US 12,461,275 B2
Identifying target regions in a cognitive reservoir system
Zackary H. Nolan, Los Angeles, CA (US); and Shahram Farhadi Nia, Glendale, CA (US)
Assigned to BEYOND LIMITS, INC., Glendale, CA (US)
Filed by BEYOND LIMITS, INC., Glendale, CA (US)
Filed on Sep. 1, 2022, as Appl. No. 17/901,629.
Application 17/901,629 is a continuation of application No. 16/157,764, filed on Oct. 11, 2018, granted, now 11,454,738.
Claims priority of provisional application 62/571,150, filed on Oct. 11, 2017.
Prior Publication US 2022/0413182 A1, Dec. 29, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. G01V 20/00 (2024.01); G06F 30/20 (2020.01); G06N 3/04 (2023.01); G06N 3/043 (2023.01); G06N 3/045 (2023.01); G06N 3/08 (2023.01); G06N 3/088 (2023.01)
CPC G01V 20/00 (2024.01) [G06F 30/20 (2020.01); G06N 3/04 (2013.01); G06N 3/043 (2023.01); G06N 3/045 (2023.01); G06N 3/08 (2013.01); G06N 3/088 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for modeling target regions of reservoir nodes, the method comprising:
receiving observed data points at a static modeler, the observed data associated with well trajectory data of a reservoir;
generating a plurality of three-dimensional (3D) populated logs corresponding to a 3D space associated with a volume of the reservoir using a neural network trained to propagate values to points across the volume, wherein the static modeler changes the points to generate each of the 3D populated logs;
generating a static model of the reservoir that includes clusters of one or more rock types in accordance with one or more of the values at each voxel representation of the 3D space;
transforming the static model into a dynamic model based on a dynamic modeler using artificial intelligence trained to identify cluster connectivity, the dynamic model including a graph representing the clusters as vertices that include a contiguous voxel set;
mapping a set of input features from the static model and the dynamic model associated with a node of the reservoir via a fuzzy inference engine, the set of input features mapped into linguistic variables; and
outputting a drilling assessment of the node associated with a target region of the reservoir based on the set of input features mapped using fuzzy logic, wherein the set of input features are represented within a continuum and evaluated against a set of one or more rules.