Unearthed Melbourne 201718 Mar - 25 Mar
From Idea to Prototype on Resource Challenges in just a Weekend
This competition has finished.
Our prototype predicts downtimes before they happen using machine learning (specifically an isolated forest). It can predict, with certainty, 16% of all downtime failures 5 minutes before they happen. Downtime in Cadia accumulates ~$250 million in losses per annum, by detecting 16% of all failures, Uptime could potentially save Newcrest $40 million. It should be noted that the median downtime at Newcrest is roughly ~3 minutes for every downtime event.