System and method of typing heterogeneous reservoir rocks

1. A method of rock typing comprising:

selecting a plurality of core samples from a region of interest;

obtaining, by a computer processor, mercury injection capillary pressure (MICP) data regarding the region of interest from the plurality of core samples using an MICP measurement device;

determining, by the computer processor, a plurality of distributions from the MICP data;

computing, by the computer processor, a distance matrix for the plurality of distributions using a statistical distance metric;

generating, by the computer processor and using the distance matrix, a cluster tree of the plurality of distributions;

performing a petrographic analysis of the plurality of core samples to identify a number m of rock types represented in the region of interest,

adjusting, by the computer processor, the cluster tree using as input the number m of rock types to produce an adjusted cluster tree; and

determining pore structure types of the region of interest based on the adjusted cluster tree.

2. The method of claim 1 , wherein the statistical distance metric is Wasserstein Distance.

3. The method of claim 2 , wherein generating, by the computer processor and using the distance matrix, a cluster tree of the plurality of distributions comprises generating the cluster tree by hierarchical agglomerative clustering.

4. The method of claim 1 , wherein adjusting, by the computer processor, the cluster tree based on the petrographic characteristic to produce the adjusted cluster tree comprises partitioning the cluster tree to produce a number of clusters that matches the number of rock types.

5. The method of claim 1 , wherein the plurality of distributions are MICP curves.

6. The method of claim 1 , wherein the plurality of distributions are pore throat distributions.

7. The method of claim 1 , further comprising outputting, by the computer processor, the adjusted cluster tree to a well-log-based rock typing.

8. The method of claim 1 , wherein the region of interest comprises a carbonate reservoir.

9. A method of rock typing comprising:

selecting a plurality of core samples from a region of interest;

acquiring mercury injection capillary pressure (MICP) data from MICP testing of the plurality of core samples;

obtaining a plurality of mercury saturation as a function of mercury injection pressure curves (MICP curves) from the MICP data;

computing a distance matrix for the MICP curves using Wasserstein distance as a distance metric;

generating a cluster tree of the MICP curves;

performing a petrographic analysis of the plurality of core samples to identify a number m of rock types represented in the region of interest;

extracting a number k of clusters from the cluster tree based on a number of m of rock types represented in the region of interest, wherein k>1, and wherein the number k matches the number m; and

outputting the clusters to a well-log-based rock typing to classify pore structure types of the region of interest.

10. The method of claim 9 , wherein generating the cluster tree of the MICP curves comprises generating the cluster tree by hierarchical agglomerative clustering.

11. A non-transitory computer-readable medium including one or more sequences of instructions that when executed by at least one processor to perform the method recited in claim 9 .

12. A non-transitory computer-readable medium including one or more sequences of instructions that when executed by at least one processor cause the at least one processor to:

obtain mercury injection capillary pressure (MICP) data from an MICP measurement device regarding a region of interest;

determine a plurality of distributions from the MICP data;

compute a distance matrix for the plurality of distributions using a statistical distance metric;

generate a cluster tree of the plurality of distributions using the distance matrix;

receive as input a number m of rock types obtained from a petrographic analysis of a plurality of core samples from the region of interest;

adjust the cluster tree based on the number m of rock types to produce an adjusted cluster tree; and

determine pore structure types of the region of interest based on the adjusted cluster tree.

13. The non-transitory computer-readable medium of claim 12 , wherein the at least one processor computes the distance matrix for the plurality of distributions using Wasserstein Distance as the statistical distance metric.

14. The non-transitory computer-readable medium of claim 13 , wherein the at least one processor generates the cluster tree of the plurality of distributions using the distance matrix and hierarchical agglomerative clustering.

15. The non-transitory computer-readable medium of claim 13 , further comprising the at least one computer processor outputting the adjusted tree to a well-log-based rock typing.