AI is ready to prove its value as a tailings management tool

From a January dam failure at a jade mine in Myanmar to another landslide-related failure in May at a gold mine in the Philippines, to a June incident at a Chilean copper mine, 2024 has brought several loud reminders that tailings dam failures remain a persistent and costly issue for the mining industry.

Just how persistent? The organisation World Mine Tailings Failures estimates there are as many as 35 000 mine tailings dumps globally, and that there were 156 catastrophes resulting from collapses at these sites between 1960 and 2023. The cost of these disasters in terms of human health impacts, environmental damage, and clean-up is immeasurable.

Now, however, a new generation of AI-powered capabilities is at hand, giving mining companies a decidedly modern tool to better manage this age-old threat to public health, safety, financial standing, and brand reputation. By using AI-powered information-gathering, modelling, and analytics capabilities, companies can manage tailings more proactively and predictively – and in the process, perhaps even turn the materials trapped in those waste reservoirs into a revenue source.

Taking advantage of AI capabilities requires a solid foundation. That includes cloud readiness (the cloud is where most AI capabilities reside), and a single system of record within the company for managing and housing data. Companies that rely on a patchwork of different processes, spreadsheets, and systems will have difficulty marshalling the fresh, accurate, and relevant historical and real-time data that AI tools require to generate insight. In short, to leverage AI, you need to have a good handle on your data, plus the ability to gather data from internal and external sources, then feed it to the AI capabilities you have on hand.

With that foundation in place, it is time to start exploring specific use cases for AI on the tailings management front, including:

• Predictive analysis of dams and equipment to identify potential failures before they occur. Real-time data coming from a sensor-equipped piece of equipment could identify a potential issue with that equipment that can be addressed proactively. With a pump, for example, an analysis of energy consumption, temperature and flow data could quickly identify an anomaly that suggests it is time to replace a part inside the pump. On a larger scale, intelligent modelling tools could identify a particular portion of a dam that is susceptible to a heavy rain event or a seismic shift.
• Predicting potential dam and equipment stressors based on data from internal and external sources. AI models can take your own historical data about a tailings dam (design specifications, maintenance history, etc.) and couple it with highly localised weather forecast information, for example, to alert a company that they should proactively take steps to prevent undue pressure on a dam, pumps and the like. Likewise, using real-time data from sensors, AI can identify changing conditions in and around a dam and provide advance alerts to the appropriate personnel.
• Using AI-monitored remote cameras to identify conditions that could lead to equipment or dam failures. AI can analyse live video streams to identify potentially troubling changes in water levels, for example, or shifts in topography that could be problematic.
• Rapid access to documentation and history about a specific piece of equipment. By prompting a generative AI co-pilot, a technician can quickly access maintenance information and other critical data about equipment.
• Compliance and reporting. Another area where AI can help mining companies is in complying with new standards for tailings storage facilities, such as the Global Industry Standard on Tailings Management. Because of the mining industry’s involvement in so many value chains and downstream industries, along with the proliferation of sustainability-related regulations, mining companies face a growing responsibility to provide data to customers and regulators about the origin of the materials they extract, how they were extracted, etc. All these efforts are predicated on collecting, standardising, and sharing data. That is where AI can help.

There is another motivation for companies to prevent dam failures and keep tailings in their proper place: the potential to recover valuable commodities from these waste pools. Vale said it expects that by 2030, 10% of its iron ore output will come from the reuse of tailings. The company also has launched Agera, a venture that sells sand produced from the treatment of tailings. Intelligent modelling capabilities can show companies how to optimise these circular processes.

Importantly, intelligent tools like these are available today, so mining companies now have the means to turn tailings management from a headache into a genuine business opportunity.

Read the article online at: https://www.globalminingreview.com/environment-sustainability/23122024/ai-is-ready-to-prove-its-value-as-a-tailings-management-tool/