Select Page

If you are to believe technology analysts and trade publications, we are entering the era of Industry 4.0. Machine Learning driven Predictive Maintenance will replace the need for Run-Till-Failure Unscheduled Maintenance. Here’s the problem: in the real world, not everyone shares this vision, especially folks tasked with deployment.
At Presenso we work with asset maintenance teams at some of the leading industrial plants in the world). We have listed some of the most common misconceptions about IIoT Predictive Maintenance and how to answer these objections.

#1 Artificial Intelligence is no match for Human Intelligence
Yes. Even today people will tell you that no computer can replace the experience of your employees. Maintenance crews know the machines cold and can figure out what’s causing a machine to fail. It’s the intuition versus Machine Learning argument.
How do you address this topic? The argument that Artificial Intelligence can analyze terabytes of data in real time may not resonate when emotions are involved. The reality is that in many industries, the experienced machine-whisperers are approaching retirement age and are not easy to replace. Progress is inevitable and Artificial Intelligence is playing an increasingly important role in today’s industrial plant.

#2 We can’t justify new technologies for old equipment
Many industries (oil & gas, paper, chemical) are struggling with equipment dating back to the 1970’s and earlier. The logic is clear: investment in expensive technologies to support end-of-life equipment may not worthwhile. In some cases, the argument is meritorious.
If you are sitting across the table from a vendor pitching a new IIoT infrastructure required to run a Digital Twin, expect a hefty price tag. Remember that if you go the route of the Digital Twin, an exact virtual replica of the physical machine must be created using blueprints. If older assets were modified over time without updating blueprints, the discrepancies the physical asset and blueprint add significant complexity.
Perhaps a more compelling counter-argument to objections related to equipment age is that software as a service (SaaS) IIoT Predictive Maintenance solutions do not require upfront investment in infrastructure.
A more accurate way of viewing Predictive Maintenance is that it can help extend the useful life of an asset. If anything, a SaaS solution offers the benefit of the Digital Twin without the upfront cost.

#3 We have extensive Preventive Maintenance programs already
It’s obvious that Preventive Maintenance is an essential element of asset maintenance. With an automobile, we have no choice but to take a car into service once a year. Based on mileage, transmission fluid, spark plugs, and oil changes are scheduled.
The comparison between an industrial plant and automobile is easy to make. However, its superficial and misses the mark.
A car is a single unit with a finite number of systems that can be inspected and serviced once a year. An industrial plant needs to operate continuously. About 80% of plant machinery degradation or failure are not based on age and occur for unknown or random reasons. The enormous volume and complexity of equipment is beyond the scope of a human mind.
Perhaps the most over-looked downside of Preventive Maintenance is that in many cases it does more harm than good. Human error such as using the wrong lubrication or misreading dials or controls leads to more breakdowns.
In a study of fossil-fueled power plants, it was revealed that the majority of maintenance outages occurred in less than a week after a maintenance outage (1,772 of 3,146 maintenance outages occurred after an outage). In other words, the error of one maintenance activity leads to the need for further maintenance.
No one is suggesting that we eliminate Preventive Maintenance. But don’t use Preventive Maintenance programs as a pretext for avoiding Machine Learning for Predictive Maintenance.

#4 There is no need for Predictive Maintenance for new assets
At face value, this argument makes sense: Our assets are new and are unlikely to fail in the foreseeable future. Predictive Maintenance is great, but it’s just not relevant for us.Machine Failure Pattern

The figure above depicts six different failure patterns. Although there is one pattern titled “Best New,” in both the “Bathtub Curve” and “Worst New” failure pattern there are higher incidents of Infant Mortality. These are caused by defects designed into or built into a product.
Based on these failure patterns, a different scenario for Predictive Maintenance occurs – that Predictive Maintenance can predict OEM defects or installation errors and prevent early failure.

Are you interested in learning more about how Presenso was able to send failure alerts to a wind farm operator, 52 hours before downtime? Click here to view our case study..

Like this article? Please share it using these icons....Share on FacebookShare on Google+Tweet about this on TwitterShare on LinkedIn
Avi Nowitz

Avi Nowitz

Avi Nowitz writes about the financial impact of Machine Learning on the industrial sector. Avi started his career on Wall Street as an equity analyst, providing institutional investor research on manufacturing companies. After completing his MBA in Finance from New York University, he joined the consulting division of Peppers and Rogers Group where his international clients included Bertelsmann, Ford Motor Company and Doğuş Holdings. For more than a decade, Avi led the Microsoft Consulting Practice at New Age Media and managed the execution of initiatives in EMEA and APAC.