Within the framework of Perumin and as part of the TIS Forum -Technology, Innovation, and Sustainability- Ariel Medina, Director of Digital Business Development, together with César Moscoso, Technological Projects Engineer, and Álvaro Rendón, Director of ECN Automation, will present the results of the Operational Digital Twin in SAG Mill, a development that already shows good results in the field after its testing phase.
Currently, much of the development of technological tools for SAG grinding aims to optimize mill performance, maximizing production and process stability. To achieve this, comprehensive monitoring of the mill inputs along with its operational state, represented by operational variables, is required. This scenario has promoted the sensorization of the main grinding equipment and the great rise of non-invasive technologies based on artificial intelligence and advanced analytics such as virtual sensors, computer vision, and among others, the so-called hybrid models. The latter appear as a solution capable of delivering timely (real-time) information on variables that still present major challenges today, as their measurement generally involves prolonged process stoppage in addition to costly logistics. This is the case for monitoring the internal mill load, whose main variables are the ball filling level (Jb) and total mill filling (Jc). These currently require stopping, washing, and scanning the mill to be measured, so it is common to find no more than 3 or 4 point measurements during an entire campaign or liner cycle.
Classical deterministic models have demonstrated the importance of knowing Jb and Jc for predicting operational parameters. Furthermore, their applied development to the process allows precise estimation of both variables; however, this is only done discretely, generating analyses with historical data that do not respond to the current process condition. The above situation improves considerably by applying a hybrid approach, where the classical model can be used together with computational techniques, such as some optimization models, whose fast computing speed is compatible with real-time application, making the hybrid model a powerful online analysis tool that serves as the basis for operational intelligence platforms such as digital twins.
The Operational Digital Twin consists of a digital platform that groups different prediction, simulation, and operational analysis tools, several of them based on hybrid modeling. The interesting results of this development will be presented on Thursday, September 29th at 13:10 hrs in the Luis Tokumine – Adolfo Médico room, as part of the TIS Forum, at the Cerro July fairgrounds, Arequipa.
We invite you to learn more about our area https://www.me-elecmetal.com/soluciones/me-digital-lab/#productos
