Predicting well markers from artificial neural-network-predicted lithostratigraphic facies

1. A computer-implemented method for predicting well markers, comprising:

separating neural-network (NN)-predicted facies output associated with a plurality of wells into two sets, a first set of NN-predicted facies output of training wells and a second set of NN-predicted facies output of target wells;

calculating, for each training well of the plurality of wells, a sameness score between zones of NN-predicted facies output and finer zones, wherein the finer zones are human-identified lithostratigraphic units;

calculating a mean sameness score for the finer zones for all training wells;

identifying finer zones with a mean sameness score greater than a threshold value as dominant facies zones; and

iterating, by a hardware processor, over each target well to calculate a top and a start depth position and an end depth position of each dominant facies zone determined based upon the NN-predicted facies output of the target well.

2. The method of claim 1 , further comprising, for the second set:

grouping and stacking scattered interbedded facies into finer zones for target wells; and

determining depth ratios of the finer zones following the grouping.

3. The method of claim 1 , further comprising:

calculating a mean coverage score for the finer zones for all training wells; and

calculating a mean of proportion ratio for finer zones for all training wells.

4. The method of claim 1 , further comprising, for finer zones with a mean sameness score less than or equal to the threshold value, grouping and stacking scattered interbedded facies into finer zones for target wells.

5. The method of claim 1 , further comprising:

feeding dominant facies detection statistics aspects for target well zones and well markers prediction into a memory; and

receiving dominant facies identified from the training wells and statistical characteristics of all finer zones in preparation for processing the target wells.

6. The method of claim 2 , further comprising calculating a sameness score and a coverage score between NN-predicted facies output and grouped and stacked finer zones for each target well, the grouped and stacked finer zones for each target well obeying a stacking order of finer zones corresponding to the training wells.

7. The method of claim 6 , further comprising moving a window of observation (WOO) to calculate a sameness score for each depth range associated with the NN-predicted facies output and grouped and stacked finer zones for each target well.

8. The method of claim 2 , further comprising determining a depth for remaining slots above, in-between, and below the calculated start depth positions and end depth positions of each dominant facies zone for each target well.

9. The method of claim 8 , further comprising, filling in zones in available slots above, below, and in-between each dominant facies zone for each target well, the filling of the zones proportional to the determined depth ratios of the finer zones.

10. A non-transitory, computer-readable medium storing computer-readable instructions for predicting well markers, the instructions executable by a hardware processor and configured to:

separate neural-network (NN)-predicted facies output associated with a plurality of wells into two sets, a first set of NN-predicted facies output of training wells and a second set of NN-predicted facies output of target wells;

calculate, for each training well of the plurality of wells, a sameness score between zones of NN-predicted facies output and finer zones, wherein the finer zones are human-identified lithostratigraphic units;

calculate a mean sameness score for the finer zones for all training wells;

identify finer zones with a mean sameness score greater than a threshold value as dominant facies zones; and

iterate, by a hardware processor, over each target well to calculate a top and a start depth position and an end depth position of each dominant facies zone determined based upon the NN-predicted facies output of the target well.

11. The medium of claim 10 , further comprising, for the second set, instructions to:

group and stack scattered interbedded facies into finer zones for target wells;

determine depth ratios of the finer zones following the grouping;

calculate a sameness score and a coverage score between NN-predicted facies output and grouped and stacked finer zones for each target well, the grouped and stacked finer zones for each target well obeying a stacking order of finer zones corresponding to the training wells; and

move a window of observation (WOO) to calculate a sameness score for each depth range associated with the NN-predicted facies output and grouped and stacked finer zones for each target well.

12. The medium of claim 10 , further comprising instructions to:

calculate a mean coverage score for the finer zones for all training wells; and

calculate a mean of proportion ratio for finer zones for all training wells.

13. The medium of claim 10 , further comprising, for finer zones with a mean sameness score less than or equal to the threshold value, instructions to group and stack scattered interbedded facies into finer zones for target wells.

14. The medium of claim 10 , further comprising instructions to:

feed dominant facies detection statistics aspects for target well zones and well markers prediction into a memory; and receive dominant facies identified from the training wells and statistical characteristics of all finer zones in preparation for processing the target wells.

15. The medium of claim 11 , further comprising instructions to:

determine a depth for remaining slots above, in-between, and below the calculated start depth positions and end depth positions of each dominant facies zone for each target well; and

fill in zones in available slots above, below, and in-between each dominant facies zone for each target well, the filling of the zones proportional to the determined depth ratios of the finer zones.

16. A computer-implemented system for predicting well markers, comprising:

at least one hardware processor interoperably coupled with a memory storage and configured to:

separate neural-network (NN)-predicted facies output associated with a plurality of wells into two sets, a first set of NN-predicted facies output of training wells and a second set of NN-predicted facies output of target wells;

calculate, for each training well of the plurality of wells, a sameness score between zones of NN-predicted facies output and finer zones, wherein the finer zones are human-identified lithostratigraphic units;

calculate a mean sameness score for the finer zones for all training wells;

identify finer zones with a mean sameness score greater than a threshold value as dominant facies zones; and

iterate, by a hardware processor, over each target well to calculate a top and a start depth position and an end depth position of each dominant facies zone determined based upon the NN-predicted facies output of the target well.

17. The system of claim 16 , further configured, for the second set, to:

group and stack scattered interbedded facies into finer zones for target wells;

determine depth ratios of the finer zones following the grouping;

calculate a sameness score and a coverage score between NN-predicted facies output and grouped and stacked finer zones for each target well, the grouped and stacked finer zones for each target well obeying a stacking order of finer zones corresponding to the training wells;

move a window of observation (WOO) to calculate a sameness score for each depth range associated with the NN-predicted facies output and grouped and stacked finer zones for each target well;

determine a depth for remaining slots above, in-between, and below the calculated start depth positions and end depth positions of each dominant facies zone for each target well; and

fill in zones in available slots above, below, and in-between each dominant facies zone for each target well, the filling of the zones proportional to the determined depth ratios of the finer zones.

18. The system of claim 16 , further configured to:

calculate a mean coverage score for the finer zones for all training wells; and

calculate a mean of proportion ratio for finer zones for all training wells.

19. The system of claim 16 , further configured, for finer zones with a mean sameness score less than or equal to the threshold value, to group and stack scattered interbedded facies into finer zones for target wells.

20. The system of claim 16 , further configured to:

feed dominant facies detection statistics aspects for target well zones and well markers prediction into a memory; and

receive dominant facies identified from the training wells and statistical characteristics of all finer zones in preparation for processing the target wells.