Automatic Determination of Lithology Boundaries from Downhole Geophysical Logs

Coal geologists undertake drilling to assist them in estimating both the lateral and vertical extents and quality of their coal.   

When coal geologists drill a borehole, they usually collect two types of data:

• Geological logs recording the chips or core samples of the rock;
• Geophysical logs that measure a variety of physical characteristics of the rock.

Both types of data provide important information regarding:

• The geological characteristics of the rock down the borehole;
• The location down the borehole of the rock with the observed characteristics.

Even though both types of data provide both types of information, in general:

• The rock samples from cored bores provide considerably more information regarding the rock characteristics than the geophysical data;
• The geophysical data from non-cored bores provide more accurate information regarding the location of the rock sample data than the geologist's log.

One of the responsibilities of the coal geologists is to merge both types of data into a single geological log that can then be used to understand the overall geology of the project and to estimate the quantity and quality of the coal. This merged log needs to account for both the rock characteristic and the location information from both types of data and weight this information according to its source.  Over the last twenty years, a number of software packages have been developed to make this task easier by displaying the geology as it has been recorded from the chip or core samples and the relevant geophysical logs on a computer screen and thus enabling the use of the mouse and keyboard to adjust the boundaries and the unit lithologies in the geology log.  The geological log before such adjustment is generally referred to as the Raw Geology log and the resulting log following adjustment as the Adjusted or Corrected Geology log.

Despite the introduction of these software systems, this adjustment process still remains somewhat tedious and time-consuming and subjective and thus not reproducible.

This project has developed methods to improve both the efficiency and objectivity of this process. Using the algorithms described in Chapter 4 of the report, the user can now generate and display computer “suggested” matches between the Raw Geology and the Geophysics for the roof and floor of coal seams, and the mid-points of both thin partings within the coal and thin coal seams within the interburden. The geologist will probably be able to accept a large majority of these “suggested” matches and will be able to devote most of the log adjustment time to deciding on appropriate matches for the minority, more difficult matches. It will be much quicker to perform the majority of matches and they will have been deemed on an objective and reproducible basis.