Mineralised rocks shift around during blasting, and this movement significantly impacts every mine. However, while all mines encounter dilution and losses during blast movement, the style of mineralisation dictates the scale of lost revenue.
While material within a blast typically moves in an overall dominant direction, rocks also move locally in many directions and to variable degrees within different parts of the blast. This complex movement makes modelling difficult.
Accurately modelling blasting is a high-value problem, given that the highest proportion of dilution and losses typically occurs at the blasting to mining stage. Conventional prediction techniques allow for the overall blast movement to be approximated in the dominant 2D direction of throw. However, such methods are unable to successfully model a blast with any accuracy.
Augment Technologies has designed the OMP 3D modelling solution to solve this. OMP uses hundreds of thousands of vectors and machine learning algorithms to produce a high-resolution, 3D block model of a blast. This model can be configured to contain as many attribute fields as required, and re-blocked to mining scale. A new post-block mark-out can then be created.
Figure 2: Using hundreds of thousands of vectors and machine learning algorithms, movement is defined in a high-resolution model containing as many attribute fields as required. This model can be re-blocked to mining scale and a new post-blast mark-out created.