Claims
- 1. A method for identifying environmental effects in well log data, comprising:
selecting a plurality of models, each of the models including at least one parameter to be optimized; fitting each of the models to the well log data by optimizing the at least one parameter, the fitting producing a plurality of optimized models; and determining an indicator of goodness of fit for each of the plurality of optimized models.
- 2. The method of claim 1, further comprising selecting a model representing a best fit of the well log data based on the indicator.
- 3. The method of claim 2, further comprising generating a set of parameters representing the selected model.
- 4. The method of claim 1, wherein the well log data include information related to formation resistivity.
- 5. The method of claim 1, wherein the at least one parameter is for borehole dimension, invasion zone, dielectric property, formation anisotropy, boundary distance, or tool eccentricity.
- 6. The method of claim 1, further comprising generating a corrected data set in which environmental effects have been removed.
- 7. The method of claim 1, further comprising pre-processing the well log data.
- 8. The method of claim 7, wherein the pre-processing comprises one selected from the group consisting of tool failure check, bit size borehole correction, and eccentricity check.
- 9. The method of claim 1, wherein the indicator of goodness of fit is a function representing a root mean square of errors between the optimized models and the well log data.
- 10. The method of claim 1, further comprising post-processing.
- 11. The method of claim 10, wherein the post-processing comprises checking for consistence with additional petrophysical or geological constraints.
- 12. The method of claim 10, wherein the post-processing comprises generating a confidence indicator for the selected model.
- 13. The method of claim 12, wherein the confidence indicator is related to Confidence=Werr/Max[Err1,c1]+Wdif×Min[(Err1−Err2),c2] wherein Werr and Wdif are two weighting factors, Err1 is a least error among the plurality of optimized models, Err2 is a next least error among the plurality of optimized models, and c1 and c2 are two constants.
- 14. A system for identifying environmental effects in well log data, comprising:
a computer adapted to store a program, wherein the program includes instructions executable by the computer for selecting a plurality of models, each of the models including at least one parameter to be optimized; fitting each of the models to the well log data by optimizing the at least one parameter, the fitting producing a plurality of optimized models; and determining an indicator of goodness of fit for each of the plurality of optimized models.
- 15. The system of claim 14, wherein the program includes instructions for selecting a model representing a best fit of the well log data based on the indicator.
- 16. The system of claim 15, wherein the program includes instructions for generating parameters representing the selected model.
- 17. The system of claim 14, wherein the program includes instructions for generating a corrected data set in which environmental effects have been removed.
- 18. The system of claim 14, wherein the program includes instructions for preprocessing the well log data.
- 19. The system of claim 18, wherein the pre-processing comprises one selected from the group consisting of tool failure check, bit size borehole correction, and eccentricity check.
- 20. The system of claim 14, wherein the indicator of goodness of fit is a function representing a root mean square of errors between the each of the plurality of optimized models and the well log data.
- 21. The system of claim 14, wherein the program includes instructions for post-processing.
- 22. The system of claim 21, wherein the post-processing comprises checking for consistence with additional constraints.
- 23. The system of claim 21, wherein the post-processing comprises generating a confidence indicator for the selected model.
- 24. The system of claim 23, wherein the confidence indicator is related to Confidence=Werr/Max[Err1,c1]+Wdif×Min[(Err1−Err2),c2] wherein Werr and Wdif are two weighting factors, Err1 is a least error among the plurality of optimized models, Err2 is a next least error among the plurality of optimized models, and c1 and c2 are two constants.
CROSS REFERENCE TO RELATED dAPPLICATIONS
[0001] This invention claims benefit to U.S. Provisional Patent Application Serial No. 60/395,018, filed on Jul. 11, 2002. This Provisional Application is hereby incorporated by reference in its entirety.
Provisional Applications (1)
|
Number |
Date |
Country |
|
60395018 |
Jul 2002 |
US |