Metals and mining case study
Industry: Steel manufacturing
Location: South America
Machine Type(s): Continuous Steel Casters
Components: Bending Roll, Breakout, Pinch Roll
A major South American steel manufacturer was underperforming its revenue targets because of unscheduled plant downtime. SKF Enlight AI’s AI-driven Industrial Analytics solution triggered alerts up to 8 days before failure occurrence. The failure prediction rate of 93% has helped the manufacturer improve yield rates and reduce maintenance and overhead costs.
Lost production / revenue due to unscheduled downtime High O&M costs from overtime labor expenditures
The plant began testing SKF Enlight AI on its continuous steel casters. Process data generated by over 400 sensors was analyzed by SKF Enlight AI’s advanced Automated Machine Learning (AutoML) algorithms. Based on detection of anomalous behavioral patterns, the solution provided predictions of evolving failure and indications of the suspected root cause.
Business Impact / ROI
SKF Enlight AI triggered alerts up to eight days before failure occurrence with an accurate prediction rate of 93%. Based on this and once fully deployed, it is estimated that the solution will reduce unplanned downtime by 30% and operational costs by 15%.
Data generated by over 400 sensors was streamed to SKF Enlight AI’s cloud. Enlight AI applies advanced Machine Learning algorithms (AutoML) to the data. Based on detection of anomalous behavioral patterns, the solution provided predictions of evolving failure and indications of failure root cause.
What is Automated Machine Learning?
Within the Machine Learning discipline, there are multiple manual processes that are dependent on data scientists. AutoML applies advanced algorithms which automate manual Machine Learning processes, thereby reducing the need for human labor.
With AutoML industrial plants can scale Predictive Maintenance solutions across multiple plants within a few days.