Method for generating a 3D earth model

1. A computer implemented method for generating a 3D earth model for determining at least one characteristic of an area formed by a plurality of formations and comprising drilling positions of real or planned wells, the method comprising the steps of:

constructing, using a computer, an initial 3D earth model on the basis of measurement data and known properties of the area by combining solutions for a set of single 1D earth models with solutions for a set of 2D earth models, each of the 1D earth models corresponding to a drilling position of a real or a planned well and covering a number of formations along a bore of the respective well, and each of the 2D earth models covering a formation covered by at least one of the 1D earth models; and

optimizing, using the computer, the constructed initial 3D earth model by:

(i) determining an optimal set of formations for each 1D earth model by analyzing an effect of each of the formations along the bore on the characteristic being determined and removing formations weakly affecting the characteristic being determined by merging those formations with host formations, and

(ii) determining an optimal set of adjustable model parameters by analyzing an effect of 1D earth model parameters on the characteristic being determined and removing those parameters whose variations weakly affect the characteristic being determined.

2. The method as set forth in claim 1 , wherein the step of determining an optimal set of adjustable model parameters comprises removing the parameters whose variations weakly affect the characteristic being determined by fixing said parameters at their most likely values.

3. The method as set forth in claim 1 , wherein, in the step of constructing an initial 3D earth model, one or more of the 2D earth models contain structural elements corresponding to geological features of the formations covered by 2D earth models, and wherein the step of optimizing comprises analyzing the effect of the structural elements on the characteristic being determined and removing the structural elements weakly affecting the characteristic being determined.

4. The method as set forth in claim 1 , wherein the step of optimizing comprises successively removing the formations on the basis of a sensitivity analysis by merging the formations with host formations.

5. The method as set forth in claim 4 , wherein for each 1D earth model the step of successively removing formations comprises:

calibrating the 1D earth model by inverting measured well data;

calculating a sensitivity coefficient vector having a number of sensitivity coefficients equal to the number of formations along the bore of the respective well;

removing a formation corresponding to a lowest sensitivity coefficient of said sensitivity coefficients by merging said formation with a host formation; and

determining an error introduced by said removal; and

if the error is less than a predetermined modeling accuracy threshold, repeating the steps of calibrating the 1D earth model, calculating a sensitivity coefficient vector, removing a formation corresponding to a lowest sensitivity coefficient, determining an error introduced by said removal and checking if the error is less than a predetermined modeling accuracy threshold.

6. The method as set forth in claim 1 , wherein the step of optimizing comprises removing model parameters to determine the optimal set of adjustable model parameters on the basis of a sensitivity analysis.

7. The method as set forth in claim 6 , wherein said step of removing model parameters for each 1D earth model comprises:

calculating a sensitivity coefficient vector that has a number of sensitivity coefficients equal to a total number of model parameters for the optimal set of adjustable model parameters, and specifying an initial threshold value for the sensitivity coefficients;

removing the model parameters for which the corresponding sensitivity coefficients are less than the threshold value by fixing those parameters at their most likely values;

determining an error introduced by said removal; and

if the error is less than a predetermined modeling accuracy threshold, correcting the threshold value for the sensitivity coefficients until a permissible modeling error level is satisfied, and repeating the steps of calculating a sensitivity coefficient vector, removing the model parameters for which the corresponding sensitivity coefficients are less than the threshold value and determining an error introduced by said removal and checking if the error is less than a predetermined modeling accuracy threshold.

8. The method as set forth in claim 7 , wherein the step of optimizing further comprises determining, from the optimal sets of formations for the 1D earth models, those laterally hydrodynamically-coupled formations to be included into the combined 3D earth model by recognizing, on the basis of analysis of the sensitivity coefficient vector calculated for the model parameters, the formations for which efficient lateral conduction of a target solution is highly sensitive.

9. The method as set forth in claim 1 , wherein the characteristic being determined is overpressure, the formations for which the 2D earth models are constructed are laterally hydrodynamically-coupled formations, and any remaining formations are laterally water-impermeable.

10. The method as set forth in claim 9 , wherein the step of constructing an initial 3D earth model applies a basin time scale, the step comprising:

specifying coordinates of drilling positions of the real or the planned wells and model parameters for a set of 1D earth models for the wells;

obtaining solutions of forward problems for the 1D earth models on the basis of the specified coordinates and the model parameters;

on the basis of the solutions obtained for the 1D earth models, determining parameters corresponding to intervals of the laterally hydrodynamically-coupled formations in the 1D models;

obtaining solutions of forward problems for the 2D earth models on a regular grid by interpolating said model parameters on said grid; and

constructing an initial combined 3D earth model and obtaining a 3D forward problem solution by combining the obtained solutions of forward problems for the 1D earth models and the obtained solutions of forward problems for the 2D earth models on the regular grid in the basin time scale.

11. The method as set forth in claim 10 , wherein the step of constructing an initial 3D earth model comprises accounting for faults in the laterally hydrodynamically-coupled formations for which 2D earth models are constructed, and the step of optimizing further comprises analyzing an effect of fault segments on the overpressure being calculated and removing those segments weakly affecting the overpressure.

12. The method as set forth in claim 11 , wherein the fault segments are removed on the basis of a sensitivity analysis.

13. The method as set forth in claim 12 , wherein said step of removing the fault segments comprises:

calculating, for each 2D earth model and the respective laterally hydrodynamically-coupled formation, a sensitivity coefficient vector that has a number of sensitivity coefficients equal to a number of fault segments of said laterally hydrodynamically-coupled formation, and specifying an initial threshold value for the sensitivity coefficients;

removing those segments which have corresponding sensitivity coefficients that are smaller than the threshold value;

determining an error introduced by said removal; and

if the error is less than a predetermined modeling accuracy threshold, correcting the threshold value for the sensitivity coefficients until a permissible modeling error level is satisfied, and repeating the steps of calculating a sensitivity coefficient vector, removing those segments which have corresponding sensitivity coefficients that are smaller than the threshold value and determining an error introduced by said removal and checking if the error is less than a predetermined modeling accuracy threshold.

14. The method as set forth in claim 13 , wherein the fault segments are removed by successively removing nodes of the regular grid used for modeling the laterally hydrodynamically-coupled formation.

15. A method for predicting overpressure while drilling, comprising the steps of:

constructing a 3D earth model for an area in which drilling is performed, in accordance with a method of claim 9 ;

pre-calibrating the constructed 3D earth model by joint inversion of measured well data, on the basis of data obtained from calibration wells and obtaining a set of model parameters for the 1D earth model;

predicting overpressure at a target well position by obtaining a solution of a respective forward problem for the pre-calibrated 3D earth model;

determining an actual overpressure from the measured data obtained while drilling and checking a match between the determined and predicted overpressures;

if a mismatch between the predicted and the determined overpressures exceeds a predetermined threshold, updating the 3D earth model by obtaining a respective inverse problem solution and calculating from the updated 3D earth model an updated overpressure prediction below present drill bit position by obtaining a forward problem solution to the updated 3D earth model.

16. The method as set forth in claim 15 , further comprising the step of adjusting a drilling process when the actual overpressure approaches a hazardous level below the present drill bit position, as determined from the updated overpressure prediction.

17. The method as set forth in claim 15 , wherein the area comprises at least three calibration wells,

the pre-calibration of the constructed 3D earth model being performed on the basis of the measured data gathered in said calibration wells, using at least one inversion operator; and

the prediction of overpressure evolution comprising interpolating the set of model parameters obtained at the pre-calibration step for said calibration wells to the target well position and calculating an overpressure versus depth curve;

and wherein said method further comprises:

continuously checking the match between the calculated overpressure curve and an overpressure curve determined using measured data obtained in the drilling process and determining a mismatch between the calculated and the determined overpressure curves;

if the mismatch between the calculated and the determined overpressure curves exceeds a predetermined threshold, updating the 3D earth model by re-calibration to minimize said mismatch, the mismatch being used as an input to the inversion operator, wherein an updated calculated overpressure curve calculated from the updated 3D earth model is used to update the overpressure prediction below the present drill bit position.

18. The method as set forth in claim 17 , wherein the calculation of the overpressure versus depth curve further comprises calculating curves which restrict a range of permissible overpressure variability between maximal and minimal scenarios.

19. The method as set forth in claim 17 , wherein the overpressure prediction is further updated on the basis of lithologic and stratigraphic data gathered while drilling and related to a current formation by updating the model parameters relevant to the target well on the basis of the data.

20. A system for predicting overpressure during a drilling process in accordance with the method of claim 17 , comprising:

at least one data acquisition device for gathering data while drilling;

a computing device for performing calculations in accordance with the method and taking account of data gathered by said at least one data acquisition device, said computing device including:

difference signal generation means for generating a difference signal comprising an indication of a mismatch between the calculated overpressure curve and the overpressure curve determined using measured data obtained in the drilling process; and

prediction updating means for real-time updating of the overpressure prediction below the present drill bit position, said prediction updating means being enabled in response to said difference signal and adapted to recalibrate the 3D earth model in real-time by inputting said difference signal into the inversion operator, and to update the overpressure prediction by calculations on the basis of the re-calibrated 3D earth model;

a control signal generation means for generating a control signal containing recommendations and instructions for adjusting the drilling process responsive to results of the real-time updating of the overpressure prediction below the present drill bit position performed by said computing device.

21. The system as set forth in claim 20 , wherein data gathered by said at least one data acquisition device is further processed by the computing device to represent the data in a form most suitable for 3D earth model updating.

22. The system as set forth in claim 20 , wherein said control signal generation means transmits the generated control signal to a local drilling process control means.

23. The system as set forth in claim 20 , wherein said control signal generation means transmits the generated control signal to a remote drilling process control means.

24. The system as set forth in claim 20 , wherein said control signal generation means is integrated into the computing device.