The customer is a global mining leader. With a history of nearly 150 years, it has become one of the world’s largest producers of mineral resources, operating in some of the most remote parts in the world.
Mining Case Study
With mining operations around the world, the company invested heavily in IoT sensors for greater insight into the health and efficiency of its equipment. With tens of thousands of sensors, the company needed an advanced data analytics solution to support diverse time series data (both structured and unstructured). The analytics solutions needed to ingest IoT data – on site and at scale – and generate actionable insights in near real time to enable predictive maintenance, process optimization, operational alarms, root-cause analysis, and best-response recommendations.
The company established a data lake on AWS as the foundation for its big data and real-time analytics. The global infrastructure provided them with the ability to rapidly ingest, store, and analyze its IoT data from sites around the world. Moreover, it gave them easier access to artificial intelligence (AI) and machine learning (ML) capabilities on the cloud for deeper insights. The company engaged 47Lining to address some of the data engineering challenges with large-scale ingest of IoT data to their data lake on AWS.
From here, they engaged 47Lining to set up a series of “challenges,” through which they experimented on different ways to optimize selected industrial process optimization use cases. The general idea was to establish different proof-of-concepts (PoCs) that would enable them to prototype and validate operational improvements that could then be pushed to production, or fail fast and move on to new ideas. 47Lining helped them structure a series of experiments for rapid learnings.
The roadmap of “challenges” was diverse. For example, the mining company operates “pit to port” railways to get resources to customers. Maintenance issues on rail carts are extremely expensive. To mitigate these costs as much as possible, 47Lining used sensor data to develop a predictive maintenance model to help make informed decisions around maintenance scheduling. Other use cases focused on efficiency of material transport within mines by safely increasing load factor and effective utilization of haul trucks, and dynamic scheduling of material flow when downtime events occur.
These PoCs enabled 47Lining to implement predictive maintenance capabilities and optimize processes leveraging real-time analytics. Predictive analytic capabilities and dynamic scheduling enabled them to pull equipment out of rotation in known intervals, allowing them to budget time and resources accordingly, helping to minimize service downtime.
The mining company took all of the successful PoCs and deployed them into production. The real-time feedback gained from industrial process analytics helped the company improve mineral extraction efficiency and significantly reduce maintenance costs, improving net cash flow by tens of millions annually.