After the excess investment and focus on production volume during the commodities supercycle, the new mantra of the mining sector is all about productivity.
And analytics have a big role to play in helping miners deliver, says Dirk Claessens, IBM’s general manager, global industrial products.
While the adoption of analytics has not yet happened in a big way in the mining industry, more and more data is being generated at mine sites from increasingly automated processes and the growing number of sensors on machinery. That increasing network connectivity of objects has been referred to as the “Internet of Things” (IoT).
Claessens explains that IoT is a platform or suite of technologies that is able to deal with real-time, massive amounts of data. That data can now be structured and analyzed, to help miners make better operating decisions — facilitating the next step in productivity gains.
In addition to having big implications for productivity, analytics have the potential to help mining companies become more responsive to changes in commodity prices and demand, to increase employee safety, and save money on energy and maintenance — and usually involves no capex and manageable opex costs.
In this interview, Claessens, who has worked with mining clients for more than a decade, helps us make sense of the potential for analytics to transform the mining industry, and explains why it’s mining companies themselves who need to define a vision of what they want to do with analytics.
Mining Trends & Developments: Can you fill us in on IBM’s history as a service provider to the mining sector?
Dirk Claessens: The mining sector has been there with us for as long as we’ve existed pretty much — IBM is 100 years old. We’ve been providing technologies to companies across industries for as long as we’ve existed. But really it’s the past five or six years when IBM declared analytics as one of its four or five major strategic imperatives, that we’ve become really relevant to the industry and started sharing our solutions proactively with the mining industry. It’s one of our fastest growing areas now.
MTD: Who are some of your clients in the mining space?
D.C: It’s a mixture — the main companies we’re doing business with are the mining majors, who for us are a nice mix of all the offerings we have. We have the more traditional service offerings with them — we’re outsourcing the back office for finance and HR for one of them — but at the same time we’re doing a lot of this more analytics sort of work. We have a broad offering around predictive analytics, both on the supply chain level as well as on an asset management and a number of other areas we’re discussing with them as we speak.
We have a mining major in Latin America where we have done a large chunk of their ERP (enterprise resource planning) implementation, which again is more of a traditional services offering. At the same time, we have our research engaged to deliver (an optimization project) using spatial-temporal analysis in an open-pit mining environment.
We are engaged with a number of second-tier clients in South Africa, particularly with platinum players. One of them is a company where you have a CEO at the top who’s very visionary and very technologically savvy. He wants us to work with him on this Mine on Demand concept where you integrate the mine design with the mine optimization and then optimize it from end to end over the life span of the mine.
MTD: Can you give me an example of how you use analytics to help your clients in the mining space?
D.C. One project we did was for an iron ore miner in Western Australia, where there was fleet of car dumpers at a mining port that failed unexpectedly, but no one understood the reason for the failure.
It’s hugely expensive for them to fail, it’s like 72 hours of down time, which cost this particular company millions and millions of dollars in lost production and lost shipping.
One car dumper failed, and then a second one failed, so we had a very rich set of inputs to really look at the root cause of this thing. We looked at the historic data weeks and months before the event: We looked at alerts coming through, we looked at outliers through the sensor data, and performance data — energy consumption, etc.
Not only did we find the root cause of what happened, we also saw that this could have been very clearly predicted a few weeks before the car dumper failed. The energy consumption of the one that was going to break down spiked up slowly but surely, trying to compensate for the fact that there was a problem with the bearings which was caused by a grease line that wasn’t functioning properly.
Data like this can use to provide maintenance advice.
MTD: What kind of return on investment can clients expect when they adopt one of your products or services?
DC: In the previous example, the 72-hour down time over two car dumpers would have saved anything between $5 million and $10 million. The cost of that project is a fraction of that, maybe $2 million.
As a rule, these analytics projects pay themselves back easily within a year.
We did an energy optimization at a crusher at a copper mine in Latin America. So, where is energy being spent, what is the energy balance, how can we start to become more predictive about energy consumption and fine tune energy consumption to the production schedule and anticipate so that you can take advantage of energy prices. The energy saving from that exercise was between 5 and 10% of the energy saved on an annual basis, which is extremely high.
With the Mine on Demand example, which involves a combination of better mine design and better scheduling of the tasks over the life of the mine, that would have added 3-4% of NPV over the mine life.
MTD: You’ve made the point that companies are not really seriously thinking about how analytics can help them drive productivity throughout the whole company. What are the barriers there — why aren’t they taking that step?
DC: I think they underestimate the challenge — there’s a lot more change management involved than people expect. We did a project with a miner in Australia who was asking us to be preventive but also prescriptive: If we want to get more lifetime out of this truck, what we can do so we can also finish this important order? For instance, reducing speed of the truck, reducing the load so it doesn’t break down. You would think it’s a no-brainer, but operators just don’t act that way: It’s extremely difficult to start entertaining differences in behaviour on the back of analytics, which come from a technology computer system. The proof of concept, the numbers are overwhelming. But when you present the results to the business unit leadership or whatever and they say fine, all of a sudden somebody comes out of the woodwork to shoot down the project at the last minute and the whole thing gets put on hold. As much as the bare numbers will tell you: ‘You should do this,’ you’d be amazed at how much resistance you come across.
MTD: How is the business model changing for mining companies and how can companies adjust to this new environment where they really have to focus on return on investment rather than just volume of production?
DC: The business model’s changing for sure. People through the commodity supercycle weren’t worried about value. The problem is during that time they built in structural productivity defects because all they worried a
bout was expansion, not thinking about how to really operate efficiently, and now people are trying to fix them.
The other thing I would argue, it’s not just productivity, it’s agility as well. Mining companies live in a world where a lot can happen — the price and demand are dictated by clients one step or two steps or three steps down the line, and so they just have to deal with it. That’s true to a large extent, but how quickly are you dealing with changes? Isn’t there a way to be much more agile and be adaptive around the way you design the mine, the way you schedule the tasks in the mine? When platinum and palladium prices change, for example, it becomes more interesting to go after the palladium than the platinum. Do you have process technology organization that allows you to do so? The answer right now is ‘no.’
People just do a mine design which is fixed once a year with a certain cutoff grade and they just start mining. Then two months later the world has changed, but there’s no way to deal with it. It’s a very static affair, and there’s a whole host of excuses but there are people around who really understand this should change.
On the agility side, for instance, with the variety of ore grade and minerals we have, how do we become much more responsive earlier to changes in both demand and price? Until you start mining with the order book in your hand, prioritizing, there’s a lot of work to be done, but we believe it can be done.
MTD: A lot of your clients are larger companies. What practical advice would you have for medium-sized or smaller companies with regard to how to use all the data that they have available to them in a smart way?
DC: I would say there are absolutely no inhibitors for small or second-tier mining companies to deal with this — provided they have a minimal quantity of data they can look across, there’s always insights to be had to optimize their productivity further. They just need to identify the issue or challenge they’re dealing with.
One of the platinum miners we work with for example, it’s a small company, a second-tier miner. They had a problem, to become much more responsive to the market they’re in — the platinum price was going up and down — palladium as well. So their CEO came to us for help.
I would say the advantage in this company is they’re much more quick on their feet in deciding, ‘We need to do this.’
So they can really be visionaries in the mining world. They have the opportunity to become much more agile in this respect compared to the big guys. Again, the capex outlay tends to be zero, the opex outlay, if it’s offered as a service it makes it really easy to digest and get going. If anything, I would say this suite of solutions is very suited to small companies because they’re quicker to respond, and because of the low level of investment required.
— This article originally appeared in the November 2015 issue of Mining Trends & Developments.
Be the first to comment on "How analytics is transforming mining"