The hole truth? - Computer modelling for mining

Materials World magazine
,
1 Feb 2010
The DeBeers Venetia mine, South Africa

Calculating ore resources and planning for their safe and efficient extraction is an essential part of designing and funding a mine. This estimation is the financial foundation of a mine and the hook on which a developing company’s reputation hangs.

Computer modelling programmes can help resolve estimation by takinggeochemical values from the surface and underground to provide a calculated resource. In addition, detailed geometallurgy can help determine therecoverable metal from the ore.

Mine design requires reliable geotechnical data both from the ore deposit and surrounding country rock in order to plan an extraction method, the economics of which determine the cut-off grade. Once established, refinement by the geologist transforms the resource value into a reserve figure, according to the stock exchange code where the money is being raised.

Surface mapping has reached new levels of accuracy, aided by global positioning satellite (GPS) devices and lasers, resulting in a highly defined surface expression of mineralisation.

A little below the surface, this level of accuracy falls away and geology mapping is reliant on geophysics and, most critically, the careful recovery and alignment of core for interpretive analysis. The principal below-surface sampling tool is the diamond drill rig. The core that it recovers is used to define the metal content of the ore, as well as the sub-surface structural geology. Unfortunately, not all cores are perfect, as ground conditions and driller competence all take their effect, particularly on core loss and borehole deviation.

One of the principal concerns is the depth inconsistencies inherent in core drilling, requiring detailed records to be maintained. While the depth of each drill run is accurately measured to the cutting face of the drill bit, a corestub of variable length always remains in-hole, leading to significant errors in logging the sample.

Furthermore, the quality of any geotechnical logs depends on the accuracy with which the core is aligned, or oriented, so that true strikes and dips are measured for all geological structures found in the core – given an accurate borehole survey. With regard to the latter, several ‘tools’ are available to survey the borehole path as it penetrates the underlying strata, each withdifferent degrees of accuracy, and therefore error, in plotting the logged datain 3D space.

 

Pace yourself

Although rigorous monitoring procedures are required to ensure that quality core is delivered to the geologist, the most effective method to get reliable data is to ensure logging and data analysis keep pace with the drilling. This enables errors to be quickly identified and corrected so that the value of the drilling programme is not compromised.

However, manual logging methods, using a tape-measure and clipboard with agoniometer to measure the alpha/beta angles of geological structures relative to the core axis, are slow, skill intensive and require experienced loggers with geological knowledge for a positive outcome. It is also a once-off activity, requiring both selective re-logging for verification, as well as dataentry before any analysis is undertaken.

John Orpen, in South Africa, claims to have developed a faster and morere producible method of logging core. The software, StereoCore PhotoLog, uses handheld digital photography of the core trays, which are calibrated for scale and perspective to construct an accurate virtual 3D model.

‘This allows structures to be measured in much the same way as down-hole geophysical scans – by matching the planes with a half-ellipse trace drawn onthe image,’ he says. ‘The geometry of the trace is then automatically calculated to derive the attitude of the plane relative to the core axis, otherwise commonly referred to as the alpha/beta angles. After loading the borehole survey, these data are automatically rotated to derive the dip and dip direction for each plane, the poles for which are then automatically plotted on a contoured equal area stereogram.’

He continues, ‘It is then a small step to compare the stereogram, either in its entirety or filtered, to plot the poles for one structure type at a time, against stereoplots derived from adjacent boreholes or surface mapping. Discrepancies can then be investigated and, if they are not due to a change instructural domain, possible errors in either the core orientation or borehole survey can be timeously corrected’.

Orpen says the process is faster than manual logging and has an added advantage over down-hole scans in that each structure can also be confidently described geotechnically after physical examination of the core.

Known dimensions

To ensure that the photogrammetry is easily implemented, the system developed by Orpen uses a hand-held digital camera that records an image of the boxed core within a simple reference frame of known dimensions. This enables the photographs to be taken beside the drill rig, while the core is fresh and unspoilt.

To speed up the logging process, image processing and depth registration can be undertaken by technicians who continuously update the project file for the geologist, who is left free to concentrate on structure identification and description so that logging keeps pace with the drilling. Descriptive formats are fully customisable enabling logging codes to conform to existing company standards.

Orpen says the digital format allows for immediate e-mail file transmission for data analysis and visual confirmation of data quality. It also allows easy comparison with independent down-hole geophysical scans. If they correlate, this results in a greatly enhanced geological model.

During its development phase, StereoCore PhotoLog was used to comprehensively audit over 69,000m of geotechnical logging, mainly for the DeBeers Venetia Mine in South Africa. All manually measured alpha/beta angles were converted to half-ellipse traces and plotted at their logged depths on the processed imagery of the core. Each trace was then compared with its corresponding fracture in the image and its colour code changed from blue (confirmed) to red (queried) if the match was poor. The results were presented stereographically and histographically along with a calculated overall percentage confidence in the data set.

The annotated imagery enabled the core to berevisited for the queried fracture measurements to be easily located, confirmedand corrected. This process established the software’s accuracy and it was subsequently used to rapidly measure the alpha/beta angles directly from thecore photography of 16,512 gneissic foliation planes that were ignored in thefirst round of logging 27,962m of core. In addition, 276 non-penetrative lineations were measured on the imagery, and both the foliation and lineation data were successfully used to significantly enhance the country rock model’s predictive capability.

Further information: Ground Modelling Technologies