When I started my career as a mechanical engineer, I experienced the consequences of top-down decision making. Oftentimes, senior management seemed to ignore facts on the ground. I wholeheartedly agree with Peter Drucker’s statement that the view that “only senior executives make decisions or that only senior executives’ decisions matter” is a “dangerous” miscalculation.
In the past few years, I have pursued a non-linear, entrepreneurial route. I am older, hopefully wiser, and my perspective is more nuanced.
Let’s put aside Drucker’s legitimate concern. Executives need to balance both business stability and future growth. When technology disrupts the production and distribution value chains, existing models become redundant and sometimes even obsolete.
Executives from major industrial and critical infrastructure sectors are rapidly embracing Industry 4.0. Yet when I speak to a plant reliability engineer or a maintenance technician, the outlook is less optimistic. Specifically, there is a common belief that core elements of Industry 4.0 are not yet fully-baked. We are told that there is insufficient data to feed Artificial Intelligence algorithms that drive analytics. We’re also told that there is a lack of requisite skills at the factory floor level to operationalize executives’ strategies.
Within the smart factory, IIoT Predictive Maintenance is low-hanging fruit.
Why? IIoT Predictive Maintenance is an extension of existing maintenance processes (e.g., Condition Based Monitoring). The existing infrastructure, people and processes to repair evolving machine asset failures is in place. It does not matter whether the breakdown is predicted by a breach of a SCADA system threshold or the results of a geothermal analysis. Repair crews can fix machines regardless of why they break down.
If you accept my position that IIoT Predictive Maintenance is a no-brainer, then what is the purchase risk? In other words, what do I mean by “risking your business” in the title of this article?
How an Executive can make the Wrong Predictive IIoT Purchase Decision
One of my favorite Oscar Wilde quotes is: “Experience is simply the name we give our mistakes.” Since there are multiple ways to make mistakes, I will outline three of the most common ones to avoid.
#1 Avoiding an ROI calculation.
It is not unusual for startups to claim that Return on Investment is not relevant for disruptive technologies. Similarly, because there are so many line items in a downtime cost estimate, most industrial plants struggle with this calculation. I do not agree with either excuse.
New technology investments always need a financial justification. Yes, there is a lack of case studies for IIoT Predictive Maintenance. Nevertheless, high-level calculations can quantify the opportunity cost of lost production. O&M expenses can be estimated based on variables such as hourly wages and average wrench time.
What is the best way to generate an ROI? Conduct a Proof of Concept (POC) with an IIoT Predictive Maintenance vendor and to use the results for internal ROI forecasts. Here is an example from an SKF Enlight AI case study of a wind farm operator.
- Correct Alerts (True Positives): 38 out 40 alerts
- False Alerts (False Positives): 2 out of 40 alerts
- Missed Failures (False Negatives): 8
Alerts were triggered ~52 hours in advance on average before downtime.
Data from this type of POC can be extrapolated and applied to an entire production facility and then used to calculate ROI.
#2 Excluding Reliability and Maintenance professionals from the decision-making process.
Many executives view digitalization and IIoT Predictive Maintenance as strategic considerations. Ultimately these impact earnings-per-share and return on equity. This makes sense. If the fourth industrial revolution lives up to its name, new product offerings and business models will emerge.
We are already seeing examples of this in the industrial machinery category. For instance, more OEM’s are working to make their products “smart” and to migrate to a “Hardware as a Service” model. At the same time, bypassing local plants in the decision-making process can have disastrous consequences.
Here is just one example. There are multiple high-powered IIoT predictive maintenance solutions applicable to specific machine assets. In theory, these solutions can provide valuable, high impact insights. In practice, plants lack the internal capabilities to install multiple solutions that are a drain on engineers’ time.
We are a sponsor of the Future of IIoT Predictive Maintenance Study. What we learned from this study surprised us. Reliability and maintenance professionals oppose the idea of using multiple IIoT Predictive Maintenance solutions: 94% of respondents preferred one solution installation that fits multiple types of machines. Only 6% preferred multiple solutions for different machine types.
#3 Failing to develop a comprehensive IIoT Roadmap.
New IIoT technologies are released at a dizzying pace. It’s tempting to get caught up in all the hype. But executives need to consider a 5-to-10-year outlook and define their version of the future smart factory. There is not one inclusive definition of IIoT or even IIoT Predictive Maintenance.
At a recent conference, Frost and Sullivan evaluated 14 different IIoT Platforms. Each competes for the same industrial customers. The result? A complex IIoT landscape impossible to navigate. There are no obvious winners or losers.
Before embarking on the Predictive Maintenance journey, executives need to take a step back and consider the big picture. One of the interesting findings from the above-mentioned Emory survey of maintenance and reliability professionals was the high level of expectation that repair scheduling is likely to become fully automated.
If this is the case, then Predictive Analytics, Automated Failure Reporting and Automated Repair Scheduling will need to be fully integrated. Without this type of long-term vision, the likelihood of purchasing the wrong IIoT Predictive Maintenance solution increases exponentially.
Summary and Conclusion
Let’s go back to my original question: Can you acquire an IIoT Predictive Maintenance Solution without risking your business? The answer is yes, but it’s not simple.
With Industry 4.0, we are entering unchartered territories. There are no guarantees of success. However, there are guarantees of failure. These include bypassing financial analysis, failing to conduct due diligence, and poor planning.
Industrial plants generate terabytes of process data. SKF Enlight AI is a SaaS Predictive Maintenance solution that uses Automated Machine Learning to identify emerging asset failure patterns within this data. It provides early warnings and sensor-level intelligence to help avert unplanned downtime and meet production goals. For more information on how SKF Enlight AI can improve performance and reliability, click here.