US 12,481,079 B2
Systems and methods for generating elastic property data as a function of position and time in a subsurface volume of interest
Yang Zhang, Sugar Land, TX (US); and Mark Allan Meadows, Danville, CA (US)
Assigned to CHEVRON U.S.A. INC., San Ramon, CA (US)
Filed by CHEVRON U.S.A. INC., San Ramon, CA (US)
Filed on Sep. 22, 2020, as Appl. No. 17/028,818.
Prior Publication US 2022/0091290 A1, Mar. 24, 2022
Int. Cl. G01V 1/28 (2006.01); G01V 1/30 (2006.01); G06F 30/27 (2020.01); G06N 20/00 (2019.01)
CPC G01V 1/282 (2013.01) [G01V 1/306 (2013.01); G06F 30/27 (2020.01); G06N 20/00 (2019.01); G01V 2210/6242 (2013.01)] 21 Claims
OG exemplary drawing
 
1. A computer-implemented method for training an elastic model to generate elastic property data as a function of position and time in a subsurface volume of interest, the method being implemented in a computer system that includes a physical computer processor and electronic storage, the method comprising:
obtaining, from the electronic storage:
a first training subsurface dataset comprising synthetic 4D subsurface data and a first set of subsurface energy values obtained at a first set of times from a first set of field data sensors located at a first set of positions, the synthetic 4D subsurface data having been generated by converting synthetic elastic property data based on a rock-physics model, and
a first set of elastic property values corresponding to the first training subsurface dataset;
obtaining, from the electronic storage, a first initial elastic model and a first set of elastic parameters based on a first set of elastic relationships between the first training subsurface dataset and the first set of elastic property values, wherein the first set of elastic parameters comprises a set of rock matrix coefficients and a geomechanical compaction coefficient, the first initial elastic model comprising a structure parameter configured to adjust the accuracy of the first initial elastic model, wherein the first set of elastic relationships comprises a mapping between a set of elastic properties and the first training subsurface dataset, and wherein the first set of elastic parameters is selected as a subset of the set of elastic properties based on the mapping;
determining (i) changes to the first set of subsurface energy values as a function of position and time and (ii) changes to the first set of elastic property values as a function of position and time, wherein position and time are determined from the first set of positions and the first set of times, respectively;
generating, with the physical computer processor, a first conditioned elastic model by training the first initial elastic model using the first training subsurface dataset and the changes to the first set of elastic property values as a function of position and time;
storing the first conditioned elastic model in the electronic storage;
obtaining, from the electronic storage, a first target subsurface dataset comprising a second set of subsurface energy values obtained at a second set of times from a second set of field data sensors located at a second set of positions in the subsurface volume of interest, and a second set of elastic property values corresponding to the second set of subsurface energy values;
determining changes to the second set of elastic property values as a function of position and time in the subsurface volume of interest;
generating, with the physical computer processor, a first target elastic property dataset based on the second set of subsurface energy values as a function of position and time in the subsurface volume of interest by applying the first conditioned elastic model to the first target subsurface dataset;
generating a representation of the first target subsurface dataset as a function of position and time in the subsurface volume of interest using visual effects to depict at least some of the first target subsurface dataset as a function of position and time; and
displaying the representation via a display.