01
Challenge
The monthly model cycle was slow and manual, and its accuracy against blastholes was not measured.
02
Problem
The monthly model cycle was slow and manual, and its accuracy against blastholes was never measured. Misclassification of ore and waste went unquantified, and few scenarios were tested because each one took too long to run.
03
Solution
We mapped the client's current workflow and rebuilt it on the platform as an indicator and grade kriging workflow, with and without local directions. Twenty-four monthly models were back-tested against blasthole-based models, side by side with the client's own.
04
Deliverables
The outcome is an automated workflow in the platform and a clear accuracy comparison against the client's existing models.
Annapurna ChallengeIndicator krigingWorkflows
