Integration of Downhole Geophysical and Lithological Data from Coal Exploration Drill Holes

Abstract

The primary variable of interest in a coal resource study is the volume of coal as estimated from the coal thicknesses in each drill hole. It is therefore essential to accurately determine, down to the centimetre level, the thickness of each seam. To attain this accuracy, each drill hole is geophysically logged as these logs provide a considerably more accurate indicator of seam boundary depths than the geologist’s log. Currently, coal geologists spend a large amount of their time integrating their logs with depth information from the geophysical logs. They do this by displaying the two logs next to each other and then manually changing the depths in their logs. Most of this process is relatively routine and thus rather tedious and boring but like many seemingly simple cognitive tasks, not easily transformed into a computer algorithm. The manual method also suffers from being subjective and often non-repeatable. Previous methods to automate this process have used multivariate statistical techniques to assign lithologies down the hole based on the geophysical values at each reading depth. However, despite these methods having been developed and publicized for over thirty years they still have not been widely adopted as they still do not integrate the information from the two types of data. Following the generation of a lithology log from the geophysics, geologists still need to manually integrate it with their log. This current study has successfully managed to develop algorithms to automatically determine both coal/non-coal and clayey/non-clayey boundaries based on the gradients and inflection points of the geophysical logs and then integrate this information with the geologist’s log.