How can we leverage big data technology to run efficient mines?

big-data

Big data example. Copyright: Kami Phuc

This week, I had a conversation with a colleague that led me to think about how little we are using the available data at mines to facilitate better decision-making. And, consequently, how little advantage we take of the large amounts of data available at a modern mine to help run more efficient mines. In this post, I want to share with you some of my thoughts on how the mining industry is using “big data” technology and the ways I think this can be improved.

Over the last decade or so, there has been a significant increase in the amount of data collected at businesses, in general, and mines are no exception. From equipment telemetry to enterprise software, mines have installed many systems that collect enormous amounts of data. Most of these systems are sold with some specific functionality and they provide the information and reports to achieve those goals. However, they often, as a matter of necessity or convenience, also collect a lot of other information from sensors and systems that are not displayed in the default reports.

For example, my research team at Missouri University of Science & Technology (Missouri S&T) has done research on energy efficiency using data from various equipment monitoring systems that were not necessarily designed for energy efficiency. In one instance, working with Drive Control Systems, we were able to build custom PLC code to store drag, hoist, and swing energy for each dragline cycle into three unused data slots. And that extra information added to the other information already collected by AccuWeigh we were able develop approaches for understanding operator effects on energy efficiency of dragline operations. I blogged about this work a while back and you can catch up on that here.

The volume, velocity, variety, variability and complexity of the data were are dealing with now is different for this data compared to what we were used to in the past. This is the challenge of using big data (extremely large data sets that may be analyzed computationally to reveal patterns, trends, and associations, especially relating to human behavior and interactions). In my opinion, the mining industry has not been able to fully take advantage of big data technology and analytics because:

  1. OEMs have limited innovation by keeping the data proprietary and restricting access;
  2. Access to the data for third-party analysis have been limited by warranty concerns;
  3. The development of analysis tools have been led by equipment manufacturers and has therefore over-emphasized equipment reliability over other mining business goals;
  4. Mine engineers often lack the data analysis (statistical, including data mining) skills to analyze such data; and
  5. Mining leaders have not fully recognized the strategic benefit of big data beyond what the OEMs are doing.

These issues can be addressed when OEMs, mine engineers and mining leaders recognize the value in extending data mining and big data analysis beyond the current applications. I have often made the argument to OEM executives that when they open up the data to researchers and third-party developers, they will increase the usefulness of their systems. The example I often use is how the mobile technology providers have increased the usefulness their systems by allowing third-party developers to develop apps.  They are, therefore, able to charge their customers for the mobile phones and tablets.

Mine engineers should seek out training opportunities to improve their analytical skills. Advanced courses, either as traditional university or short courses, are a good way to improve an engineer’s statistical and data mining skills. I find that most engineers do not have enough understanding of statistics to really handle large data sets with high variability. Most engineers I come across struggle with concepts like expectations, confidence limits, and hypothesis testing. If you are one of those, I will advise you look out for an online graduate level course (e.g. Missouri S&T’s distance graduate programs in mining engineering and other courses) or short courses by reputable consultants and professors (e.g. Sphinx Mining Systems).

We all need to continue the discussion on the value of big data. As that discussion continues, we will refine our views on its value relative to all the other things that take resources. If we decide it is important, then mining leaders need to provide resources and leadership to spur R&D to extend big data analysis beyond equipment reliability.