MINING TYRES
Making tracks
From AI in tyre management and planning, to large tyre recycling and the latest tooling, Paul Moore looks at some developments in the mining tyre industry
Artificial intelligence( AI) is pervading many aspects of the mining industry and tyre management is no exception – a good example being its introduction into Kal Tire’ s highly successful Tire & Operations Management System( TOMS).
TOMS was released back in 2016- fundamentally it’ s a predictive planning and reporting system which has been growing in use rapidly as Kal Tire itself has grown. Mark Goode, Kal Tire’ s Mining Tire Group Director, Business Insights told IM:“ When we launched it, phase one was paper based – our service teams would do the work and write down the info, then an admin or supervisor would put that into TOMS. It’ s vastly evolved since then- increasingly, we’ re getting connectivity with other systems, whether it’ s customer dispatch systems, sensor systems, TPMS or Pitcrew AI thermal imaging. Secondly, we’ ve moved from paper entry to mobile. Initially we started using the standard app from the software provider but we’ ve now built and are continuing to evolve a Kal Tire specific app built specifically for TOMS and tyre service management.”
From a mining point of view, TOMS today is used on more than 150 mine sites, divided between open pit and underground mines. If you include quarries and waste management sites, the figure is much higher at over 1,000 locations. Goode:“ All these sites today can use TOMS via tablets or smartphones. One
Kal Tire has applied AI to its proprietary Tire Operations & Management System( TOMS)
of the real plus points of TOMS is that you can make data comparisons across sites. Historically, customer sites may have been on different systems from different suppliers and using different databases Now, because it’ s one system with one common data language, you have that capability- because TOMS uses the same data approach for example in terms of classification of tyre types and sizes, across all the sites. This can be a real benefit for customers with multiple locations.”
So where did the focus on AI potential in TOMS come from? Goode told IM:“ I would say we started thinking about the use of AI pretty early on. TOMS is well suited to it as a planning system. And when we rolled it out, we saw that the system helps you plan, but you also need to configure the system to build the work orders. You need to aggregate the work and that is very manual. The average Kal Tire supervisor is trying to manage at any one time 400 installed tyres and up to 2,000 installed positions on the bigger sites. Trying to identify and plan when all the various tyres need changing, linking them together, and maintaining sufficient stock is really challenging manually. We recognised we would need to put a lot of planners in place to support the sites to plan with TOMS which drives a lot of cost into the business. AI allows you the possibility to automate that, and ensure everything is included in the plan, using a degree of machine learning and algorithms.”
Goode says it is possible to code relatively simply a majority of what a tyre management planner does into the system and automate it. And use machine learning for the rest.“ We’ ve been developing it over the last three years and for the last 12 months we’ ve had it testing on nine sites. AI is today already within TOMS- it’ s already operational. We’ re now at the stage of enabling it at all our sites because the results of the test have essentially said it makes the job of planning far easier.”
It’ s also an evolving system.“ We have tweaked it a lot. And I think the key reason I would say why we did that is one of trust. When we started building this and when we got it to a certain stage, we were testing it fairly widely across a variety of different sites in different environments and cultures. I was very cognisant that if people didn’ t understand it, they wouldn’ t trust it. And if they didn’ t trust it, they wouldn’ t use it. So we wrote algorithms to encourage the AI to do things at a time that people would likely expect. We did not want to just develop a black box where people don’ t understand it in the future. Once we’ ve got a good level of adoption and people are saying yes this makes my life easier and the trust is built up, then it may be possible to introduce more pure AI based prediction.”
Goode says a good example is that it is industry standard to rotate the front tyres on mining trucks at roughly a third life, and within safety limits, you bring those rotation points forward or back to maintain spare stock.“ So that was something very easy to codify it into the AI. You’ re going to rotate at roughly a third life and you will move work backwards or forwards depending on your forecast of stock level. And when the model creates work orders at that time, our people can understand that.”
He adds:“ We’ re now developing predictive tools of when tyres are going to fail based on when they were fitted, the speed of the trucks, the payload, climate, the amount of rainfall, the pressure in the tyre, and what we pick up from Pitcrew AI thermal imaging in terms of damage; there’ s lots of things that we can see on the tyre that might allow us to predict when the tyre will fail. But we need to introduce this level of prediction in a staged way. By using AI, we’ re helping our folks get back to hands on work with less time behind the computer, and I don’ t think AI is replacing jobs, it’ s helping the people we have become more efficient without adding extra cost to our business or customers.”
As with any AI model the devil is in the quality of data- for a new fleet, the AI will be challenged but it will learn as it goes along. Goode:“ If I have a fleet of trucks and I add another truck, the model will make
International Mining | JULY 2025 41