US 12,140,725 B2
Determining reservoir formation properties
Marwah Mufid AlSinan, Al Qatif (SA); Xupeng He, Thuwal (SA); Hyung Tae Kwak, Dhahran (SA); Hussein Hoteit, Thuwal (SA); and Yiteng Li, Thuwal (SA)
Assigned to Saudi Arabian Oil Company and King Abdullah University of Science and Technology, Dhahran (SA)
Filed by Saudi Arabian Oil Company, Dhahran (SA); and King Abdullah University of Science and Technology, Thuwal (SA)
Filed on Dec. 15, 2022, as Appl. No. 18/082,017.
Prior Publication US 2024/0201413 A1, Jun. 20, 2024
Int. Cl. G01V 3/38 (2006.01); G01V 3/14 (2006.01); G01V 5/12 (2006.01); G06F 30/27 (2020.01); G06N 3/0442 (2023.01)
CPC G01V 3/38 (2013.01) [G01V 3/14 (2013.01); G01V 5/12 (2013.01); G06F 30/27 (2020.01); G06N 3/0442 (2023.01)] 33 Claims
OG exemplary drawing
 
1. A computer-implemented method for determining one or more reservoir formation properties, comprising:
generating, with one or more hardware processors, a plurality of digital models of a plurality of core samples taken from one or more reservoir formations based on a corresponding plurality of images of the plurality of core samples;
determining, with the one or more hardware processors, a pore throat size distribution of the plurality of core samples based on the plurality of digital models;
determining, with the one or more hardware processors, corresponding capillary pressure curves and corresponding nuclear magnetic resonance (NMR) value distributions of the plurality of core samples with one or more numerical simulations;
generating, with the one or more hardware processors, one or more machine-learning (ML) models based on the pore throat size distribution, the corresponding capillary pressure curves, and the corresponding NMR value distributions of the plurality of core samples;
adjusting, with the one or more hardware processors, the one or more ML models with one or more reservoir data of the one or more reservoir formations;
generating, with the one or more hardware processors, adjusted capillary pressure curves and adjusted NMR value distributions from the one or more adjusted ML models; and
determining, with the one or more hardware processors, a reservoir formation specific pore throat size distribution from at least one of the adjusted capillary pressure curves or the adjusted NMR value distributions.