Predictive Model for Relative Permeability Using Physics-Based Artificial Neural Networks

Hanif Farrastama Yoga, Prakash Purswani, Russell Taylor Johns

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Scopus citations

Abstract

Hysteresis of transport properties like relative permeability (Kr) can lead to computational problems and inaccuracies for various applications including CO2 sequestration and chemical enhanced oil recovery(EOR). Computational problems in multiphase numerical simulation include phase labeling issues and path dependencies that can create discontinuities. To mitigate hysteresis, modeling Kr as a state function that honors changes in physical parameters like wettability is a promising solution. In this research, we applythe state function concept to develop a physics-informed data-driven approach for predicting Kr in the spaceof its state parameters. We extend the development of the relative permeability equation-of-state (kr-EoS) to create a predictivephysics-based model using Artificial Neural Networks (ANN). We predict kr as a function of phasesaturation (S) and phase connectivity , as well as the specific path taken during the displacement, while maintaining other state parameters constant such as wet tability, pore structure, and capillary number. We use numerical data generated from pore-network simulations (PNM) to test the predictive capability of the EoS. Physical limits within space are used to constrain the model and improve its predictability you tside of the region of measured data. We find that the predicted relative permeabilities result in a smooth and physically consistent estimate. Our results show that ANN can more accurately estimate kr surface compared to using a high-order polynomial response surface. With only a limited amount of drainage and imbibition data with an initial phase saturation greater than 0.7, we provide a good prediction of kr from ANN for all other initial conditions, over the entire space. Finally, we show that we can predict the specific path taken in the space along with the corresponding kr for any initial condition and flow direction, which makes the approach practical when phase connectivity information is not available. This research demonstrates the first application of a physics-informed data-driven approach for prediction of relative permeability using ANN.

Original languageEnglish (US)
Title of host publicationSociety of Petroleum Engineers - SPE Improved Oil Recovery Conference, IOR 2022
PublisherSociety of Petroleum Engineers (SPE)
ISBN (Electronic)9781613998502
DOIs
StatePublished - 2022
Event2022 SPE Improved Oil Recovery Conference, IOR 2022 - Virtual, Online
Duration: Apr 25 2022Apr 29 2022

Publication series

NameProceedings - SPE Symposium on Improved Oil Recovery
Volume2022-April

Conference

Conference2022 SPE Improved Oil Recovery Conference, IOR 2022
CityVirtual, Online
Period4/25/224/29/22

All Science Journal Classification (ASJC) codes

  • Energy Engineering and Power Technology
  • Geotechnical Engineering and Engineering Geology

Fingerprint

Dive into the research topics of 'Predictive Model for Relative Permeability Using Physics-Based Artificial Neural Networks'. Together they form a unique fingerprint.

Cite this