--> Production Forecasting: Improved Understanding of Why Sparse Data, Static and Dynamic Reservoir Modeling Limitations, and Human Bias Leads to Optimistic Recovery Forecasts

2019 AAPG Annual Convention and Exhibition:

Datapages, Inc.Print this page

Production Forecasting: Improved Understanding of Why Sparse Data, Static and Dynamic Reservoir Modeling Limitations, and Human Bias Leads to Optimistic Recovery Forecasts

Abstract

Reservoir production forecasts used to sanction project approvals are typically optimistic, sometimes significantly optimistic. Nandurdikar and Wallace’s 2011 SPE paper, which was based on a large number of project lookbacks, noted that the production shortfall for projects that were found to have reservoir-related “issues” such as optimistic OOIP or more than expected reservoir compartmentalization or heterogeneity typically produced only about 55% of the volumes projected at time of project sanction. A portion of the forecast optimism, perhaps 15-25%, may be explained by the impact of sparse data, particularly in the early phases of development when the number of wells is limited. The typical parameters used to build reservoir models may contribute 20-40% of the forecast optimism particularly if relatively coarse grids and/or significant horizontal and/or vertical upscaling is done prior to dynamic modeling. Well location optimization workflows may contribute 10-25% of the observed forecast optimism. Human biases such as the real or perceived need to move a project forward, likely contribute 30-40% to the observed forecast optimism. Mitigation of most of the mentioned sources that contribute to the observed production forecast optimism may be mitigated through better understanding of the impact of static and dynamic modeling parameters on the resulting forecast. For example: (1) Use the smallest possible grid cell size when building the initial geological model; (2) limit the amount that the geological model is upscaled as the dynamic model is constructed; and (3) consider the potential bias introduced as a result of the location of delineation/appraisal wells. Finally, the use of truly independent peer reviews may significantly reduce the impact of human bias, particularly in cases where there may be a “management-induced” bias to advance or approve a particular project. Note that the observations reported above are based on a large number of projects, particularly early development and mature fields undergoing waterflooding or steamflooding to maintain or improve production volumes.