A client once told me that Big Data is like an exclusive gym membership; just because you’ve paid for it, doesn’t mean you will ever use it. Sure, when you signed up for it you made a serious commitment and every month, the privilege of membership has a price. It’s low enough that you don’t cancel and apply for a refund, but high enough to know that you are making a mistake.
From my experience in the trenches, Big Data is worse. At least with the gym membership, you technically can use it and if you don’t, your excuses are flimsy. The problem with Big Data is that it’s a big-ticket item that is purchased but is usually not even usable.
When do you reach for your Big Data? Sometimes never. Usually, when it’s too late.
Long before my career at SKF, I was a mechanical engineer and worked as a hardware specialist and systems support engineer. Every morning I would come to work with my passport in hand and by afternoon I would be on a flight to Singapore, Germany or the UK. It was always for the same reason: to troubleshoot a hardware failure in a manufacturing plant that led to machine and plant shutdown.
It was during this period that I first developed a love/hate relationship with data. The first task of a support engineer is to review historic sensor data in order to identify the root cause of a malfunction. Very often the only time that sensor data is ever accessed is after a machine failure has occurred. To (over) use our analogy above, it’s similar to visiting the gym after your first heart attack.
A factory owner or GM is likely to invest millions in a state-of-the-art manufacturing facility. Within the plant equipment lie sophisticated sensors that generate masses of data. However, these sensors mostly service the machines’ control processes during operations, and the data they produce is rarely utilized for predictive maintenance purposes.
Big Data is typically accessed after a machine has broken down and a post-mortem is conducted.
This data is usually stored locally and not shared outside of the plant. If you have multiple facilities, it’s unlikely that you are able to apply insights from data generated from one machine to another machine in a different plant.
Who owns the data? Theoretically, the data is owned by the plant. However, at a practical level, the machine vendor is often the gatekeeper of data that is paid for by their customers. I’ve had discussions with CEOs of large manufacturing companies that are hesitant to request access to their own data. Whether it’s fear of the unknown or a derivation of Stockholm Syndrome, the practical effect is that most Big Data simply goes unused.
Now let’s turn to the crux of this article: shameless self-promotion. SKF Enlight AI’s approach to Big Data Machine Learning offers factory owners a new way to gain operational insights from their existing data investments.
The Enlight AI cloud-based solution is able to continuously extract sensor data from the Historian database. Our advanced artificial intelligence algorithm detects abnormal data patterns (or correlations of data patterns). These insights are used to give factory owners advanced notice of emerging failure threats. At any given time, both operational staff and management can access a dashboard that details the state of each piece of plant equipment.
The confluence of cloud-based Big Data machine learning and the move towards digitalization gives factory owners a new opportunity to tap into their existing investments in data. Big Data will form the foundation of the emerging Smart Factory and it’s up to factory owners to figure out a plan to access, operationalize and ultimately monetize their big data assets.