Determination of mechanical properties of a geological formation using deep learning applied to data acquired while drilling

1. A method comprising:

receiving, by at least one processor, data acquired during drilling a geological formation;

generating, by the at least one processor, features of the data acquired during drilling, the features indicative of mechanical properties of the geological formation;

segmenting, by the at least one processor, the features of the data acquired during drilling into sequences readable by a trained temporal convolutional network (TCN);

predicting, by the at least one processor, subsequences of a synthetic geophysical log corresponding to the geological formation using the TCN based on the sequences obtained from the features of the data; and

outputting, by the at least one processor, a final synthetic log by combining the predicted subsequences.

2. The method of claim 1 , wherein the data acquired during drilling comprises at least one of a Gamma Ray log, a rate of penetration, a number of revolutions per minute, a weight on bit, a stand-pipe pressure, a hook height, a hook load, fluid flow values, a type of a drill bit, near-bit vibrations, a rotational drilling speed, a mud motor speed, a drilling torque, an area of the drill bit, a temperature of a drilling mud, a weight of the drilling mud, a nozzle diameter of the drill bit, or a number of cutters of the drill bit.

3. The method of claim 1 , wherein the mechanical properties comprise at least one of compressional sonic properties, shear sonic properties, density of the geological formation, porosity of the geological formation, an unconfined compressive strength, a Young's modulus, or a Poisson's ratio.

4. The method of claim 1 , further comprising:

performing, by the computer system, sensitivity analysis on the features of the data acquired during drilling; and

ranking, by the computer system, the features with respect to the mechanical properties of the geological formation based on the sensitivity analysis.

5. The method of claim 1 , further comprising capturing, by the computer system, geologic patterns using a convolutional inception model applied to the sequences obtained from the features of the data, wherein predicting the subsequences using the TCN is further based further on the geologic patterns.

6. The method of claim 1 , further comprising training, by the computer system, the TCN to generate the mechanical properties of the geological formation based on the sequences obtained from the features of the data.

7. The method of claim 1 , further comprising:

permuting, by the computer system, at least one feature of the features of the data acquired during drilling; and

determining, by the computer system, a change in the mechanical properties of the geological formation responsive to permuting the at least one feature.

8. A non-transitory computer-readable storage medium storing instructions executable by a computer system, the instructions when executed by the computer system cause the computer system to:

receive data acquired during drilling a geological formation;

generate features of the data acquired during drilling, the features indicative of mechanical properties of the geological formation;

segment the features of the data acquired during drilling into sequences readable by a trained temporal convolutional network (TCN);

predict subsequences of a synthetic geophysical log corresponding to the geological formation using the TCN based on the sequences obtained from the features of the data; and

output a final synthetic log by combining the predicted subsequences.

9. The non-transitory computer-readable storage medium of claim 8 , wherein the data acquired during drilling comprises at least one of a Gamma Ray log, a rate of penetration, a number of revolutions per minute, a weight on bit, a stand-pipe pressure, a hook height, a hook load, fluid flow values, a type of a drill bit, near-bit vibrations, a rotational drilling speed, a mud motor speed, a drilling torque, an area of the drill bit, a temperature of a drilling mud, a weight of the drilling mud, a nozzle diameter of the drill bit, or a number of cutters of the drill bit.

10. The non-transitory computer-readable storage medium of claim 8 , wherein the mechanical properties comprise at least one of compressional sonic properties, shear sonic properties, density of the geological formation, porosity of the geological formation, an unconfined compressive strength, a Young's modulus, or a Poisson's ratio.

11. The non-transitory computer-readable storage medium of claim 8 , wherein the instructions further cause the computer system to:

perform sensitivity analysis on the features of the data acquired during drilling; and

rank the features with respect to the mechanical properties of the geological formation based on the sensitivity analysis.

12. The non-transitory computer-readable storage medium of claim 8 , wherein the instructions further cause the computer system to capture geologic patterns using a convolutional inception model applied to the sequences obtained from the features of the data, wherein predicting the subsequences using the TCN is further based further on the geologic patterns.

13. The non-transitory computer-readable storage medium of claim 8 , wherein the instructions further cause the computer system to train the TCN to generate the mechanical properties of the geological formation based on the sequences obtained from the features of the data.

14. The non-transitory computer-readable storage medium of claim 8 , wherein the instructions further cause the computer system to:

permute at least one feature of the features of the data acquired during drilling; and

determine a change in the mechanical properties of the geological formation responsive to permuting the at least one feature.

15. A computer system comprising:

one or more computer processors; and

a non-transitory computer-readable storage medium storing instructions executable by the one or more computer processors, the instructions when executed by the one or more computer processors cause the one or more computer processors to:

receive data acquired during drilling a geological formation;

generate features of the data acquired during drilling, the features indicative of mechanical properties of the geological formation;

segment the features of the data acquired during drilling into sequences readable by a trained temporal convolutional network (TCN);

predict subsequences of a synthetic geophysical log corresponding to the geological formation using the TCN based on the sequences obtained from the features of the data; and

output a final synthetic log by combining the predicted subsequences.

16. The computer system of claim 15 , wherein the data acquired during drilling comprises at least one of a Gamma Ray log, a rate of penetration, a number of revolutions per minute, a weight on bit, a stand-pipe pressure, a hook height, a hook load, fluid flow values, a type of a drill bit, near-bit vibrations, a rotational drilling speed, a mud motor speed, a drilling torque, an area of the drill bit, a temperature of a drilling mud, a weight of the drilling mud, a nozzle diameter of the drill bit, or a number of cutters of the drill bit.

17. The computer system of claim 15 , wherein the mechanical properties comprise at least one of compressional sonic properties, shear sonic properties, density of the geological formation, porosity of the geological formation, an unconfined compressive strength, a Young's modulus, or a Poisson's ratio.

18. The computer system of claim 15 , wherein the instructions further cause the computer system to:

perform sensitivity analysis on the features of the data acquired during drilling; and

rank the features with respect to the mechanical properties of the geological formation based on the sensitivity analysis.

19. The computer system of claim 15 , wherein the instructions further cause the computer system to wherein the instructions further cause the computer system to capture geologic patterns using a convolutional inception model applied to the sequences obtained from the features of the data, wherein predicting the subsequences using the TCN is further based further on the geologic patterns.

20. The computer system of claim 15 , wherein the instructions further cause the computer system to train the TCN to generate the mechanical properties of the geological formation based on the sequences obtained from the features of the data.