Machine Learning and optimization were part of the key areas announced by ABB within the framework of the 19th version of the International Conference on Mineral Processing and Geometallurgy, Procemin Geomet 2023. An activity held between October 4 and 6 at the Sheraton Hotel in Santiago, Chile.
At the event, David Gallego, Digital Solutions Architect at ABB in Switzerland, spoke about the remote monitoring of the Gearless Mill Drives (GMD) that drive the SAG mills and ball mills in large-scale mining. The executive highlighted the remote diagnostic service, arguing that GMDs are crucial equipment in the grinding process, therefore, “ensuring their maximum availability constitutes a primary and essential task to obtain the best possible indicating values for OEE (Overall Equipment Effectiveness)”.
The expert pointed out the relevance that this technology has for top mining countries, such as Peru and Chile, which are the nations with the largest installed base of GMDs in the world. He highlighting that many of ABB's clients in Chile and Peru, through the Long Term Service Agreement (LTSA) offered by the company, have access to and benefit from the Predictive Maintenance for Grinding platform, ABB Ability™.
Our platform is cloud-based and uses the most advanced services offered by Microsoft Azure, especially in terms of Machine Learning. Specifically, with use of the Azure AutoML service that was recently launched, it allows us to evaluate the prediction against 30 different models in real time and select the prediction that best fits our criteria,” indicated the Digital Solutions Architect from ABB Switzerland.
The executive also explained that the possibility of accessing the transient files of the Cycloconverter controller was recently deployed, for tasks A, B, and C. They contain high-frequency data from different signals and are normally difficult to access. However, they are instrumental at the moment to perform a correct diagnosis when performing troubleshooting tasks.
Regarding the advances in the material presented by the company, the executive highlighted that they are currently “monitoring almost 40 GMDs from various clients around the world, and with the strategies presented, our estimate is that, with the five main cases detected, it has been possible to avoid 120 hours of unexpected shutdowns for the grinding equipment”.
