Industrial Operations
Global gold producer - autoclave control
High operator-to-operator variability in managing complex autoclave processes was reducing recovery rates and driving unplanned downtime
The challenge
A global gold producer faced high operator-to-operator variability in managing complex autoclave processes. Suboptimal control choices led to reduced recovery rates and unplanned downtime, with significant financial impact - and the experienced operators who knew the process intimately were close to retirement.
What we built
We developed a temporal deep learning model trained on years of plant data that recommends optimal control sequences for water injection, oxygen flow, temperature, and hold time. The system provides operators with actionable set points while maintaining very low recommendation error, codifying the expertise of senior operators into a system every shift can use.
What changed
Productivity
through optimized control sequences
Unplanned downtime
Recommendation error rate
Built with
Deep Learning · Temporal Modeling · Industrial IoT · Process Control · Time Series Analysis