Digital underground

OptiMine, Sandvik’s modular production management system for visualizing and managing various data sources linked with mine instrumentation and for controlling mining operations, has evolved. A new component, OptiMine Analytics, transforms the acquired data into predictive insights and actionable dashboards. OptiMine Analytics builds on the IBM Watson artificial intelligence platform and merges its analytic and predictive modelling capabilities with Sandvik’s extensive knowledge of mining operations and equipment. Thanks to the joint expertise, the predictive modelling power of the system is clearly superior to generic analytics solutions.

Safe protocols and encryption technologies are used for all data exchange and storage operations to create a secure cloud environment. Redundant access control systems ensure that users can only access their own data. Sandvik was one of the first mining equipment suppliers to provide an Interoperability Policy to outline the principles of data accessibility, fleet data compatibility and data privacy in compliance with the GDPR. The backbone of OptiMine Analytics is the descriptive component that shows the current and historical values for the equipment, operator and productivity data integrated from all the available data sources. Based on these machine- and operator-linked KPIs and availability rates, the system indicates and visualizes the overall equipment efficiency (OEE) in a few uncluttered parameters: How does the actual production tonnage compare with the target? What is the breakdown of OEE losses in terms of their main causes? But OptiMine analytics is not simply a performance dashboard. It builds on the descriptive analytics data to offer predictive and prescriptive analytics. While the descriptive component answers questions such as “What was the availability figure for this loader during the last six months?”, the predictive analytics offer answers to questions like “Which component in this unit is likely to need unplanned maintenance next month?” The prescriptive analytics, meanwhile, advise the operator on how to avoid the predicted issues, such as by replacing the component that is likely to become faulty.

A longer-term approach can include pre-emptive maintenance schedules, which again helps to minimize unplanned downtime and improves productivity. A major strength of OptiMine Analytics is that these predictive models do not rely on data from a single mine or customer. Through My Sandvik, the system leverages the big data obtained from Sandvik’s entire customer base. Thanks to this force multiplier, the accuracy of the predictions is superior to other comparable solutions, and this also keeps the capabilities of the system progressively improving. As with all technology, software and hardware alone are simply tools. Gaining real impact and results also requires qualified people to operate the systems. In addition to building up in-house capabilities, mine operators can make use of OptiMine 365 service. This practically means that Sandvik mining experts and data scientists join forces to help mine managers find bottlenecks or other issues critical for their specific operations. “The data is already out there,” Mannonen says. “Sensors and systems are producing it and databases are storing it all the time. The real question is how to translate data into actionable insights, and that is where we can offer a truly unique solution. OptiMine Analytics essentially processes data to information, knowledge and, at the end of the day, real-life OEE and productivity improvements.”