5 Best Practices for Big Data and Mining Maintenance

Whether miners realize it or not, they know quite a lot about working with big data. The knowledge and practices developed throughout a mining career are surprisingly helpful in identifying useful data and applying it to produce better results.

Miners are inherently collaborative. They know how to approach tasks based upon situational context. They have experience thinking of development from a long-term perspective. These are skills acquired from working in the mining industry, and it's these very same skills that can be applied to utilizing big data to promote technological change.

Develop Systems and Processes to Extract Maximum Data Value

The digital age has arrived for the mining industry, and the amount of data it is generating is massive. And yet, that wealth of data isn't paying off in the ways many would expect, even with such experiencing unplanned breakdowns and equipment lives that are not living up to expectations.

One of the biggest reasons for this is that people are diving into data without having the systems and processes in place that will allow them to extract insight from it and apply what they've learned in the field. Real-time data from onboard monitoring is useful and provides instant gratification to customers. Still, the payoff from it is minimal if you don't already understand how it will be economically useful.

Let's imagine that you've identified an imminent problem using data. Despite this valuable information, the issue will go unresolved if you haven't already implemented the systems and processes required to address it. Putting these initiatives in place ahead of time allows you to establish a connection between live data and real-world action to ensure that insight provides productive results.

Distinguish Big Data from Thick Data

Using technical data to determine the viability of an orebody is about more than just the data itself. Location, altitude, rail and port access, and other considerations must be evaluated as well. As a result, miners have become extremely good at understanding technical data within the context of these factors rather than just trusting it outright.

That same lesson can be applied to the use of big data. No matter how large your data sets are, they won't be nearly as effective if you're not evaluating them within the context of the mining industry and the people who work in it. This is where thick data, like work orders and rebuild reports, comes into play. In addition to quantitative cause and effect insights, thick data also considers the human element.

That isn't to say that you should only be using thick data. Big data and thick data must be used in conjunction to gain the most benefit from each. Big data allows you to identify patterns and develop insights based on them. From there, thick data is used to determine what steps need to be taken based upon what you've already learned.

Base Data Models on What You Know, Then Build from There

The most important outcome of big data is that it allows you to make smarter decisions, and that's why it's necessary to develop data models to guide you towards producing desired outcomes. You may not realize it, but you've already got all the information you need to start building useful data models.

Start by developing models that relate to high-value or high-cost events. These are the events for which you will have the most data, and as a result, they will produce the most accurate models. From there, you can extend these established models to patterns and events that are not as well understood while always keeping the known events as references to connect back to.

As you continue to train your data models, they will become more efficient at connecting data, actions and events. They can also be further augmented by the addition of human verification and oversight. This will allow the models to correlate data with real-world outcomes and actions, resulting in greater confidence levels and predictive power.

Break Through Organizational Barriers

It is all too common for big data within the mining industry to be viewed as separate from the work taking place out in the field. This results in an unfortunate rift being created between data scientists and technologists and those who operate and maintain the equipment they work to improve. This divide often leads to reliability improvement methods that are poorly implemented and end up not working at all.

That's why organizations must begin breaking down the internal barriers in their companies that separate those working in the field from the people using big data to make their jobs easier. By establishing greater contact between technology teams and operators and maintainers, those teams will gain the insights needed to develop improvements that produce real-world outcomes.

Human Touchpoints Are Critical for Success

Big data is invaluable for delivering actionable insights, but it's what you do from there that counts. No matter how valuable the insights you receive, using that information in conjunction with human interaction will yield the most beneficial results.

Consider this example of the importance of the human element when implementing technological change period an analyst detected a fuel dilution problem in a large shovel in a mine in Nevada. The Superintendent had already finished work for the day, but the analyst could track him down to explain how serious the issue was. As a result, the shovel was shut down and the mine was saved from $500,000 in parts in cost alone.

Yes, the data helped detect the problem, but there's no guarantee that the alert would have been acted upon without that human touchpoint. Even the most promising technologies will fall flat if they are not implemented in a way that works to incorporate human expertise into the process.

Contact DINGO for a brief Asset Health Consultation or use our Savings Calculator to estimate the annual savings that your operation could achieve with Dingo's industry-leading predictive maintenance solutions.