual Convention and Exhibition, Houston, Texas, April 2-5, 2017