System and method for automatically correlating geologic tops
A system and method are provided for automatically correlating geologic tops. The system receives well logs from different well bores and one or more user seed picks identifying a well top to be correlated. Each of the seed picks is added to a priority queue ordered by each pick's confidence. User selected picks are assigned the highest level of confidence. The system performs correlation by selecting a window of well log data about a user's manual pick, selected from the top of the priority queue, and then finding the best optimal match with a corresponding window in a neighboring wellbore. That new pick is then estimated in the target well using a correlation function. A quality value and a confidence value may then be calculated for each pick using some correlation function, for example dynamic time warping, and added to the priority queue according to the confidence value. The system may be configured so that picks that fall below a preset quality or confidence value may be discarded and not added to the queue. The system may then move on to the next pick in the priority queue.
1. A method for automatically correlating geologic tops using at least one processor, the method comprising:
receiving a first well log of a first well bore and at least a second well log of a second well bore;
receiving at least one seed pick designating a particular sequence of data in the first well log as a well top;
determining, by the processor, at least one neighbor for each well log;
defining, by the processor, a highest confidence series of well top picks by retrieving a highest confidence pick wherein the confidence of the pick is a non-increasing function of path length from the at least one seed and a non-increasing function of a quality of the pick, determining a new pick by performing correlation on each neighbor of the well log, assigning the new pick a pick quality value and generating the highest confidence series of well top picks; and
displaying, on a display connected to the processor, a result of generating the highest confidence series of well top picks.
2. The method of claim 1 , wherein performing correlation further comprises performing dynamic time warping on each neighbor of the well log.
3. The method of claim 1 further comprising displaying the first well log and the at least second well log and selecting, by a user, a seed pick for the display of the first well log and the at least second well log.
4. The method of claim 1 , wherein receiving at least one seed pick further comprising assigning a high confidence value to the at least one seek pick.
5. The method of claim 3 , wherein selecting the seed pick further comprises assigning a high confidence value to the at least one seek pick.
6. The method of claim 1 further comprising displaying the highest confidence series of well top picks.
7. The method of claim 1 , wherein determining at least one neighbor for each well log comprises adding each well log to a complete weighted graph, wherein each well log is represented by a vertex, and each edge weight represents a distance between a first well log and a second well log and determining at least one neighbor for each vertex.
8. The method of claim 1 , wherein receiving the first well log and the at least second well log further comprises selecting, by a user, the first well log and the at least second well log.
9. The method of claim 7 , wherein the determining at least one neighbor for each well log further comprises using Delaunay triangulation.
10. A method for automatically correlating geologic tops using at least one processor, the method comprising:
receiving a first well log and at least a second well log;
receiving at least one seed pick designating a particular sequence of data in the first well log as a well top;
adding, by the processor, the at least one seed pick to a confidence priority queue, wherein the confidence priority queue is configured to assign a priority based on a confidence value and wherein at least one seed pick is assigned a maximum priority and wherein the confidence value of the pick is a non-increasing function of path length from the at least one seed and a non-increasing function of a quality of the pick;
determining, by the processor, at least one neighbor for each well log;
defining, by the processor, a highest confidence series of well top picks by traversing the confidence priority queue until the queue is empty by retrieving a first element from the confidence priority queue and removing the first element from the queue, determining a new pick by performing correlation on each neighbor of the first element, assigning the new pick a pick quality value, adding the new pick to the confidence priority queue according to the confidence value and generating the highest confidence series of well top picks; and
displaying, on a display connected to the processor, a result of generating the highest confidence series of well top picks.
11. The method of claim 10 , wherein adding the new pick to the confidence priority queue comprises adding the new pick to the confidence priority queue when a quality value of the pick exceeds at least one of a quality threshold and a cumulative confidence threshold.
12. The method of claim 11 , wherein the quality value of the pick is determined by:
q
=
∑
i
=
1
n
(
X
i
-
X
_
)
(
Y
i
-
Y
_
)
∑
i
=
1
n
(
X
i
-
X
_
)
2
∑
i
=
1
n
(
Y
i
-
Y
_
)
2
where X and Y represent the correlated data sequences entered in the warping function.
13. The method of claim 10 wherein adding the new pick to the confidence priority queue further comprises:
searching the confidence priority queue for an element corresponding to the new pick;
updating the element corresponding to the new pick when the confidence priority queue contains the element corresponding to the new pick and the pick confidence value exceeds an element confidence value defining the confidence of the element; and
adding the new pick to the confidence priority queue when the confidence priority queue does not contain the element corresponding to the new pick.
14. The method of claim 10 , wherein determining at least one neighbor for each well log comprises:
adding each well log to a complete weighted graph, wherein each well log is represented by a vertex, and each edge weight represents a distance between a first well log and a second well log; and
determining at least one neighbor for each vertex using Delaunay triangulation.
15. The method of claim 10 , wherein the quality value is determined using a Pearson quality measure.
16. The method of claim 15 , wherein the new confidence value is a function of the Pearson quality measure and the confidence value.
17. The method of claim 10 , wherein performing correlation further comprises performing dynamic time warping on each neighbor of the well log.
18. The method of claim 10 further comprising displaying the first well log and the at least second well log and selecting, by a user, a seed pick for the display of the first well log and the at least second well log.
19. The method of claim 10 further comprising displaying the highest confidence series of well top picks.
20. The method of claim 10 , wherein receiving the first well log and the at least second well log further comprises selecting, by a user, the first well log and the at least second well log.
21. An apparatus, comprising:
a computer system having at least one processor, memory and display, wherein the processor is configured to:
receive a first well log of a first well bore and at least a second well log of a second well bore;
receive at least one seed pick designating a particular sequence of data in the first well log as a well top;
determine at least one neighbor for each well log; and
define a highest confidence series of well top picks by retrieving a highest confidence pick wherein the confidence of the pick is a non-increasing function of path length from the at least one seed and a non-increasing function of a quality of the pick, determining a new pick by performing correlation on each neighbor of the well log, assigning the new pick a pick quality value and generating the highest confidence series of well top picks; and
the display displaying a result of generating the highest confidence series of well top picks.
22. The apparatus of claim 21 , wherein the processor is configured to perform dynamic time warping on each neighbor of the well log.
23. The apparatus of claim 21 , wherein the computer system has a display that displays the first well log and the at least second well log.
24. The apparatus of claim 23 , wherein the computer system has an input device that is configured to allow a user to select the seed pick for the display of the first well log and the at least second well log.
25. The apparatus of claim 21 , wherein the processor is configured to assign a high confidence value to the at least one seek pick.
26. The apparatus of claim 23 , wherein the display displays the highest confidence series of well top picks.
27. The apparatus of claim 21 , wherein the processor is configured to add each well log to a complete weighted graph, wherein each well log is represented by a vertex, and each edge weight represents a distance between a first well log and a second well log and determine at least one neighbor for each vertex.
28. The apparatus of claim 21 , wherein the processor is configured to add the at least one seed pick to a confidence priority queue, wherein the confidence priority queue is configured to assign a priority based on a confidence value and wherein at least one seed pick is assigned a maximum priority.
29. The apparatus of claim 28 , wherein the processor is configured to define the highest confidence series of well top picks by traversing the confidence priority queue until the queue is empty by retrieving a first element from the confidence priority queue and removing the first element from the queue, determining a new pick by performing correlation on each neighbor of the first element, assigning the new pick a pick quality value, adding the new pick to the confidence priority queue according to the confidence value and generating the highest confidence series of well top picks.
30. The apparatus of claim 21 , wherein the processor is configured to add the new pick to the confidence priority queue when a quality value of the pick exceeds at least one of a quality threshold and a cumulative confidence threshold.
31. The apparatus of claim 30 , wherein the quality value of the pick is determined by:
q
=
∑
i
=
1
n
(
X
i
-
X
_
)
(
Y
i
-
Y
_
)
∑
i
=
1
n
(
X
i
-
X
_
)
2
∑
i
=
1
n
(
Y
i
-
Y
_
)
2
where X and Y represent the correlated data sequences entered in the warping function.
32. The apparatus of claim 21 , wherein the processor is configured to search the confidence priority queue for an element corresponding to the new pick, update the element corresponding to the new pick when the confidence priority queue contains the element corresponding to the new pick and the pick confidence value exceeds an element confidence value defining the confidence of the element and add the new pick to the confidence priority queue when the confidence priority queue does not contain the element corresponding to the new pick.
33. The apparatus of claim 21 , wherein the computer system further comprises one of a computer and one or more computing resources.
34. The apparatus of claim 33 , wherein the one or more computing resources are one of one or more server computing devices and one or more cloud computing resources.
35. The apparatus of claim 34 further comprising a computing device used a user to access the one or more server computing devices.
36. The apparatus of claim 27 , wherein the processor is configured to determine the at least one neighbor for each vertex using Delaunay triangulation.