Methods for determining fluid volumes from NMR T1-T2 log data

1. A method of characterizing fluids in a subterranean formation traversed by a borehole, comprising:

a) measuring NMR data by a downhole NMR tool that is conveyed in the borehole that traverses the formation;

b) processing the NMR data of a) to derive T1-T2 maps at different depths of the formation using a processor executing instructions stored in a non-transitory, machine-readable medium;

c) processing, using the processor, the T1-T2 maps at different depths of the formation of b) to generate a cluster map of different fluids that are present in the formation, wherein the cluster map is a two-dimensional array of grid points in the T1-T2 domain with each grid point being assigned or classified to a specific fluid; and

d) using the processor, using the cluster map of different fluids that are present in the formation of c) to characterize properties of the formation.

2. A method according to claim 1 , wherein:

the operations of d) involve applying the cluster map of different fluids that are present in the formation of c) to a T1-T2 map corresponding to a particular depth of the formation to calculate fluid volumes for the different fluids that are present in the formation at the particular depth.

3. A method according to claim 1 , wherein:

the operations of d) involve applying the cluster map of different fluids that are present in the formation of c) to T1-T2 maps corresponding to a number of depths of the formation to calculate fluid volumes for the different fluids that are present in the formation at the number of depths.

4. A method according to claim 1 , wherein the operations of d) involve:

applying the cluster map of different fluids that are present in the formation of c) to at least one of T1-T2 distributions, T1 distributions, and T2 distributions corresponding to a depth in the formation to extract the respective T1-T2 distributions, T1 distributions or T2 distributions for the different fluids at the corresponding depth; and

using the extracted T1-T2 distributions, T1 distributions or T2 distributions of the different fluids at the corresponding depth for additional analysis and interpretation of the formation.

5. A method according to claim 1 wherein the operations of d) involve:

applying the cluster map of different fluids that are present in the formation of c) to at least one of T1-T2 distributions, T1 distributions, and T2 distributions corresponding to a number of depths in the formation to extract the respective T1-T2 distributions, T1 distributions or T2 distributions for the different fluids at the number of depths; and

using the extracted T1-T2 distributions, T1 distributions or T2 distributions of the different fluids as a function of depth for additional analysis and interpretation of the formation.

6. A method according to claim 1 , wherein:

the NMR data of a) and corresponding T1-T2 maps of b) characterize a particular interval of interest within the formation; and

the cluster map of different fluids that are present in the formation of c) is used to characterize properties of the interval of interest in the formation.

7. A method according to claim 1 , wherein the cluster map of different fluids that are present in the formation of c) is determined by:

constructing a matrix of T1-T2 maps from the T1-T2 maps of b);

performing a non-negative matrix factorization (NNMF) procedure on the matrix of T1-T2 maps to determine an initial cluster map for an initial guess of Nc different fluids; and

performing a hierarchical clustering process that involves non-negative matrix factorization (NNMF) procedure on the matrix of T1-T2 maps for different values of Nc to determine a cluster map for a solved Nc of different fluids dictated by a constraint of connectedness with regard to all clusters in the cluster map; and

assigning specific fluid types to the clusters of the cluster map that is determined by the hierarchical clustering process.

8. A method according to claim 7 , wherein:

the cluster map includes data elements that correspond to grid points in the matrix of T1-T2 maps, wherein the data elements for each cluster of the cluster map are assigned to a unique cluster identifier based on likelihood values of the associated grid points as determined from the non-negative matrix factorization (NNMF) procedure; and

the constraint of connectedness does not break any cluster as identified by a particular cluster identifier into multiple disconnected pieces.

9. A method according to claim 7 , wherein:

the matrix of T1-T2 maps is constructed in a reduced T1-T2 domain determined from analysis of an average T1-T2 map based on a number of T1-T2 maps of b) at different depths.

10. A method according to claim 7 , wherein the cluster map of different fluids that are present in the formation of c) is further determined by:

evaluating results of the hierarchical clustering process by sub-sampling NMR data multiple times.

11. A method according to claim 10 , wherein:

a subset of T1-T2 maps is taken for analysis at any given iteration.

12. A method according to claim 1 , wherein:

the downhole NMR tool comprises a wireline tool.

13. A method according to claim 1 , wherein:

the downhole NMR tool comprises a measuring-while-drilling tool.

14. A system for characterizing fluids in a subterranean formation traversed by a borehole, comprising:

a downhole NMR tool that is conveyed in the borehole that traverses the formation and measures NMR data;

a non-transitory, machine-readable medium storing code to perform one or more operations; and

at least one processor that is configured to perform the one or more operations, wherein the one or more operations comprise:

i) processing the NMR data to derive T1-T2 maps at different depths of the formation;

ii) processing the T1-T2 maps at different depths of the formation of i) to generate a cluster map of different fluids that are present in the formation, wherein the cluster map is a two-dimensional array of grid points in the T1-T2 domain with each grid point being assigned or classified to a specific fluid; and

iii) using the cluster map of different fluids that are present in the formation of ii) to characterize properties of the formation.

15. A system according to claim 14 , wherein:

the operations of iii) involve applying the cluster map of different fluids that are present in the formation of ii) to a T1-T2 map corresponding to a particular depth of the formation to calculate fluid volumes for the different fluids that are present in the formation at the particular depth.

16. A system according to claim 14 , wherein:

the operations of iii) involve applying the cluster map of different fluids that are present in the formation of ii) to T1-T2 maps corresponding to a number of depths of the formation to calculate fluid volumes for the different fluids that are present in the formation at the number of depths.

17. A system according to claim 14 , wherein the operations of iii) involve:

applying the cluster map of different fluids that are present in the formation of ii) to at least one of T1-T2 distributions, T1 distributions, and T2 distributions corresponding to a depth in the formation to extract the respective T1-T2 distributions, T1 distributions or T2 distributions for the different fluids at the corresponding depth; and

using the extracted T1-T2 distributions, T1 distributions or T2 distributions of the different fluids at the corresponding depth for additional analysis and interpretation of the formation.

18. A system according to claim 14 , wherein the operations of iii) involve:

applying the cluster map of different fluids that are present in the formation of ii) to at least one of T1-T2 distributions, T1 distributions, and T2 distributions corresponding to a number of depths in the formation to extract the respective T1-T2 distributions, T1 distributions or T2 distributions for the different fluids at the number of depths; and

using the extracted T1-T2 distributions, T1 distributions or T2 distributions of the different fluids as a function of depth for additional analysis and interpretation of the formation.

19. A system according to claim 15 , wherein:

the at least one processor or portion thereof is part of the downhole NMR tool.

20. The method of claim 1 , wherein processing the T1-T2 maps comprises generating the T1-T2 maps using unsupervised machine learning.