Abstract
A combining mechanism for borehole logging tool data that employs modality merging to combine the output data of various borehole logging tools to provide a combined result and automated interpretation is provided, said mechanism comprising: at least one mechanism for assigning interpretive values to individual processed data types; at least one mechanism for combining the interpretive value data sets; and, at least one mechanism for providing an interpretation. A method of combining borehole logging tool data that employs modality merging to combine the output data of various borehole logging tools to provide a combined result and automated interpretation is also provided, said method comprising: assigning interpretive values to individual processed data types; combining the interpretive value data sets; and, providing an interpretation.
Claims
1. A combining mechanism for borehole logging tool data that employs modality merging to combine the output data of various borehole logging tools to provide a combined result and automated interpretation, comprising: at least one mechanism for assigning interpretive values to individual processed data types; at least one mechanism for combining the interpretive value data sets; and at least one mechanism for providing an interpretation.
2. The combining mechanism of claim 1, wherein said mechanism is configured to process ultrasound or acoustic data from a borehole logging tool.
3. The combining mechanism of claim 1, wherein said mechanism is configured to process x-ray data from a borehole logging tool.
4. The combining mechanism of claim 1, wherein said mechanism is configured to process neutron porosity data from a borehole logging tool.
5. The combining mechanism of claim 1, wherein said mechanism is configured to process neutron activation data from a borehole logging tool.
6. The combining mechanism for borehole logging tool data of claim 1, wherein said borehole tool is a wireline-based tool.
7. The combining mechanism for borehole logging tool data of claim 1, wherein said borehole tool is a logging-while-drilling-based tool.
8. The method of combining borehole logging tool data that employs modality merging to combine the output data of various borehole logging tools to provide a combined result and automated interpretation, comprising: assigning interpretive values to individual processed data types; combining the interpretive value data sets; and providing an interpretation.
9. The method of claim 8, wherein said method further comprises processing ultrasound or acoustic data from a borehole logging tool.
10. The method of claim 8, wherein said method further comprising processing x-ray data from a borehole logging tool.
11. The method of claim 8, wherein said method further comprises processing neutron porosity data from a borehole logging tool.
12. The method of claim 8, wherein said method further comprises processing neutron activation data from a borehole logging tool.
13. The method for borehole logging tool data of claim 8, wherein said borehole tool further comprises wireline-based tool.
14. The method for borehole logging tool data of claim 8, wherein said borehole tool further comprises a logging-while-drilling-based tool.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0032] FIG. 1 illustrates an example ultrasonic wellbore tool combined with an x-ray-based wellbore tool being lowered into a well by means of wireline conveyance.
[0033] FIG. 2 illustrates an example of how ultrasound data may be combined with x-ray data using a modality merging technique to provide a better interpretation than both data-types being examined separately.
[0034] FIG. 3 illustrates an example of how ultrasound data may be combined with x-ray data and neutron porosity data using a modality merging technique to provide a better interpretation than all three data-types being examined separately.
BRIEF DESCRIPTION OF SEVERAL EXAMPLE EMBODIMENTS
[0035] The methods described herein use the outputs of various log data provided from borehole logging tools, which have been deployed in the same well, and combine them in such a way that the combination of the log data produces a more accurate interpretation of the data as compared to interpretation of each data log in a side-by-side comparison.
[0036] With reference now to FIG. 1, in one example embodiment an ultrasonic logging tool [102] is accompanied by an x-ray cement evaluation and/or neutron porosity tool [102] by wireline conveyance [103] into a cased borehole, wherein the cemented section of the well [104] is logged through the inner-most casing or tubing [105]. The cemented section of the well [104] can contain discontinuities/anomalies [106] that can adversely affect well integrity.
[0037] FIG. 2 illustrates how inverted and processed ultrasound images or logs [201], are further processed as a function of depth in order to assign an interpretive modal value to the data [202] as a function of log depth, which reflects the quality of the bond and the acoustic impedances (relating to density) within the cement mass surrounding the casing. For example, an assigned value of “0” would indicate a perfect cement bond, and a highly homogeneous cement volume attached to the casing at that depth, whereas a value of “4” indicates that there is no cement bond whatsoever in that region. Values in between would be based upon the analysis of the acoustic/ultrasound log(s) at that depth such that “1” could indicate a good cement bond having issues within the cement mass are noted, and “3” could indicate a nominal cement bond having significant issues in the cement mass. Processed three-dimensional x-ray density data [203] is further processed to assign an interpretive modal value to the data [204], as a function of log depth, which reflects the completeness/homogeneity of the cement mass and the likelihood of major anomalies or channels within the cement. For example, an assigned value of “0” would indicate a perfectly homogeneous and isotropic cement density distribution within the cement volume located in the annulus, whereas a value of “4” would be representative of significantly missing cement in the annulus, or the indication of a continuous channel through the cement volume within that interval/depth-range. An assigned value of “2” could indicate a shallow anomaly (close to the casing), whereas “3” may indicate a deep anomaly within the cement volume. After ensuring that both model matrices are correlated and matched for depth, the modal matrices are merged through a further processing step [205] which produces a resultant value based upon the weighting of the significance of the data to the overall solution. The resultant value matrix (as a function of logged depth) [206] can then be filtered as necessary [206, 207] to produce an informative log [207] that indicates the most likely solution of the merging of the modal matrices. For example, an ultrasound modal value of “4” (indicating no cement bond) combined with an x-ray modal value of “1” would indicate that the cement volume was mostly complete, but that delamination (micro-annulus) had occurred between the casing and the cement—which would be typical within an old well (during a plug and abandonment operation). The resultant merged value of “4” would fall within the “delaminated cement” [210] region of the interpreted cement log [208]. A very high value on the informative log [207], for example above “12”, would indicate that both ultrasound modal data and x-ray modal data concur that there is no cement present behind the casing at that depth interval. A very low value, such as “1” or “2” would indicate that both ultrasound modal data and x-ray modal data concur that the cement is well bonded and there is a homogeneous and isotropic volume of cement within the annular region behind the casing, represented by the “good cement” [209] region on the interpreted cement log [208]. Regions on the interpreted cement log [208], could indicate “good cement” [209], “delaminated cement” [210], “deep channel present” [211], “shallow channel present” [212], and “no cement present” [213].
[0038] With further reference to the attached figures, FIG. 3 illustrates how inverted and processed ultrasound images or logs [301], are further processed as a function of depth to assign an interpretive modal value to the data [302], as a function of log depth, which reflects the quality of the bond and the acoustic impedances (relating to density) within the cement mass surrounding the casing. Processed three-dimensional x-ray density data [303] is further processed to assign an interpretive modal value to the data [304], as a function of log depth, which reflects the completeness/homogeneity of the cement mass and the likelihood of major anomalies or channels within the cement. Processed three-dimensional neutron porosity data [305] is further processed to assign an interpretive modal value to the data [306], as a function of log depth, which reflects the porosity within the cement mass and the likelihood of major anomalies or channels within the cement. This can be compared and correlated directly with the x-ray data modal values [304] to ascertain whether a region which is indicating the presence of a fluid channel also correlates with a relative increase in cement porosity within the same depth interval. After ensuring that all model matrices are correlated and matched for depth, the modal matrices are merged through a further processing step [307] which produces a resultant value based upon the weighting of the significance of the data to the overall solution. The resultant value matrix (as a function of logged depth) [308] is then filtered as necessary [308, 309] to produce an informative log [309] that indicates the most likely solution of the merging of the modal matrices. Regions on the interpreted cement log [310], could indicate “good cement” [311], “delaminated cement” [312], “deep channel present” [313], “shallow channel present” [314], and “no cement present” [315].
[0039] In a further embodiment, other input log types, such as open-hole gamma logs, pulse-echo measurements, gamma-gamma density logs and/or neutron-gamma density logs can be combined with ultrasound and other acoustic techniques and nuclear methods through the technique of modal merging to provide more accuracy to the automatic interpretation algorithm.
[0040] In a further embodiment, machine learning can be employed to analyze the results of the modal matrix merging and quality index flags (produced from the inversion of the raw data) to determine whether the selection of data applicability weightings were appropriately selected during the modal merging process, and improve/iterate the possible combination scenarios to improve the overall matrix merging algorithm based upon the study of historical inputs and outputs to the scheme, in addition to machine input ‘quality feedback’ from a human operator.
[0041] The foregoing specification is provided only for illustrative purposes, and is not intended to describe all possible aspects of the present invention. While the invention has herein been shown and described in detail with respect to several exemplary embodiments, those of ordinary skill in the art will appreciate that minor changes to the description, and various other modifications, omissions and additions may also be made without departing from the spirit or scope thereof.
