Rationalizing the concept of net-to-gross with respect to reservoir heterogeneity and flow behavior utilizing machine learning analyses

Sheldon Gorell, Justin Andrews, Jim Browning

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Though seemingly straightforward, the concept of “net-to-gross” (NTG) is often a source of confusion. Its proper use is still being debated in some portions of the oil and gas industry. NTG is a method to account for non-reservoir quality rock when calculating oil volumes within a reservoir. This is normally accomplished by applying cutoffs to calculated quantities, such as porosity, which then get excluded from the volumetric calculation. To the extent there have been recent discussions of this, the focus has been primarily on how to determine appropriate cutoffs. There has been very little mention of the implications of using NTG in flow equations within a reservoir simulator. The paper discusses the derivation and implied assumptions for the simulator NTG formulation and possible errors and proposes modifications to account for inconsistencies. Resolving the NTG flow equations can be viewed as an upscaling problem, subject to implied assumptions about reservoir continuity. Many fine-scale reservoir simulations were run to test this and to calibrate the NTG equations. The underlying attributes were sampled from a bimodal distribution, which represent pay and non-pay. The results show the effects of NTG ratio, values of fine-scale attributes and spatial correlation on steady state, single phase effective permeability and immiscible flow displacements. They demonstrate errors in effective horizontal and vertical permeability when using NTG within a simulator. These errors cause potentially significant differences in production responses between underlying detailed fine-scale models and coarser models. The results demonstrate a possible need for corrections to the simulator net-to-gross formulations due to underlying implied assumptions and inconsistencies. Some possible modifications are also presented. Both standard and machine learning techniques were used to analyze the results.

Original languageEnglish
Title of host publicationSociety of Petroleum Engineers - SPE Eastern Regional Meeting 2019, ERM 2019
PublisherSociety of Petroleum Engineers (SPE)
ISBN (Electronic)9781613996768
DOIs
StatePublished - 2019
EventSPE Eastern Regional Meeting 2019, ERM 2019 - Charleston, United States
Duration: Oct 15 2019Oct 17 2019

Publication series

NameSPE Eastern Regional Meeting
Volume2019-October

Conference

ConferenceSPE Eastern Regional Meeting 2019, ERM 2019
CountryUnited States
CityCharleston
Period10/15/1910/17/19

Fingerprint Dive into the research topics of 'Rationalizing the concept of net-to-gross with respect to reservoir heterogeneity and flow behavior utilizing machine learning analyses'. Together they form a unique fingerprint.

Cite this