Can Shovel-By-Shovel Data Analysis Make Mines More Sustainable?
May 28, 2020
As published on Forbes.com, written by Jeff Erickson
Sprawling mining operations are feeling the squeeze — new challenges from the pandemic and economic slowdown are only compounding pressures that have been grinding on the industry for some time. These include mining ore deposits in ever harder to reach locations, and myriad changes aimed at improving sustainability in their business. One consensus: The way forward is to make operations much more efficient. They’re looking to new digital technology for options.
“We’ve seen that a lot of the mining companies now have built ‘innovation teams’ to really look for ways that they can mine more efficiently and sustainably,” says Frank Hoogendoorn, chief data officer at MineSense, a Vancouver, Canada-based tech company that’s carving out a niche in helping mines become more sustainable.
Mining companies have been chipping away at waste for years — with geologists, engineers, and data scientists building detailed models that point to where digging should be most fruitful. MineSense fills in a missing piece of the puzzle by showing operators exactly what grade of ore — whether copper, zinc, nickel, or iron — is in each dip of their massive shovels.
This information has two powerful effects: It can help miners avoid throwing away useful ore, and if the ore grade they’re digging is lower than what they modeled, they can avoid wasting time, water, and energy trying to process it. Either way, MineSense’s feedback helps improve resource planning. “We allow them to mine much more precisely and much more surgically,” says Hoogendoorn. MineSense is already working with some of the world’s largest mines in Canada, Chile, and Peru, and was recently named one of the Top 100 Global Cleantech companies.
MineSense works by attaching a device to the shovel bucket that uses x-rays to virtually sift through the bucket’s contents and inform the operator in real-time what’s there. Onboard the shovel, software does a quick calculation; is the ore grade as high as expected or not? That data flows to the mine’s fleet management system, which directs the load confidently toward processing or away from it.
That data also gets sent to a cloud-based data warehouse where MineSense data scientists have applied different types of analytical and machine-learning models. Those insights are shared with the mine operators in near real-time dashboards, Hoogendoorn says.
Getting the Job
It’s not a given that Minesense’s innovative approach would find acceptance with mine operators. “Lots of people think they have the killer solution for mines,” but tight production schedules mean shift superintendents rarely stop or slow operations and test them. “Our goal is to reduce the friction and be as noninvasive and nonintrusive as possible,” he says. “Particularly on the data side.”
That’s easier said than done. “We’re scanning constantly and produce a large amount of IoT data as soon as the shovel’s done digging,” Hoogendoorn says. While the system is making decisions onboard the shovel, it’s also streaming data to Oracle Cloud Infrastructure, “where we will continue to collect large amounts of data and scale up very, very quickly.”
Hoogendoorn’s team uses Oracle Autonomous Data Warehouse to store data for access by data scientists, and to deliver client-specific dashboards using Oracle Application Express, or APEX, which is Oracle Database’s popular built-in, low-code development framework. For Hoogendoorn, this is an unbeatable combination.
“As a chief data officer, my number one priority is to get clean data in front of as many people as possible so they can use it,” including PhDs in computer science, applied statistics, and geophysics who work for MineSense or for its customers.
Oracle Autonomous Data Warehouse deploys, tunes, secures, and backs itself up with no human intervention—letting Hoogendoorn’s team meet their ambitious goals while spending less on database maintenance and more on development and reporting using Oracle APEX.
“The way MineSense works is very dynamic,” says Adrian Png, from Insum, a firm that helps MineSense architect its data solutions. “They are ambitious about going out and ingesting new datasets” from mining operations, he says. “And their data scientists are always experimenting and doing discovery across entire datasets.” Oracle Cloud Infrastructure lets them scale up to handle the work, he says, and provides a playground of other services that MineSense’s inquisitive developers and data scientists can try.
For now, says Hoogendoorn, “Oracle Application Express is an absolute killer feature. As far as I know, the Oracle platform, particularly with Autonomous Data Warehouse, is the only platform out there with a built-in, fully integrated application development framework that lets you build and deploy applications all from within the platform, without having to install anything else.”
Hoogendoorn recalls the time his team was starting a high-stakes project with a large copper mine in Chile, which is run by one of the largest mining companies in the world. The team needed to get data in front of their research team, “and they were expecting that this was going to be a weeks or months-long project, and we literally did it in about two days,” he says. “We gave them a full-featured web portal with access to all the data, complete with interactive reporting and data visualization features, and they were just absolutely gobsmacked that we got that up and running so quickly.”
With Autonomous Data Warehouse and Oracle APEX on Oracle Cloud Infrastructure, “I know now that I can continue to deliver as we bring on big new clients and their analysts,” he says. “We’re going to get a data tsunami very, very quickly,” as more of the world’s largest mining operations prepare to use the system. “And I’ve got zero concern that we’re not going to be able to handle it,” he says. That despite the fact that his team is planning new types of sensors “that are going to give us completely new and quite large data feeds.”
“We are fundamentally a data company,” says Jeff More, CEO of MineSense. “The data we produce, combined with our advanced machine learning algorithms, is at the very core of the value we generate for mines. We needed a powerful platform to demonstrate and deliver that value, which we now have.”
It’s the combination of analytical power and easy access to data that makes it powerful. “I’ve got a world class, scalable super secure, super powerful database engine, and with the build-in application development tools, I can build and deploy applications right away so that I can get people access to data,” he says, “I don’t know of any other platform where I can do that ‘out-of-the-box’.”
Thus, Hoogendoorn and MineSense have the confidence to help move the rugged mining industry toward an efficient, digital future.