Despite strong commodity prices, the need has never been greater for the mining industry to innovate. Technology development can address the challenges of a capital-intensive industry with ongoing cost pressures and staffing in remote locations, but only if the will exists to make it work.
A mining operation is complex and continually changing, not to mention physically and mentally demanding for the people involved in the work — more than in manufacturing plants where advanced technology such as robots are typically used. So, the operating environment is a tougher challenge for new equipment and systems than the conditions in a conventional plant. New technology must integrate smoothly with existing operations, and the initial capital cost must be balanced with high utilization and reliability to provide an adequate return.
Despite the complex challenges for developing and deploying new technologies, there are many opportunities to improve mining methods and reduce the ancillary costs of mining.
Mining and mineral processing operations have to contend with variable feedstocks, disruptions in feed and multiphase process streams with unusual properties. A model-based analysis of these processes is needed to help operators to improve controllability, along with reliable sensing of key feedstock parameters such as grade and contaminants that would affect downstream separation processes. In this way, separations can be robust to feedstock variability without relying on energy intensity and expensive ore- blending strategies.
Automation of process control is a logical evolutionary development for mining, not only underground but also on the surface. This technology will address a number of process performance issues. Increasingly, mining and mineral processing are becoming tightly coupled, and will become more so as new technologies are introduced to integrate more of the process in the mine to reduce material transport costs. As mines grow, the ton-kilometres of material transport becomes a higher operating and maintenance cost.
Robotic mining technologies have reached the point of commercial deployment, but their first implementation may not be for automatic excavation and haulage, but for niche applications to reduce risks to workers, such as: capping and reclaiming tailings impoundments; applying wear-resistant overlays to increase the useful service life of ground-engaging tools, bins and liners; and inspecting hoists, structures and vessels in remote or hazardous locations. In this way, the technology can be integrated and debugged without putting the key production processes at risk. Such incremental innovations are not only simpler to retrofit, but they also provide a foundation of understanding and technology that can lead to a real leap forward.
Step changes in technology change the nature of the business, such as enabling production from inaccessible or hazardous locations, delivering a level of performance not achievable by humans or existing technology. For example, shovels and trucks are limited by operator safety and comfort as much as by available power and structural loading limits. It is only through new technology that reserves will be increased, by being able to access ore that is not possible with today’s technology.
The real commercial opportunities for robotic technology will be in production applications: automating most mining activities so that there is tight operational control of an integrated process. Some enabling technologies are required for this to happen, in organizations that are ready to innovate.
Improved sensing systems will most likely be based on non-contact methods (vision). For machine navigation and control, this is necessary from automatic construction of unstructured environments, with scene refreshing so that the machine knows when changes occur and can automatically avoid collisions. Limited use of spectrometry and isotope detection is already used for ore grade estimation. Machine vision is used to detect lost shovel teeth and oversized lumps of ore on conveyor belts.
New ore classification techniques are being developed; a real-time estimate of oil sands grade-to-fines ratio, with a real-time method to estimate bitumen recovery, finally brings the industry to the point of feed-forward control of extraction. Near real-time ore characterization is also necessary at the mine face to reduce dilution. The best way to reduce the amount of dispersed clays in tailings is to avoid mining interburden in the first place.
Improved communication is necessary to allow people to operate equipment from a distance. This has already been achieved underground, where radio spectrum allocation is not an issue. Navigation underground has been the challenge. Giving a person a good sense of what is going on requires a lot of bandwidth; but it is now possible to have the control room in another time zone. Where possible, machines should be given increased autonomy for a greater number of tasks with on-board control schemes to deal with the inherent uncertainties in a mining environment. In that way, multiple robots can work in a coordinated fashion, with the advantage of delivering a high level of production with smaller, less expensive equipment, with smaller impact if one machine needs to shut down for service.
Operating strategies will have to be developed for combined autonomous and human-operated systems, including design for ruggedness to different operating conditions, and delivering reliable service, through strategies for fault-tolerant control, error recovery and overall system reliability, and intelligent systems for machinery diagnostics and prognostics for predictive maintenance, based on relationships between production and reliability.
This culture of innovation will have a significant impact on organizational structure and operating philosophy. In learning how to operate a mine like a factory, mining companies will develop best practices that can be used effectively now.
Mike Lipsett is a professor in the Department of Mechanical Engineering at the University of Alberta.