Reservoir Modeling – An Insider's History of a Key Enabling Technology

Abstract

As the capability of computers and software increased and cost significantly decreased in the early 1980’s, it did not take long for geological workstations and reservoir modeling software to become a key enabling technology for the industry. The initial tools included integrated and interactive applications that allowed geologists to generate cross sections, maps, and 3D reservoir property models with relative ease facilitated by databases that could be easily updated and revised. Early adopters were generally project teams working on large assets with hundreds to thousands of wells for whom the workstation environment provided a clear benefit in terms of efficiency, technical quality, and cross-discipline cooperation. The “cultural” gap between the geoscience and reservoir engineering disciplines began to shrink in the early 1990's as technology improvements enabled easy use of increasingly detailed 3D reservoir property models to be readily up-scaled for the dynamic models used by reservoir engineers to evaluate development options and generate production forecasts. The 1990’s also witnessed the rapid acceptance of the use of a variety geostatistical algorithms (e.g. kriging, conditional simulation, multiple-point modeling, object-based modeling, and process-mimicking modeling) to populate the increasing detailed reservoir models. The ability to generate very large and very detailed reservoir models gave rise to the still unresolved issue of how much model complexity is actually useful – an issue variously referred to as “fit-for-purpose” modeling or, somewhat divisively, as “Gilligan vs. Frankenstein” modeling. The incorporation of a variety of geostatistical algorithms also led to significant improvements in the industry’s assessment and use of uncertainty in reservoir development decisions. By the early 2000’s the reservoir modeling “toolkit” moved largely from proprietary software to vendor-provided software. This change significantly improved cooperation and decision making among private and national oil companies. In less than four decades, the industry reservoir modeling capability went from reservoir models with a few thousand grid cells with dimensions on the order of hundreds to thousands of feet to today’s reservoir models that may have up to a few billion cells (the so called “giga-cell" models) with grid dimensions of a few tens of feet or smaller.

AAPG Datapages/Search and Discovery Article #90323 ©2018 AAPG Annual Convention and Exhibition, Salt Lake City, Utah, May 20-23, 2018