Global site

ABB's website uses cookies. By staying here you are agreeing to our use of cookies. Learn more

Data analytics in mining - from a digital native's perspective

ABB DataLab for mining: creating value through data-driven process performance optimization

By Sophie Emmrich  LinkedIn
Digital Products and AI

Data lights the way to innovation

As a data scientist, it’s been really interesting for me to read the blogs written by my ABB colleagues, where they talk about their ‘boots on the ground’ experiences with customers. Although I have travelled the world with ABB, my primary focus has not been meeting customers in person. Instead, I meet them through their data.

Every data set tells a story. Finding these stories is what AI enthusiasts like me do best. And, perhaps more important, we can help avoid costly mistakes.

I was one of the first students to work in ABB’s DataLab for Mining in Cottbus, Germany, which was founded two years ago by Dr. Martin Handreg - as a workspace for co-creation with remarkable mining experts and data scientists. Our mission is to design new algorithms, systems and products that create value through process optimization.

The DataLab environment is one that really nurtures innovation. I’ve been with ABB since I was working on my master thesis, which explored machine learning’s ability to predict alarms within complex systems. My thesis findings were actually directly used in ABB products, which gives you an idea of how that spirit of experimentation works to benefit our customers. We are getting data analytics out of the labs into the real world, to the mining companies who are brave enough to scale, embrace agile, iteration, co-creation and change.

Ready or not, it's time to embrace data analytics

The readiness of customers for change varies a lot. And I am not talking about technological readiness but rather the right mindset to innovate. 

Because pushing boundaries means to be confronted with conventional barriers in people's minds. This extends to technological advancement in general. For instance, everyone knows someone –– and my mother used to be very much like this –– who refuses to get a smart phone because they “don’t know what to use it for” or they insist they just want a simple phone that makes calls and nothing else. Why over-complicate? 

As anyone with a smart phone can tell you, technology can add significant value to your life. So much so that you wonder how you ever got along without it. And when you add data and AI to the mix, you’re creating something very powerful, something that can make life easier, faster and more intuitive.  

Of course, once my mum finally got a smartphone, she loved it. She makes bank transactions on her phone without blinking an eye! And this is not because she has an engineer’s understanding of the technology but because she knows and trusts the practical things it can do for her. 

The industrial world has the same issue to overcome: data analytics for process optimization offers manufacturers an incredible opportunity. The main obstacle is not implementing the tech, it’s changing how we think about it. It’s about creating a clear context in which data works for us.

When miners meet data miners

Mining in Germany is a very traditional industry. From the 16th century on until today miners greet each other with the words “Good Luck”. They’re wishing each other safe return to the surface. And it makes sense: their  work is sometimes extremely dangerous, so even a casual greeting is colored by this knowledge.

When I think about this, it makes the work I do seem even more urgent. It’s my job to listen to customers and their data so we can create solutions that lift people out of that kind of everyday danger. Since mining evolved from working by hand with pickaxes to huge, automated machines, miners depend on the technology to perform their work with minimum risk. Sensor data is the voice of the machinery, telling us about dangerous situations and possible harms.    

The phrase “Good luck” also refers to the wish that new lodes (of ore) may be open. The mining work back in the days was accompanied with constant faith that next time down there will be the lucky day of finding treasures. And this belief is an explanation why ‘data science’ was called ‘data mining’ during its rising popularity phase in the early 00’s. Digging deep into the data eventually reveals unknown treasures. For the highly evolved mining industry the ‘treasures’ transformed from hidden lodes into hidden saving potentials, bottlenecks or reduced carbon footprints.

Once we start thinking of data as a digging ground , the possibilities start to get really exciting. The question stops being: "Why would I use that?" and transforms into “What possibilities does this create for me?"

ABB DataLab: the power of ‘trial ‘n’ error’

There will be a structural change of industries in Germany, mainly because of the future shutdown of coal-fired power plants in 2038. ABB and open-pit mining companies came together to explore new innovations and to find a way to share their knowledge and experience with mining companies all over the world - that's how the DataLab idea was born.

What was really magical about the DataLab is that it started with playing around with algorithms and solutions for customer pains and then evolved into a modern DevOps Software development team.

I call myself lucky that I could be part of this transformation. The process of developing a new analysis feels very much like “trial ‘n’ error”. Nobody knows for sure what solution might be best or what its full potential will turn out to be. As DataLab is small and agile, we can draw a lot from the experts within ABB. My colleagues have an incredible knowledge about the industry, from the simplest things, right up to customers’ big pain points. We're also involved in the product management, which includes research with our customers so we can figure out where a product’s value truly lies.

It’s an exciting time to work in the process industry because there’s so much to discover. Yes, Google and the finance industry are visibly shipping some highly sophisticated artificial intelligence, but there’s so much untapped potential in the industrial world.  

There’s a real sense that we are only at the very beginning of what the technology can offer and, together with our customers, we can actually shape the future of AI within this space.  

ABB wants to be your partner when it comes to digital transformation. The combination of long-term expertise in industries and integrated digital platforms is what distinguishes us from other big players. We add the newest technologies in data science and AI and present it in a way that non-digital natives can get in touch with easily. Thereby, we lower the barriers and smoothen the change.


Diving deep for big rewards

There’s so much power in data but one thing I think gets lost is this simple truth: no matter how sophisticated the offering, the usefulness of the products and service we provide is dependent on how the customer uses them. Because in the end, the software is only as smart as the users. Certainly, within my own work, the value of bold experimentation cannot be underestimated. And when a customer takes on that spirit, the results can go beyond anything we imagined in the lab. 

When developing and training models for specific use cases there is always the motto: “The more data, the better”. However, many mining plants only track and log a fraction of data produced by the sensors and systems. We learned to work with what we have – filling the gaps, enhancing the data quality. Nevertheless, a solid data base with many sensor inputs will definitely lead to higher accuracy and better results.

Data science is like sea fishing in deep, unchartered waters. At first you don't know what works… you try things out. . Fishing alone takes time and isn’t as effective as doing it with other people that have lots of experience. Together you explore the environment, test out scenarios and define scopes. And then then, finally, you catch that rare fish.  

I have high hopes not only for those miners with safety on their minds, but also for the wider world. Creating greener, more sustainable practices and optimizing production through tools like digital twins, wearables and advanced predictive and visualization technologies.

I’m looking forward to teaching more people to dive deep into their data. Together, we can catch new ideas, applications and potential for optimization. And, the more data that’s available, the better we can all get at optimizing our world.


Looking for your use cases

For the ambitious goal to provide more and more analysis services spiced up with machine learning we are always looking for new use cases from customers who are interested in this topic.


We start with the offline analysis and detailed report. If the outcome shows interesting insights and improvement potentials it directly moves into the production environment. This way, we are growing with our challenges and explore the boundaries of applicable and (really) helpful models.

Sneak peak

Next time we'll bring you behind the scene of software development together with ABB digital solution architect....

> Get our insights in your inbox.

  • Contact us

    Submit your inquiry and we will contact you

    Contact us