Subterranean formation properties prediction

1. A method for predicting subterranean formation properties of a wellsite, comprising:

obtaining seismic data for an area of interest;

obtaining, using a computer processor, an initial seismic cube using the seismic data, wherein the initial seismic cube is a three-dimensional representation of the seismic data;

generating, using the computer processor, a plurality of shifted seismic cubes within the area of interest using the seismic data and a shifting parameter,

wherein each of the plurality of shifted seismic cubes is shifted from the initial seismic cube, and

wherein the shifting parameter defines a direction and a range that the initial seismic cube should be shifted;

generating, using the computer processor, a neural network using the initial seismic cube, the plurality of shifted seismic cubes, and well log data; and

applying the neural network to the seismic data to obtain a model for the area of interest, the model being configured for use in adjusting an operation of the wellsite.

2. The method of claim 1 , wherein generating the neural network comprises:

obtaining a plurality of seismic well logs using the initial seismic cube and the plurality of shifted seismic cubes;

obtaining the well log data associated with the area of interest;

generating the neural network, wherein the neural network comprises the plurality of seismic well logs and the well log data; and

training the neural network using the well log data to generate a weight matrix, wherein the weight matrix comprises a plurality of weights associated with the plurality of seismic well logs.

3. The method of claim 2 , wherein the weight matrix is generated using a back propagation algorithm and the well log data is used as a stopping criterion in the back propagation algorithm.

4. The method of claim 2 , wherein generating the neural network further comprises decreasing a size of at least one of the plurality of weights based on a relative importance of each of the at least one of the plurality of weights.

5. The method of claim 1 , wherein the model corresponds to an acoustic impedance model.

6. The method of claim 1 , wherein the operation is one selected from a group consisting of an exploration operation, a drilling operation, and a production operation.

7. The method of claim 1 , wherein the neural network describes a non-linear relationship between the seismic data and a geologic property of the area of interest.

8. The method of claim 1 , wherein the direction is at least one selected from a group consisting of a vertical direction and a horizontal direction.

9. A system for predicting subterranean formation properties of a wellsite, comprising:

a processing module configured to obtain seismic data for an area of interest;

a modeling unit configured to:

obtain an initial seismic cube using the seismic data, wherein the initial seismic cube is a three-dimensional representation of the seismic data; and

shift the initial seismic cube to generate a plurality of shifted seismic cubes within the area of interest based on the seismic data and a shifting parameter,

wherein each of the plurality of shifted seismic cubes is shifted from the initial seismic cubes, and

wherein the shifting parameter defines a direction and a range that the initial seismic cube should be shifted;

a training module configured to:

generate a neural network using the initial seismic cube, the plurality of shifted seismic cubes, and well log data; and

apply the neural network to the seismic data to obtain a model for the area of interest, the model being configured for use in adjusting an operation of the wellsite.

10. The system of claim 9 , wherein:

the modeling unit is further configured to obtain a plurality of seismic well logs using the initial seismic cube and the plurality of shifted seismic cubes;

the processing module is further configured to obtain the well log data associated with the area of interest; and

the training module is further configured to:

generate the neural network, wherein the neural network comprises the plurality of seismic well logs and the well log data; and

train the neural network using the well log data to generate a weight matrix, wherein the weight matrix comprises a plurality of weights associated with the plurality of seismic well logs.

11. The system of claim 10 , wherein the weight matrix is generated using a back propagation algorithm and the well log data is used as a stopping criterion in the back propagation algorithm.

12. The system of claim 10 , wherein the training module is further configured to decrease a size of at least one of the plurality of weights based on a relative importance of each of the at least one of the plurality of weights.

13. The system of claim 9 , wherein the model corresponds to an acoustic impedance model.

14. The system of claim 9 , wherein the operation is one selected from a group consisting of an exploration operation, a drilling operation, and a production operation.

15. The system of claim 9 , wherein the direction is at least one selected from a group consisting of a vertical direction and a horizontal direction.

16. A computer program product, embodying instructions executable by a computer to predict subterranean formation properties of a wellsite, the instructions comprising functionality to:

obtain an initial seismic cube using the seismic data, wherein the initial seismic cube is a three-dimensional representation of seismic data;

generate a plurality of shifted seismic cubes within an area of interest using the seismic data and a shifting parameter,

wherein each of the plurality of shifted seismic cubes is shifted from the initial seismic cube, and

wherein the shifting parameter defines a direction and a range that the initial seismic cube should be shifted;

train a neural network comprising the initial seismic cube and the plurality of shifted seismic cubes based on well log data; and

apply the neural network to the seismic data to obtain a model for the area of interest, the model being configured for use in adjusting an operation of the wellsite.

17. The computer program product of claim 16 , the instructions further comprising functionality to:

obtain a plurality of seismic well logs using the initial seismic cube and the plurality of shifted seismic cubes;

obtain the well log data associated with the area of interest;

train the neural network using the well log data to generate a weight matrix, wherein the weight matrix comprises a plurality of weights associated with the plurality of seismic well logs.

18. The computer program product of claim 17 , wherein the weight matrix is generated using a back propagation algorithm and the well log data is used as a stopping criterion in the back propagation algorithm.

19. The computer program product of claim 17 , wherein generating the neural network further comprises decreasing a size of at least one of the plurality of weights based on a relative importance of each of the at least one of the plurality of weights.

20. The computer program product of claim 16 , wherein the direction is at least one selected from a group consisting of a vertical direction and a horizontal direction.