This is the beginning of a multi-part part series on using well log data effectively from a geological perspective. Additional sections will come out over the next few months as I describe each technique for building the best geologic interpretation. (Sorry – no petrophysics in this series – maybe we will cover that in a future series).
In part, I will be describing some things I have learned along the way for gleaning information that is not on the log. Yes, what is not on a log can often be just as important as what is on the log or said another way, ignoring what is not on the log can lead to an erroneous interpretation. We’ll also look at picking tops. Future parts, I will be discussing isochore maps and their huge importance. How isochore maps should be constructed and why structure maps should often be derivative. Beyond that, we will be looking at the really important part of this discussion with putting all this in a geological framework.
How you interpret your data from well logs can have a major impact on what you can do with that interpretation and can also have a major impact on the technical quality of the displays generated from it. Picking the right tops and picking them correctly is the key to any geological study. What you do with them is also key to maximizing your knowledge and understanding of the geology.
Maps and cross-sections are usually the final displays that convey the maximum amount of information about the geology. Creating them using the best possible data and the best techniques can be the difference between success and failure.
I’ve tried to write this as generic as possible, so no matter what tool you are using, you will understand the basic concepts. Even if you have the ultimate tool that does much of this workflow for you, it is worth it, at least once, to check it and make sure it is really doing what you think it should. Most of the things in here are easy to verify and it is not worth drilling a dry hole just because you think a tool is working the way you think it should.
Pick the top of the formation not the first occurrence.
I realize this is geology 101 but it is worth repeating. When you pick the top of a formation, pick the top of it, not the first occurrence of it on the log. The top of the formation is the last rock that was deposited. It is not the first occurrence of the formation in the log. If the top of the formation has been eroded, even a little bit by an unconformity, it is a poor practice to pick it as the top of the formation. What you are essentially doing is mixing the top of the formation with the top of the unconformity. Keeping them separate will make doing detailed stratigraphic work easier. If you don’t, it will make it difficult if not impossible to do proper geologic mapping. Know the geology and what the formation should look like, top to bottom, before you start picking formation tops.
What isn’t on the log is just as important as what is on the log – be sure you are taking into account the information that isn’t on the log.
Depth of the log and depth of the formations
It is very common when drilling a well, to drill to the top of a specific formation and stop. Although wells are often permitted to a specific depth, most operators are interested in drilling through the objective formation and stop. This means they will drill into the top of the formation below and stop. Given that logs start recording 30 or more feet from the bottom the hole before they start capturing data it is very common for that bottom formation to never be recorded on the log.Knowing the bottom of the first reading on the log (not the loggers total depth) and knowing that a given formation is actually deeper than the log, can give you another data point that you do not already have – especially if the formation is conformable with the ones above.
The error that often happens is we ignore the top of a formation that is below the TD of the well. When we go ahead and map it, the mapping system will ignore it also. The algorithm that makes the grid and contours will also ignore the well and interpolate or extrapolate some value at the wellbore location. Are you sure that value of the contours that intersect the wellbore on the map is actually below the TD? It should be, because you have direct information that it is.
If the log did not penetrate a given formation you can never really know the precise depth of that formation below the wellbore, however, you do know it is below the first reading on the bottom of the log.
Almost all systems will project the grid onto the cross section view so when you see the log in cross section you might recognize the grid crosses onto the log and realize the mistake “made by the computer”. However, many geologist tend to ignore logs that don’t penetrate the formation they are interested in, so the shallow wells tend to never be used in the cross sections. If the log doesn’t penetrate the formation of interest but comes very close, it is a good idea to display it on the cross-section.
Figure 1 – Stick-section showing a problem at the middle well where the surface of the XYZ Formation incorrectly interpolated above the TD of the well.
Figure 2 – Stick-section showing the resulting and corrected version of the XYZ formation that is below the TD of the center well.
Pick the bottom most usable reading on the log. This makes for a strange looking map but will be very useful to ensure you are getting the maximum use out of the information that is not on the log.
Create a top called BottomOfLog and pick the lowest most usable reading off the log. This is almost certainly not the TD or the loggers total depth of the well. Do not use those values. Remember, the logging tool sits on the bottom of the hole with the sensor a few to many feet above that. Each curve will have its own bottom value and I tend to pick the curve that is the most useful in picking the first formation that is below the bottom value.
If the formation, for example, the XYZ formation, is below the TD and did not show up on the logs, marking it DNP (Did Not Penetrate) is another point of information you now have. (Remember: there is a huge difference between Did Not Penetrate and Not Present – do not confuse the two or mix them up).
It is important that when you map the XYZ formation, even though you do not have a formation top, it should not be mapped above the BottomOfLog. This is easily determined by constructing a simple isochore map.
Making the map
The workflow of ensuring the mapping system has constructed proper structure and isochore maps that honor this data is completely dependent on the tools you have available at your disposal. Some systems will automatically adjust the surfaces to account for this information, other systems you might have to do a bit more work and construct an isochore or two.
A very simplistic isochore map can be made using 2 well lists and the difference between the XYZ formation and BottomOfLog. In the first well list put all the wells that have a real XYZ formation top pick and post them on the map in green. In the second list put all the wells that had a DNP for the XYZ formation top and post them on the map in red. Make the isochore map between the XYZ and BottomOfLog. Contour from 0 up to positive and look for wells that are red in the positive contours – you might have a problem there.
The above is not definitive but a great place to start if your system does not have the capability to mark DNP tops and/or the ability to interpolate a value from a grid onto a wellbore (more discussion below).
If the interpretation package you have supports DNP, simply give it the XYZ formation when making the surface and give it the BottomOfLog as the TD and the system will construct the surface of the XYZ and every time there is a DNP entered for the XYZ it will ensure the surface constructed is below the BottomOfLog. However, it is a good idea to at least check this once to make sure the tool is working correctly.
If your system supports a well-interpolation algorithm (determines the value of the surface at every well location), but doesn’t support DNP, you can still make the map and the results will definitely be worth it.
Make a surface of the BottomOfLog and interpolate the value of this surface onto the wells where the value of XYZ does not exist (this must only be a list of wells that have log files). Then go to data operations and construct the isochore between the XYZ formation and the BottomOfLog (BottomOfLog – XYZ). Any place where you have a positive value, you have the YXZ exist on the grid above the BottomOfLog. Since you only interpolated where XYZ did not exist you have a dataset which includes positive values that need to be fixed.
Fixing the surface is again tool dependent. I prefer using a hands on approach and modifying the surface manually where there are known issues. This gives me the ultimate in control and I can make modifications based on information from other logs in the vicinity.
Knowing what is not in a well log can often be just as important as what is in the well log. Knowing the formation you are interested in should be below the log TD can be hugely important. Mapping systems can often ignore this data and construct structures that are seen on the log but are definitely below the TD.
Even doing something just as simple as posting the TD (in subsea) at every wellbore when displaying a structure map, can quickly tell you if you have a problem.
What do you think? Leave a comment below.
Latest posts by John Fierstien (see all)
- Getting the Most Out of Your Oilfield Raster and LAS Data - May 26, 2015
- More about Oilfield Data: Everything is Important Pt 2 - May 5, 2015
- Big Data or Small Data – in Today’s Oilfield Everything is Important - April 28, 2015