STOCHASTIC INVERSION OF GEOPHYSICAL DATA FOR ESTIMATING EARTH MODEL PARAMETERS

• US 20100185422A1
• Filed: 01/20/2009
• Published: 07/22/2010
• Est. Priority Date: 01/20/2009
• Status: Active Grant

• Current Assignee: Chevron USA Incorporated (Chevron Corp.)
• Original Assignee: Chevron USA Incorporated (Chevron Corp.)
• Inventors: Hoversten, Gary Michael
• Granted Patent: US 8,095,345 B2
• Time in Patent Office: Days
• US Class Current: 703/2
• CPC Class Codes: G01V 11/00 Prospecting or detecting by methods combining techniques covered by two or more of main groups G01V1/00 - G01V9/00
• CPC Class Codes: G01V 2210/665 using geostatistical modeling

Abstract

A computer implemented stochastic inversion method for estimating model parameters of an earth model. In an embodiment, the method utilizes a sampling-based stochastic technique to determine the probability density functions (PDF) of the model parameters that define a boundary-based multi-dimensional model of the subsurface. In some embodiments a sampling technique known as Markov Chain Monte Carlo (MCMC) is utilized. MCMC techniques fall into the class of “importance sampling” techniques, in which the posterior probability distribution is sampled in proportion to the model'"'"'s ability to fit or match the specified acquisition geometry. In another embodiment, the inversion includes the joint inversion of multiple geophysical data sets. Embodiments of the invention also relate to a computer system configured to perform a method for estimating model parameters for accurate interpretation of the earth'"'"'s subsurface.

Claims

1. A computer implemented stochastic inversion method for estimating model parameters of an earth model of a subsurface geological volume of interest, the method comprising:

a) acquiring at least one geophysical data set that samples a portion of the subsurface geological volume of interest, each geophysical data set defines an acquisition geometry of the subsurface geological volume of interest;

b) generating a specified number of boundary-based multi-dimensional models of the subsurface geological volume of interest, said models being defined by model parameters;

c) generating forward model responses of the models for each specified acquisition geometry;

d) generating a likelihood value of the forward model responses matching the geophysical data set for each specified acquisition geometry;

e) saving the model parameters as one element of a Markov Chain for each model;

f) testing for convergence of the Markov Chains;

g) updating the values of the model parameters for each model and repeating b) to f) in series or in parallel, until convergence is reached;

h) deriving probability density functions for each model parameter of the models which form the converged Markov Chains;

i) calculating the variances, means, modes, and medians from the probability density functions of each model parameter for each model to generate estimates of model parameter variances and model parameters for the earth models of the subsurface geological volume of interest which are utilized to determine characteristics of the subsurface geological volume of interest.

2. The method of claim 1, wherein the multi-dimensional models are 2D, 3D or 4D.

3. The method of claim 1, wherein the model parameters are defined at nodes, and have values comprising an X,Y and Z value for defining the node location in space and at least one geophysical property value.

4. The method of claim 3, wherein model boundaries are defined by a unique collection of nodes.

5. The method of claim 3, wherein the at least one geophysical property value includes electrical conductivity or resistivity, compressional velocity, shear velocity, density, fluid saturations, pressure, temperature and/or porosity.

6. The method of claim 3, wherein the nodes define boundaries via interpolation of the geophysical property values in space between nodes to generate the model parameters required for the calculation of the forward model responses.

7. The method of claim 1, wherein the models are projected onto a finite-difference or finite-element mesh for calculation of the forward model responses.

8. The method of claim 1, wherein the geophysical data set includes controlled source electromagnetic data, magnetotelluric data, gravity data, magnetic data, seismic data, well production data or any combination of the foregoing.

9. The method of claim 1, wherein forward model responses for more than one type of geophysical data set are calculated to generate likelihood values for the combined geophysical data sets thus producing a joint inversion.

10. The method of claim 1, wherein Markov Chain Monte Carlo (MCMC) sampling is used to generate a sequence of model parameters that form a Markov Chain.

11. The method of claim 10, wherein the MCMC sampling algorithms include Metropolis-Hastings sampling and Slice Sampling algorithms.

12. The method of claim 1, wherein the sampling algorithm is utilized at each iteration of the updating the values of the model parameters for each model operation, and the sampling algorithm is determined by a random draw of a uniform variable wherein the probability of each sampling technique being used on any iteration is assigned at the start of the inversion.

13. The method of claim 1, wherein calculating the likelihood value of the forward model responses matching each of the geophysical data sets for the specified acquisition geometry is performed in accordance with a likelihood function.

14. The method of claim 1, wherein determining convergence of the Markov Chains is performed in accordance with a potential scale reduction factor:

15. The method of claim 1, wherein determining convergence of the Markov Chains computes the within-sequence variance and the between-sequence variance.

16. The method of claim 1, wherein the models defined by the model parameter means, modes and/or medians are used to define a starting model for deterministic (e.g. least squares) inversion to generate a model with the lowest root-mean-square data misfit.

17. The method of claim 1, wherein the models defined by the model parameter means, modes and/or medians, and/or the model parameter variances are graphically displayed.

18. A system configured to execute a computer readable medium containing a program which, when executed, performs an operation comprising:

a) acquiring at least one geophysical data set that samples a portion of the subsurface geological volume of interest, each geophysical data set defines an acquisition geometry of the subsurface geological volume of interest;

b) generating a specified number of boundary-based multi-dimensional models of the subsurface geological volume of interest, said models being defined by model parameters;

c) generating forward model responses of the models for each specified acquisition geometry;

d) generating a likelihood value of the forward model responses matching the geophysical data set for each specified acquisition geometry;

e) saving the model parameters as one element of a Markov Chain for each model;

f) testing for convergence of the Markov Chains;

g) updating the values of the model parameters for each model and repeating b) to f) in series or in parallel, until convergence is reached;

h) deriving probability density functions for each model parameter of the models which form the converged Markov Chains;

i) calculating the variances, means, modes, and medians from the probability density functions of each model parameter for each model to generate estimates of model parameter variances and model parameters for the earth models of the subsurface geological volume of interest which are utilized to determine characteristics of the subsurface geological volume of interest.

19. The method of claim 18, wherein the models defined by the model parameter means, modes and/or medians, and/or the model parameter variances are graphically displayed.