How SLDM (Santa Luz) improved dilution control through predictive blast-movement modelling
Pre/post-blast modelling sharpens polygon decisions. Teams route material with more confidence and less variability.
Context
A carbonaceous orebody introduced complex feed variability, increasing dilution and making grade control more difficult. Traditional classification methods struggled to differentiate material types. A better way to characterize post-blast material was required.
Solution
Machine-learning models were applied to post-blast data to classify carbonaceous and non-carbonaceous zones more accurately. Updated polygons guided equipment operators in routing material more effectively. This reduced misclassification and improved downstream feed consistency.
Results
Classification accuracy improved, leading to more stable plant feed and better dilution control. The operation gained confidence in routing decisions supported by data-driven geological interpretation.
Data Inputs
Pre/post-blast survey, timing, design parameters