Snowden employs 60 specialist consultants who provide a range of mining project services, including resource modelling, geotechnical studies, pit optimisation, open pit and underground mine design and technical IT solutions. Other offerings include providing various specialist courses in resource estimation, grade control and technical risk assessment.
One of the company’s focus areas is resource estimation, using alternative estimation and analytical techniques that can be applied to coal, according to Mark Noppé, principal consultant geologist for Snowden in Brisbane.
Noppe said Snowden had evaluated the relationship between drilling expenditure and resource estimate confidence at a number of operations. He provided two examples where Snowden has helped companies finetune the balance between expenditure and resource definition risk.
One of these included the planning of a new high extraction longwall colliery.
“Mining experience from an adjacent operation suggested that the existing 1km by 1km drillhole spacing was insufficient to identify the expected ranges in seam thickness necessary for specifying equipment and for reliably estimating tonnage and quality parameters,” he said.
The error of estimation for Specific Energy and seam thickness for a typical block, representing one month of longwall extraction (200m by 200m), were calculated for a variety of planned drilling patterns. The estimation errors were plotted against the specific sampling configurations and sample densities as shown in the graph. The ‘I’ in the sample configuration indicates that the pattern is rotated to allow for grade or thickness trends.
Noppe said in the case of seam thickness estimates, there is a 25% improvement in estimation confidence by increasing the drilling density from 1 to 3 drillholes per 100 hectare. A further 15% improvement can be achieved by increasing the sampling density to 5 samples per 100 hectare, with an inclined grid (750m by 250mI configuration). However, there is essentially no improvement in estimation confidence to be gained by increasing the sample spacing to as much as 12 samples per 100 hectare (400m by 200mI configuration). Although the estimation error is very much lower for the specific energy, a similar conclusion may be made.
The ability to quantify the relative benefits of various sampling scenarios allows management to make much more informed decisions on programs of work. This type of analysis also provides a better interpretation of the confidence in existing resource estimates. When this work is extended to conditional simulation, the likely variability for given mine schedules and mining periods to be even more reliably assessed.
Another example of Snowden’s work in this area comes from the iron ore sector, at Iscor’s Sishen iron ore mine in South Africa. Snowden used tools such as estimation error (using kriging variance) and conditional bias (using kriging efficiency and slope of regression) to quantify the relationship between confidence in the resource estimate and drilling expenditure. The relationship is summarised as a graph depicting confidence compared with expenditure compared to the original 400 metre exploration grid.
The 20-fold increase in cost associated with 50m infill drilling results in a 64% improvement in confidence compared with the 400m drilling base case. The remaining 26% improvement is achieved by 20m infill drilling, representing a 125-fold increase in cost compared with the 400m grid. There is, however, no further increase in confidence associated with 16m drilling. The most “bang for the buck” is achieved at a drill spacing of 50m, beyond which the slope of the confidence curve flattens significantly and the benefit of additional sampling is small in comparison.
The recommendation from this work, together with knowledge of the geological and grade continuity, was that the exploration model should be based on a drilling grid of 100m for long term planning (Indicated), 50m for one- to three-year planning (Measured) and 20m for short term planning (production). The exception is that in areas of structural complexity, the 50m grid should be infilled to 25m to achieve Measured status.
Snowden’s Brisbane offices are located with those of consulting company Runge and Noppe said the two groups hoped to work together in a non-competitive and mutually beneficial manner. Other areas of specialisation Snowden plans to offer include quality assurance and data analysis techniques, and internationally recognised levels of expertise.
“Snowden’s specialists work equally well as independent consultants or as integrated members on our client’s project teams, adding value through both their input and by drawing on the extensive experience of Snowden’s world-class team,” Noppe said.
