Presentation: Going BI - Optimizing KPI Management Across the Enterprise

Anglo American Platinum is the world’s leading producer of refined platinum, supplying approximately 40% of the world’s newly refined Platinum. Unlike many mining companies, the value chain at Amplats is process intensive. Ore at grades of 1.6 grams per tonne needs to be refined to 99.99% metal. The Anglo Converting Process is the process step where all of Amplats’ production converges. As part of process monitoring and improvement the Metallurgists have Key Performance Indicators (KPI’s) monitored on a daily basis. The first KPI system built, was a success; but was non-scalable, inflexible, onerous to maintain and complex to configure. Due to its success in meeting the metallurgists’ needs more sites wanted the system, but KPI additions took 30 minutes.

A rebuild was needed. This was an opportunity to learn from the old KPI system and other systems and do a complete system redesign. A fundamental architecture and product stack review was done with OSIsoft’s Centre of Excellence (CoE) in conjunction with Microsoft. A new system was designed to give more flexibility, to improve maintainability and scalability and to promote user participation/ maintenance.

Plant data is aggregated by the PI Server. PI Asset Framework (PI AF) is used to hold the KPI configuration in a contextualized plant hierarchy. An SQL server tabular data module on a data warehouse does the aggregations from the PI AF data extractions. User experience is then dynamic, using all of Microsoft’s Business Intelligence tools with the power of PI AF’s contextualization and the PI Server’s live data management.

Maintenance and scalability improved by an order of magnitude by leveraging OSIsoft products.