27 AUGUST 2019

Split Engineering began as a school assignment more than 20 years ago. Today it's an essential component of Hexagon's integrated drill and blast solution for the mining industry.

JS:  Split Engineering began as a school assignment more than 20 years ago. Today, it's an essential component of Hexagon's integrated drill and blast solution for the mining industry. Thanks for tuning in. Hi, I'm Jose Sanchez from HxGN Radio. Hexagon recently acquired Split Engineering, a leader in coarse rock fragmentation, size measurement systems, software and services. Today we have John Kemeny, head of the Department of Mining and Geological Engineering at the University of Arizona and co-founder of Split Engineering. Professor Kemeny joins us to discuss Split Engineering's origins and its importance to Hexagon today. We appreciate you taking time to join us, John.

JK:  Well, thank you very much, Jose. Again, my name's John Kemeny. I'm the Department Head of Mining and Geological Engineering. I've been actually at the University of Arizona for about 30 years and only the head of the department for about five years. In terms of what my role as head …

Well, the role is a transition from being a faculty member that does research and teaches classes to now managing the department. And there's a lot of elements to that. There's elements of trying to engage the faculty and engage the staff and get things done that way and also create new environments for the students and we do things that way and try to engage with the industry. That also is very important. Also engage with alumni. Alumni are very important to us and so the alumni functions that we do, we have one big donor that gave us an endowment and because of that we can provide nice scholarships to our undergrad students. And that's one very happy alumni that graduated many years ago. So we hope that that continues and our alumni are very satisfied and give back to the University.

JS:  The story goes that Split Engineering was the result of a school assignment. Tell us how that happened.

JK:  Okay. What I'd like to do is actually describe two homework assignments that actually happened, and if you think about the development of something like Split Engineering, it actually was in two parts. The first part is actually the research that went on to create the software that was developed for doing the Split program. And secondly, there was the business aspect and developing the business plan. So really it started really back in the early 1990, around 1990, and basically I was new to the University and someone showed me a picture of a muck pile. It was a photograph and they had, it had little circles around all the fragments and they were doing this by hand. And then they would take a ruler and they would measure the size and they had a scale in the picture and so on. And so it seemed to me, even though I'd never done anything with image processing, it seemed to me that this had to be, you had to be able to automate this process.

So I had two PhD students and I gave them the assignment, I gave them these pictures of muck piles and I said, "Let's develop some image processing software to automatically delineate." So we basically had two PhD students, two different approaches. One student took the approached was, "Well there's these things called edge detection. They must be, this must be, this is an edge detection problem, so let's use, let's use this, try that." And so this student tried various kinds of edge detection algorithms. The other one said, "We might have to be more creative here, you know, because people have tried the edge detection. It's a very complicated problem, this picture of a muck pile, a picture of rocks on a conveyor belt."

And so he went to the journals and looked not in rocks and not in, even in engineering, but actually in biology, and he found that there was this, this problem in biology where he had pictures of chromosomes and they had to be split. And so this, this one paper described how they would take these images of chromosomes and split them. And that's where the name Split came from, was the, this algorithm.

And the way it works is when I put two circles together and they kind of overlap a little bit, then they form these little sharp points and these sharp points point to each other. And that's the idea, the algorithm and it finds those sharp points, it splits the chromosome, in the same way we split these images of rock fragments and so on. That was one of the homework assignments. One PhD student did much better than the other one, so that was the algorithm that we kept.

The second homework assignment was later on after we developed software and we had these two people, Tom (BoBo) and Brian (Norton), that were interested in, and they were in the entrepreneurship program at the Eller College of Management and they were interested in developing a business plan for this. Their assignment was, you know, develop a business plan for some new business or take a professor and market it. And so, they came to me and asked me if they could basically develop a business plan and I said, "Sure, why not?"

So I went off to sabbatical in Australia and then I get a phone call saying, "Okay, we did that and we succeeded and we won some contest and now we'd like to start this business." And I was like, "Oh, okay." And so they were very successful. They wanted to start the business, I was in Australia, I said, "Well, let's wait till I come back and let's do it." And so we did, we started the company sort of based on this homework assignment that they had, which was develop a business plan for some new enterprise.

JS:  What were your expectations when you began exploring potential solutions to optimize fragmentation?

JK:  You know, this is what we did. I mean, I remember we had a stack at one point of these photographs of muck piles and each one we would delineate by hand and then we would use these different algorithms and try it out. And then we would just compare and see which was the most accurate. So eventually we developed an algorithm that was pretty good and that could actually, you know, automate the process that you would have normally done by hand.

JS:  Why is fragmentation so important to get right?

JK:  Okay, well that's good. I should also mention that my field, I come from a not image processing background. My field is rock mechanics or geomechanics or geotechnical engineering. And so I've always had an interest in blasting and I've had interest in the fragmentation you get from blasting.

So the idea is you start with a very complex rock, it could be a granite type rock or it could be some other kind of ore deposit, and it has fractures and it's, and it's quite complicated. Then you put some explosive in there and you fragment it and you get all these fragments. And so the idea is that you have three parts, you have the initial rock you start with, you have the type of explosive and the different blasting parameters, and then you have the end result. And so it's a loop. It's a loop really. And so what's really needed in the loop is you try different kinds of explosive and you see what kind of fragmentation you get. You try different kinds and you try different rock type. And so that's how this sort of loop develops.

Why is, why is fragmentation important? Well, the main thing that we wanted with fragmentation was material handling. You blast it and you had to be able to fit it in the truck that took it to the next process, whether it was dumping on a leach pile or whether it was taking it to a primary crusher.

So material handling was the initial reason that you had to optimize the blasting. But as time went on we realized that there were other things too. In this field, what we call Mine to Mill Optimization sort of was starting to develop as we started this. And what that says is that it's not good enough just to be able to put it in the truck or put it in the dipper of the shovel. You also want to optimize it for the crushing and grinding circuit or you want to optimize it for the leaching circuit. So, so then the idea became that how, what's the optimal size so that when it goes through the crusher, you use the least amount of energy? So that's this idea of Mine to Mill Optimization.

And then the third thing came along and the third thing really was slope stability. And so the idea was, well, if smaller is better, then you just keep putting more explosive, I guess, blast the hell out of it. And then, and you, that's the best thing for crushing and grinding, that may be true, but it's not the best thing for the slopes. So you really have this optimization problem. You've got material handling, you've got crushing and grinding, and you have froth flotation and leaching and you have slope stability, so you're trying to optimize all those. And fragmentation's the centerpiece of all those.

JS:  Describe your personal satisfaction at seeing how far Split Engineering has come.

JK:  Okay. In terms of personal satisfaction, I really just want to thank the unbelievable great team that helped make Split Engineering what it is today. And, and to the, to the point where it would get bought by a company like Hexagon. So yeah, I really want to thank the founders, which was Tom BoBo, Brian Norton, and actually Kristen Gerdner, who left the company after some years. Also, the people that have been with the company a long time, Cristián Rodriguez, he started the, really became head of the, the operation down in Chile 20 years ago. We also have our programmer Jeff Handy again, been with us probably on the order of almost 20 years. We also have Janice, Janice Meyer, she's been our office manager that we stole her from the University, she was actually office manager at the, in the, in our department at the University, she's fantastic. And then you have the people in the Tucson office, other people today in the Tucson office, Don Cramer and Jonathan Dufek, Sean Klasen, unbelievable. And then the, the people that run the office in Johannesburg, Edgar King and Dean Riley. So I just want to thank all them and all the students that have helped throughout the years.

And I should say that the research didn't stop when the company get started. So we continue to do research and write research proposals from the company, we are small company so we can write what's called SBIR, Small Business Innovative Research proposals.

And so I want to just mention that we talk a lot about fragmentation, but another piece of software that also came out of one of these small business research projects was what we call Split-FX. And Split-FX is actually very important to the process as well because if you go back to the, to the idea of the, that what you start with, you start with this complicated rock and the Split-FX program basically delineates the fractures. It uses point clouds to delineate the fractures before you blast. So it really provides some valuable information before you blast. Then you blast, then you have this split analysis to look at the fragmentation.

JS:  What has Split Engineering success meant for the University of Arizona?

JK: Well, I think because we started Split Engineering quite a long time ago, it was over 20 years ago, so I think it's had a pretty big impact. Tech Transfer was, was just getting started I would say about 20 years ago at the University of Arizona. And so, I think that this Split starting of the Split Engineering company was an early, what I call, Tech Transfer success. It brought royalties into the University of Arizona with, you know, steady royalties over a 10-year period. It also brought visibility, you know, and started this idea that that, you know, that professors could start to do Tech Transfer. It also had an impact on the MG Department. I think that, you know, by seeing a successful Tech Transfer, other professors started to get interested as well. And, and there was, you know, it was a big part of my, my life back then.

I should mention that because I was a professor, I always was, limited my, my, the hours that I worked at Split Engineering, that was kind of dictated at, at the time that we started the company. I was basically, you know, eight hours per week or you know, so on. And so, I didn't have the same role that the other people at Split Engineering had. But again, it was a very important role. So again, I think in the MG Department it had an impact in today, as I mentioned before, almost all the professors or maybe all of them actually have some kind of Tech Transfer. We have a very big organization on campus called Tech Launch Arizona. They really get involved with professors early on, let them know that Tech Transfer is important, that if you have something that's patentable or that this could start a spinoff company, let's, let's hear about it. They also provide funding. We didn't have that opportunity to get some initial funding. We actually had to go out to the industry and we were very lucky that we did get funding at the beginning. So we were always able to continue the research to get it to the stage where you could do the Tech Transfer, but now they will actually provide some funding for that, which is great.

And then the last thing I'd like to say is that really in, is particularly in the College of Engineering, there's this concept of, we've always had this concept of entrepreneurship. Entrepreneurship is, "Okay, I have a great idea. I'm going to start a company, I'm going to start a, you know, start a business to sell ice cream, or I'm going to do this, I'm going to do that." But there's another thing called entrepreneurial thinking. And so even if you were getting a company like Hexagon, there's still the importance of entrepreneurial thinking, which means you're thinking like an entrepreneur. You're going to develop a new product and when you develop that new project, you might give it some sexy new name and you might figure out how am I going to make money and how is it going to integrate with the other things. So, so I think that all that in, you know, in the last 20 years has been sort of a very important process at the University.

JS:  What do you think being part of Hexagon means for Split Engineering's customers and for the mining industry?

JK:  I'd just like to say that, you know, Split Engineering getting acquired by Hexagon is outstanding. It's a perfect fit and it helps one another. I mean it's, I can't imagine a better fit than Split Engineering and Hexagon in terms of, of not, you know, the software, the software technologies that they have as well as the hardware technologies, they all fit in. So, I think again, there's sort of two things I'm thinking of. I'm thinking of the fragmentation software that fits in so well with the blasting models that, that Hexagon is working on. But I also think about the Split-FX program and how it fits in with this whole idea of characterizing the rock mass and figuring out about slope stability and things like that. So both those programs, I think just unbelievable fit.

I think that in terms of the clients that have been developed with, with Split Engineering, I think that they overlap so much with the, with the client base that Hexagon has that obviously one helps the other. So I can imagine there's a lot of the synergies there. Also, from the software development point of view, I think that again, a lot of synergies, I think it may be in not so much today, but eventually integrating all the Split's software right in to the Hexagon software, is going to be fantastic.

So again, as I mentioned before, it's kind of a closed loop. The whole, the whole system is a closed loop. You start, you characterize the rock, you start to excavate, you start to blast, you get the mineral product, it goes through the mill and it's a closed loop. You learn, you learn how to do it better, you optimize. So Split Engineering is definitely an integral part of that.

JS:  A big thank you to our guest, John Kemeny. For more information about today's topic, visit hexagonmining.com. To listen to additional episodes or learn more, visit hxgnspotlight.com. Thanks for tuning in.