Taking mining tyre management to the next level

I talk to a lot of people in my line of work, as do most journalists. And, although we rarely admit it publicly, we all have favourite experts to interview. You know the ones; super knowledgeable, speak clearly and concisely, give great quotes…

One of my favourites, and if you’ve followed my work for a few years you’ll already know this, is Dan Allan, senior vice president for Kal Tire’s Mining Tire Group.

Dan is a fount of information and has heaps of experience. So, I was delighted when he agreed to jump on a call in August to talk data collection and visualisation in the tyre management space.

“Today, miners have access to a lot more data. They’re trying to be very datacentric, especially the large ones like BHP, Rio Tinto… they’ve all made strides in terms of big data analytics,” he began.

“We’ve seen more and more customers go that total pit to port route, pushing towards autonomy or using data analytics to predict loads and schedule freights. Mines are much more connected now than they’ve ever been.

“The challenge for small companies like us [Kal Tire] is to stay relevant with all this stuff. We don’t make the tyres or the equipment, but customers look to us to have that analytic ability, to tell them what’s going on with their consumable product.

“Going forward, our value in that chain will really rely on us being able to aggregate data in a simple way, interpret it and feed it back to a customer in a manner that they can consume it properly, and that’s why we developed TOMS.”

TOMS is Kal Tire’s proprietary Tire Operations Management System. Launched in 2018, the platform collects and analyses tyre performance data from mine fleets to aid predictive maintenance and performance optimisation. The system is fast approaching deployment at 100 mine sites across the globe.

Making sense of big data

Allan continued: “One of your questions is about open-ended architecture and how easy it is to swap information back and forth… It’s getting much easier to bring in data from various systems, but I don’t see many that enable various sources of data to be easily interpreted and then fed back.”

It’s true. I’ve written multiple articles about tyre management over the years and one of the things that has always struck me as challenging for miners is navigating the number of brand-specific management and analytics systems available (almost every manufacturer has their own) and the volume of data that those collect.

As an operator, most likely with a mixed fleet of machines, fitted with tyres from multiple manufacturers and with different rims, how do you bring all that data together and get the most out of it? There’s almost too much.

“In my opinion, there’s way too much data,” Allan said. “We’ve learnt a lot about data collection; it’s not all relevant. You need to look for spikes beyond the norm, the brackets and anomalies.

“We’ve learnt on the tyre management side that with tire pressure monitoring systems (TPMS), heat recordings… that sort of stuff, you don’t need to know when its normal, you need to know when it’s not normal. You need to know the period preceding that and following that in terms of what’s happened.

“The idea of exception data reporting makes that a little bit more palatable but, for sure, mining companies are collecting tonnes of data I’m not sure they’ll ever be able to sift through.”

Better sensor technology

The other big challenge the industry faces is tyre sensor reliability and accuracy: false positives and negatives can easily corrupt datasets.

“It’s taken a long time for sensors to become more reliable,” agreed Allan. “TPMS sensors used to be in the range of 75% reliable; that’s nowhere close to being useful. It’s only in the last couple of years that has improved and that we’ve understood how to properly maintain sensors. So, you can imagine, all the data that’s been collected prior to that… 10-20% of it is just garbage.

How has sensor technology changed? I asked.

“If we go back to the history of TPMS systems,” said Allan. “We [Kal Tire] had a relationship with the Fuller Brothers back in the day, and I saw examples of mining tyre sensors that were literally the size of a soccer ball and would weigh around 14lbs.

“The sensor would be put into a truck tyre where it would roll around and collect information. These gradually evolved into system-provided sensors that didn’t have to mount inside the tyre; they used a proxy for heat. However, those had to be protected, so we started monitoring tyres on the side and on the rims.

“Both Bridgestone and Continental even tried baking sensors into new tyres; there was no sensor that could survive that exercise. Even now, you’ve got passive versus active sensor technology which has changed the game a lot.

“It’s been a long journey, but sensors are getting more reliable, they’re getting cheaper and they’re more common place. Manufacturers are slowly weeding out the things that don’t work and sticking to those that do.”

Data sharing & interoperability

To get back to the issue of interoperability…

The Goodyears, Bridgestones, Michelins of this world all have their own tyre monitoring systems. I asked Allan whether it’s possible for those types of systems to exchange data, or for one to analyse data from another manufacturer’s tyres…?

“It is,” he said. “And the issue is not one of technology, it’s one of purpose. The tyre manufacturers want the best information they can about their own tyre, so they design their system with that in mind.

“There are examples where we’ve seen – for instance – Bridgestone sensors fitted on Michelin tyres, but it’s not that common, and I don’t think the tyre manufacturers have set themselves up for that to be: a) a true profit centre on its own, or b) really reflective of what the customer’s needs are.

“Many of the tyre management systems on the market… they’re good systems but objectivity tends to be an issue for customers. And, from a manufacturer perspective, what happens to other brands of tyres is secondary. There’s a conflict of interest.”

There are also data sharing and security to consider. For example, if a major tyre manufacturer were to purchase a tyre tracking software provider or similar, then rather than a third party holding data on mining customer’s fleet performance, it would be in the hands of the tyre supplier.

Standards and ethics mean it’s unlikely that deals like this would be a problem but, in an age where data (knowledge) is power, it’s only natural to question data security.

Allan continued: “Interoperability and sharing of the data are not generally an issue. Getting it all in one place so that you can show a customer a comprehensive overview of the entire ecosystem is still a challenge.

“This is where some upstart data aggregators, companies like Newtrax which was bought by Sandvik, are making waves. It’s a really important segment that we need to keep our eye on and see where we can partner, because the gaps in these systems… that’s where the critical learning will happen.

“We’re never going to know more about CAT equipment than the Caterpillar people, and they’re never going to know as much about tyres as we do, but if, together, we look at how that tyre interacts with the suspension system or with the transmission, and how those areas interface, that’s where the real learning and value could be.”

Beyond basic maintenance

So, how could the data gathered be used beyond basic maintenance? What value does it hold?

“There are two areas that we’re hopeful about,” said Allan. “The first concerns automation. There’s not a lot of experience yet in terms of what components could have their lifetimes extended, or what’s impacted negatively because of duty cycles or wear. We’re hoping to learn more as more autonomous fleets are put into practice.

“The second is predictive analytics: the ability to say, ‘if this condition exists, then we can tell you with a degree of certainty that this event will eventually transpire’.

“With the massive amounts of data that we have, it’s still not simple to associate say, a tyre cut, with what happened in the preceding two hours, or the following two hours. And then to determine, if that condition exists, how long before the tyre will fail.

“That’s partly because of the number of variables involved but also the limited number of instances. Mining tyres traditionally last 5,000-7,000 hours, so it takes a while to build up the numbers so that you can start to make correlations.”

Which is also why automation is probably the biggest single optimisation opportunity for mines, not just in cost savings but in health and safety too; because operators are the biggest variable of all.

Incorporating AI

For TOMS, Kal Tire uses artificial intelligence-based analytics paired with optical recognition software – photos are taken of every tyre injury and fed into the system – to help train the algorithms on what data exceptions or conditions to look for.

Those are translated into alerts for operators and/or management staff which are received via a tablet or mobile phone so that swift action can be taken.

“We’ve taught the AI tool to recognise different types of tyre cuts, where they occurred, and the likelihood that they will result in a major failure,” explained Allan. “Traditionally, when a truck comes in for a fuel stop, service technicians will inspect the wheels, make sure that the temperature and pressure are correct, inspect for cracks etc.

“Imagine a world where the truck could pull into that fuelling station and, automatically, cameras perform those inspections. All the service technicians need to do is use the high level part of their brain to say, ‘no that is a smaller cut rather than a larger cut’, because the optical recognition systems have already told them what’s going on.

“That’s our vision of the future; not to supplement our team members with machines but to supplement their knowledge by having machines show them in advance, what they should be looking for.”

How far off are we from that given the rate of progress? I asked.

“I can’t tell you the company because we’ve signed an NDA, but some of that work is underway right now,” said Allan excitedly. “It’s very much a prototype project, but we are trying this with a major customer to see how far we can take the concept. Which is kind of cool!”

It really is. Because maintenance activities, which are very hands on and can require heavy lifting or working at height, are some of the most dangerous for people to work in. They are also some of the hardest to recruit skilled technicians for.

By limiting human involvement to problem solving and troubleshooting, mines can make better use of the skilled technicians they do have and keep them safer.

System evolution

I asked Allan how he sees tyre management systems and their capabilities evolving going forward?

“We’ve already seen, through Translogik, through some of the stuff that BKT has done, and through Michelin’s MEMS 4 system development, that TPMS capabilities have expanded,” he said. “They’ve gone from measuring heat and pressure to featuring gyroscopes and accelerometers… there’s a good blend of technology in these systems and they’re getting cheaper to install.

“Most are still pretty fussy to maintain though and that needs to be improved. For instance, if a sensor signal goes down, you’ve got to figure out what’s going on without bringing that truck out of service. So again, going back to sensor reliability, if it is not 95, 98, 99% reliable you could have productivity problems.

“I don’t think tyre management systems themselves will evolve much further, but I do think there are opportunities to aggregate the data they generate and use it to improve areas like haul road designs, maintenance, those sorts of things.

“We see that as a possible extension rather than an evolution of the equipment.”