Discover the key maintenance and safety innovations that will shape the mining industry this year.
According to Google, the top five concerns in mine maintenance in Australia during 2025 were occupational health and safety (especially around ground support and heavy machinery); equipment reliability and ageing machinery; environmental impact and liability; labour and skills shortages; and financial and logistical challenges. As maintenance, repair and overhaul (MRO) operations are key to delivering zero-harm production, these issues are inter-related.
It seems incredible to think that Open AI released ChatGPT 3.0 in November 2022. In the space of just three years, large language artificial intelligence (AI) models have grown to influence almost every aspect of our lives, and this includes MRO services. With the advent of agentic AI – the ability of AI bots to act like an additional worker by being able to independently plan, act (with a little robotic assistance), measure and improve – this trend is only set to accelerate. So with a little help from my crystal ball, the following are my predictions for trends set to affect the MRO world in 2026.
Occupational Health and Safety:
Diagnostics tools will continue to evolve so that it no longer becomes necessary to “open” a machine in order to detect, diagnose and specify repair actions. This will increasingly be done via smart software, often employing AI and ML (machine learning) algorithms. Some “edge” computing may be required to detect impending failures, however increasingly machine operating data and repair histories will be accessible in the cloud. This means the AI algorithms for diagnosing and specifying repair actions can be cloud-based, which greatly facilitates regular updating of the software. Removing the need for maintenance workers to access machinery in order to diagnose faults removes workers from the line of fire.
Skills shortages
AI will do for white collar workers what robotics did for the blue-collar workforce. Robert Gottliebson, seasoned business analyst for The Australian, recently wrote that AI favours “doers” over the “thinkers”. We are already seeing this play out in declining enrolments in certain sectors such as computer science, where AI can now generate reliable usable code at a fraction of the costs of young graduates. This is likely to cause young school leavers to question the relevance of a university qualification, and to think instead of obtaining trade qualifications. And so AI may have inadvertently solved the industry skills crisis. The supply of newly qualified trades peoples will not be immediate, however, as they still need to flow through their three-year qualification period.
Agentic AI, or the ability of computer chat bots to plan, think and improve, is also likely to impact the industry. It could, for example, augment the skills of maintenance planners. A former PhD graduate working in data analytics in the US recently told me that, “agentic AI is like having 20 assistants to help me”.
Prescriptive maintenance
We are also likely to see the continued roll out of AI for decision support, expanding from predictive to prescriptive maintenance via cloud-based computer apps accessible from anywhere.
Prescriptive maintenance goes beyond the anomaly-detection capability of predictive maintenance. In addition to fault detection, it adds fault diagnosis, prognosis and decision support in the form of recommended courses of action.
To keep older assets operational, research is being directed towards establishing machine and component health indices. While some of these indices will be derived from AI or ML approaches, risk-based multivariate statistical approaches are equally valid. The advent of risk-based health indices will help extend the operational lives of older machines.
Integrated supply chain informatics
There is a need to enhance integration and visibility of information along the value chain, and across the verticals.
There remains a significant “gap” between predictive alerts occurring at the mine site and the ability of distributors to supply spare parts. This is in part because, for reason of commercial proprietary, information systems do not talk to each other. This causes repair times to blow out, resulting in wrench time factors that are low by world standards. How can we improve equipment repair times?
As an industry, we need to start measuring the components of repair time (wait time, decision time, logistics time, active repair time). Once we have baselined performance, improvements can be targeted. Sharing predictive information with original equipment manufacturers (OEMs) and distributors has the potential to significantly improve time on tools.
Robotic maintenance
While the last decade has seen large advances in autonomous mining equipment in the form of autonomous haul trucks (AHTs) and autonomous rotary blasthole drills (ARDs), the journey to automate maintenance tasks has been long and arduous.
About a decade ago, Rio Tinto trailed robotic tyre change systems. They are still not a regular feature of operations today – why? Early systems were not particularly reliable, requiring considerable on-site support. The systems did not completely replicate the range of tasks a tyre fitter can supply, so labour savings were not so evident. It is arguably easier to economically justify automation for production, rather than support tasks. Nevertheless, some interesting cases exist where robotic maintenance systems can save time, money and environmental and safety risks.
Once such example is robotic truck wash down facility. These not only reduce wash down times, thus boosting fleet availability, but they also reduce water usage and operator risk exposure. Another area where repair times might be reduced is engine turbo replacement. At present, fitters have to wait two to three hours for turbos to cool down prior to performing repairs. A robot can perform this work while the turbo is still hot.
By University of Queensland School of Mechanical and Mining Engineering, professor and discipline leader mining Peter F. Knights.
This feature appeared in the January-February edition of Safe to Work.