Process And Asset Monitoring At A PGM Concentrator For UG2 Concentrate

Advanced process and equipment analytics have been around for a while. However, with accessible IT infrastructure and site maturity, it is possible to analyse, monitor, and control processing plants and equipment easily. These various disparate systems provide different reports requiring the involvement of various personnel and attention to different reporting platforms. For efficient harnessing of the systems? capabilities, as well as effective improvement and monitoring of process and equipment, a single point of entry is required. A process and equipment optimization solution was required at a concentrator in South Africa, focussing on the product stream of platinum group metal (PGM) concentrate from the UG2 Reef. The investigation of data to identify strategies for optimization was required to allow visibility into plant operations to guide operators, an easier method to identify process loop problems, identification of imminent equipment failures, and advanced process control to optimize throughput and stability. A collaborative effort between the client and General Electric Intelligent Platforms (GEIP) investigated data-driven solutions to each of the problems. With the focus on process optimization and asset monitoring, as well as equipment monitoring, a combination of approaches was utilized to address the various problem areas that were identified. Each was implemented separately in the control layer of the plant IT infrastructure. The integration was done in a centralized web-based reporting tool that integrated the detailed results from the various solutions. On the process control side, monitoring of control loops was included by calculating performance statistics that were reported daily to identify sub-optimally performing control loops and indicating potential problems. The advanced process control solution for this site includes a surge tank controller, stabilizing flow through the flotation banks, and preventing the surge tank from overflowing or running empty. Equipment monitoring was investigated, for preventative maintenance, by making use of statistical data models to indicate imminent equipment failure. These predicted alarm events included the cause for failure, drilling down, and allowing focus on a specific variable for preventative action. The results included a more stable flow through the mill, with fewer stoppages due to surge tank overflows or running empty. The causes for process drift could be identified and operators could react early. Alarms warning of equipment failure 2½ and 4 months in advance were identified in an offline analysis.