Autonomy has become a popular buzzword in the mining industry over recent years, with companies touting the benefits of self-directed machines. But the industry needs to carefully consider how such technologies are implemented and whether autonomy in itself is the ultimate goal, as they have major implications for the potential benefits and employee acceptance.
By definition, autonomy is the ability to act without external decisions, guidance or influence. Machine autonomy is achieved when a machine operates on its own, without any kind of control or monitoring. Related though slightly different is automation, which enables machines to complete actions without direct human control. Autonomy is the point at which automated machines can make all the decisions and execute all the actions inherent to its operation.
Many benefits are commonly attributed to the operation of autonomous mining machines: improved safety, increased availability, reduced variability, reduced component wear, improved productivity, et cetera. Often though, reduced labor requirements is the first benefit that comes to a mine operator’s mind when discussing automation. However, this should not even make it on the list of objectives.
Obviously, maintaining the workforce means not making machines autonomous. The benefits enumerated above, however, are not inherent to the autonomy of the machines. Rather, they result from the machines’ automation. And though the other benefits are often only afterthoughts, they are just as real, tangible, significant and measurable.
Removing operators, and thus reducing the salary mass, indisputably has a major positive financial potential. But with that comes all the major costs and risks associated with automation, such as lack of accountability and facing regulatory voids. The fact is that for most processes, automating them is relatively easy and inexpensive. Making them safe is normally not complex either. What makes automation expensive is ensuring that the systems implemented are guaranteed to be safer than manned operations in all possible circumstances. It is insufficient to make autonomous machines safer than manned machines on average. There is an expectation that autonomous machines must be safer, or at least as safe, in all possible individual situations that exist. Therefore, it is best to leave operators in control of automated machines (i.e. not making them autonomous). This enables operations to gain the benefits of automation, while leaving on employees the responsibility to react to abnormal situations when necessary.
Both the benefits and the costs increase as the level of automation increases toward autonomy. Yet they do not increase at the same rate. Benefits grow faster at the beginning, when you start collecting the low-hanging fruits; but become less significant as the major gains are realized. Similarly, successive improvements typically become more onerous as the cheaper ones get implemented.
The problem is that it is not clear what the right level of automation is for a given operation until it is implemented. Yet two things are sure: 1) Not automating guarantees not getting any of the benefits automation can provide, and 2) The more an operation automates, the higher the probability that it will hit the point of diminishing return (i.e. the level of automation where the incremental costs outweigh the incremental benefits). For most existing sites, the benefits of autonomy are not worth the investment. That does not mean one should not automate existing operations, but the change management challenge that represents the transition to autonomy is a major impediment. Change management is too often overlooked and if not addressed properly, it will fail one’s automation effort. Mine operations will be reluctant to replace seasoned operators with computers and electronic sensors, and automation technology trials will often yield poor results due to people voluntarily failing the prototypes, pushing them outside their operational range.
The industry should not start with the principle that no operator will be in charge of tomorrow’s mining machines. It might be the case, but we should progressively make machines more automated and let machine autonomy become a natural evolution as machines become able to handle longer periods of autonomy. Savings related to labor force reduction will materialize on their own as automation allows operators to do additional tasks or operate more machines at one time. Progressive implementation provides the double benefit of diminishing the change management challenge and minimizing any eventual investments beyond the point of diminishing return.
François Gariépy has extensive experience engineering software for mining applications. He is currently the director of the technology solutions for Peck Teck Consulting Ltd.
Got an opinion on one of our columns? Send your comments to firstname.lastname@example.org.
Combining geostats, AI and VR to find gold deposits
Jean-Philippe Paiement explains how SGS has combined historical geological data with machine learning and virtual reality tools to identify and visualize drilling targets