System and method of using spatially independent subsets of data to determine the uncertainty of soft-data debiasing of property distributions for spatially correlated reservoir data

Patent metadata

Field	Value
Publication number	US-9355070-B2
Application number	US-201213678218-A
Country	US
Kind code	B2
Filing date	Nov 15, 2012
Priority date	Nov 15, 2011
Publication date	May 31, 2016
Grant date	May 31, 2016

Abstract

A system and method, implemented on a computer, for determining the uncertainty of soft-data debiasing of property distributions of spatially correlated reservoir data are provided. The method includes inputting, into the computer, soft-data at a plurality of spatial locations on a grid and hard-data at a plurality of location along well paths that intersect the grid. Each location of the hard-data is collocated with soft-data values of one or more types. The method further includes calculating, using the computer, a variogram of the hard-data and creating subsets of independent hard-data using the variogram and the soft-data values that are collocated with each of the hard-data location; applying, using the computer, a bootstrap process on each of the plurality of subsets of independent hard-data to obtain a plurality of bootstrap data sets from each of the plurality of subsets of independent hard-data; calculating soft-data debiased histograms for each of the obtained plurality of bootstrap data sets using a binary decision tree and the collocated soft-data; and characterizing the uncertainty of the soft-data debiased histograms.

First claim

What is claimed is:

1. A method, implemented on a computer, for determining the uncertainty of soft-data debiasing of property distributions of spatially correlated reservoir data, comprising:
inputting, into the computer, soft-data at a plurality of spatial locations on a grid and hard-data at a plurality of location along well paths that intersect the grid, wherein each location of the hard-data is collocated with soft-data values of one or more types;
calculating, using the computer, a variogram of the hard-data and creating subsets of independent hard-data using the variogram and the soft-data values that are collocated with each of the hard-data location;
applying, using the computer, a bootstrap process on each of the plurality of subsets of independent hard-data to obtain a plurality of bootstrap data sets from each of the plurality of subsets of independent hard-data;
calculating soft-data debiased histograms for each of the obtained plurality of bootstrap data sets using a binary decision tree and the collocated soft-data;
characterizing the uncertainty of the soft-data debiased histograms; and
using the soft-data debiased histograms and the uncertainty for a quantitative oil in place assessment.

2. The method according to claim 1 , wherein the soft data comprises seismic data, seismic attributes, or seismic inversion products, or any combination thereof, and the hard data comprises down-hole well logs of petrophysical properties.

3. The method according to claim 1 , wherein the soft-data, the hard-data, or both are weighted using weight factors to form a linear combination of soft-data or linear combination of hard-data or both.

4. The method according to claim 3 , wherein the weight factors are adjusted so that the correlation between the weighted hard-data and the collocated weighted soft-data is maximized so as to achieve a more linear relationship between the weighted hard-data and the weighted soft-data.

5. The method according to claim 1 , further comprising verifying that seismic impedance in the soft-data correlates with collocated well log impedance in the hard-data.

6. The method according to claim 1 , wherein calculating the soft-data debiased histograms comprises:
inputting, into the computer, minimum and maximum confidence levels;
selecting randomly, using the computer, a confidence level between the minimum and maximum confidence levels;
sorting, using the computer, all data by soft-data value and putting all the data in one bin;
determining, using the computer, a first split soft-data value that maximizes a difference in hard-data property values in the bin above the first soft-data split value and below the first soft-data split value;
if the difference in hard-data property values is statistically significant at the selected confidence level accept the first split soft-data value;
determining a second split soft-data value for a first sub-bin above the first split soft-data value that maximizes a difference in hard-data property values in a sub-bin within the first sub-bin above the second split soft-data value and below the second split soft-data value; and
determining a third split soft-data value for a second sub-bin below the first split soft-data value that maximizes a difference in hard-data property values in a sub-bin within the second sub-bin above the third split soft-data value and below the third split soft-data value; and
repeating the determining recursively until reaching the selected confidence level to form a binary tree decision tree of classification of bins.

7. The method according to claim 6 , further comprising:
applying the binary decision tree classification of bins to each cell in a plurality of cells on the grid to predict a probability distribution of hard-data in each cell to obtain probability distributions of hard-data associated with the plurality of cells; and
calculating the debiased soft-data histogram by a cell volume-weighted average of the probability distributions in each cell.

8. The method according to claim 1 , wherein applying the variogram comprises selecting a gamma value of the variogram according to the sample petrophysical data.

9. The method according to claim 8 , wherein selecting the gamma value comprises selecting a gamma value in a range between 0.3 and 1.

10. The method according to claim 9 , wherein selecting the gamma value comprises selecting a gamma value greater than 0.5.

11. The method according to claim 1 , wherein applying the bootstrap process comprises applying a resampling bootstrap process on each of the plurality of subsets of data.

12. The method according to claim 1 , further comprising:
selecting a statistical parameter for ranking statistical distributions;
ranking the histograms by using a selected statistical parameter to obtain ranked soft-data debiased histograms; and
characterizing the uncertainty of the soft-data debiased histograms based on the ranked soft-data debiased histograms.

13. The method according to claim 12 , wherein ranking the data distributions by using the selected statistical parameter comprises applying a vector-scalar transform to the soft-data histograms to obtain scalar values, each scalar value being associated with a histogram, and ranking the soft-data histograms according to the obtained scalar values.

14. The method according to claim 13 , wherein applying the vector-scalar transform comprises applying a mean function or a standard deviation function.

15. The method according to claim 14 , further comprising applying an ensemble averaging to a portion of the histograms that are ranked within a desired ranking ensemble averaging percentage.

16. A computer program product comprising a computer readable medium having instructions stored thereon when executed by a computer performs the method recited in claim 1 .

17. A computer system for determining the uncertainty of soft-data debiasing of property distributions of spatially correlated reservoir data, the system comprising:
a storage device configured to store soft-data at a plurality of spatial locations on a grid and hard-data along well paths that intersect the grid;
a processor configured to output a graphical user interface, the graphical user interface comprising a plurality of fields for inputting parameters, wherein the processor is configured to:
calculate a variogram of the hard-data and creating subsets of independent hard-data using a variogram;
apply a bootstrap process on each of the plurality of subsets of independent hard-data to obtain a plurality of bootstrap data sets from each of the plurality of subsets of independent hard-data;
calculate soft-data debiased histograms for each of the obtained plurality of bootstrap data sets using a binary decision tree;
characterize the uncertainty of the soft-data debiased histograms; and
use the soft-data debiased histograms and the uncertainty for a quantitative oil in place assessment.

18. The system according to claim 17 , wherein the soft data comprises seismic data, seismic attributes, or seismic inversion products, or any combination thereof, and the hard data comprises down-hole well logs of petrophysical properties.

19. The system according to claim 17 , wherein the soft-data, the hard-data, or both are weighted using weight factors to form a linear combination of soft-data or linear combination of hard-data or both.

20. The system according to claim 19 , wherein the weight factors are adjusted so that the correlation between the weighted hard-data

Assignees

• Thorne Julian
• Chevron Usa Inc

Inventors

• Thorne Julian

Classifications

• G01V2210/665

  using geostatistical modeling · CPC title

• G06F17/18 Primary

  for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)} · CPC title

• G01V2210/667

  Determining confidence or uncertainty in parameters · CPC title

• G01V11/00 Primary

  Prospecting or detecting by methods combining techniques covered by two or more of main groups G01V1/00 - G01V9/00 · CPC title

• G01V2210/6244

  Porosity · CPC title