The Quest for Improved Reliability: Industry 4.0 Technology & Predictive Maintenance

Industry 4.0 and the Industrial Internet of Things (IIoT) promises to bring changes to the way predictive maintenance is done and the implications that has for how we manage factories. Predictive maintenance isn’t new, with various forms of condition monitoring being used within industry for decades, but in the Industry 4.0 environment the concept is taken to a new level due to the volume and detail of data collected, the software tools available to analyse that data, and the technologies on hand to make use of the results as actionable intelligence. For businesses willing to make the investment – not just in dollars but in time and training to get the maximum advantage from the technologies – predictive maintenance promises to deliver greatly improved up-time, significant savings and productivity benefits.

Predictive maintenance offers clear advantages over both unplanned, reactive maintenance and traditional preventative maintenance. By monitoring for potential failure well in advance of it occurring and thus scheduling maintenance at optimal times, predictive maintenance aims to bring the occurrence of expensive unplanned maintenance dramatically down while also minimising the frequency of maintenance in general by reducing the need for preventative maintenance. There are obvious cost and productivity benefits that should be easy to quantify for businesses seeking to assess the suitability of installing predictive maintenance systems in their production environment.

Traditional condition monitoring systems make use of limited amounts of data and are seldom real time, with the collection and analysis of data lagging behind the plant’s operation. By comparison, predictive maintenance systems in an Industry 4.0 environment utilise increasingly cost-effective IIoT sensors to monitor all aspects of plant operation, connected by internal networks that can collect vast amounts of data and send to the Cloud for processing. This means such systems can provide analysis in real time (or near to it) and use sophisticated modelling to predict failure and provide actionable intelligence. The combination of IT and OT has clear advantages, since focusing on collecting and analysing data pooled from multiple locations across the entire factory infrastructure allows for the development of a holistic view of the health of the total system rather than the more limited and discreet insights provided by traditional maintenance systems.

The use of such technologies as digital twins – in which the real-time, operational data from a factory’s physical plant are used to create a digital simulation of that system – opens up new options for both delivering better predictive maintenance results and enhanced training of staff. Taking data from things like vibration and temperature sensors and running that through machine learning algorithms that can learn to detect small anomalies, signifying potentially larger problems, can help maintenance staff not only reduce downtime and improve Overall Equipment Effectiveness (OEE) but by using this data in a digital twin simulation they can model different operating scenarios that may lead to better operating outcomes.

Similarly, using technologies like virtual or augmented reality (VR/AR) alongside the real physical assets can provide enhanced training opportunities for staff. The ability of an industrial AR solution to guide comparatively inexperienced operators through maintenance tasks also has obvious appeal to businesses trying to streamline maintenance budgets and remove language barriers.

In fact, the very real advantages modern predictive maintenance can quickly demonstrate to business can be a double-edged sword. For manufacturing businesses, the most important question is always: are machines available to be operated when called upon? There is no doubt that a well designed and implemented predictive maintenance system can deliver greatly improved OEE results. But that very success may divert the business away from asking the more important question: would it not be better to have more reliable equipment in the first place?

The struggle for maintenance professionals to shift management's thinking from improved maintenance to improved reliability is as old as the industrial revolution. In highly competitive business environments, it is natural for management to focus on technology that can deliver a quantifiable benefit in the short to medium term, rather than a greater benefit in the long term. But beware, the data driven-approach of Industry 4.0, and the innate desire for businesses to use new technologies to save costs on maintenance and operator training could undermine the required shift in management's mindset. Like any algorithm-based technology, Industry 4.0 can entrench an existing data-centric view of the production environment that sees predictive maintenance as an end in itself, rather than a tool to be used alongside of, and to inform, well trained and deeply experienced staff in the decision making process.

The ability to more effectively predict and pre-emptively avoid problems – and see immediate benefits on the balance sheet as a result – may lead business to fail to recognise that the problems their systems are detecting could have been avoided by designing for reliability and that even if this entailed a greater upfront cost, the long term benefit to the business would be substantially greater. It is after all, better to see incremental improvements in OEE on a system that is already running at 90% reliability than substantial improvements in one that habitually runs at 70%.

And while Industry 4.0 technologies could open up improved training scenarios that could provide staff with increased insight into plant operations there will remain a temptation by management to see it instead as an opportunity to reduce payroll by relying on the predictive maintenance system itself rather than the expertise of those skilled in interpreting the data. It may also divert attention away from the other aspects of plant design and operation – such as how operators use machinery and how procurement specifies equipment for purchase – that can have a major effect on reliability.

Embracing predictive maintenance in the Industry 4.0 environment should be an easy sell for many businesses once they run the numbers but understanding that using the technology to complement and enhance the efficiency of maintenance staff rather than replacing them and that designing for reliability across the platform is the best solution to improved maintenance outcomes is likely to remain a more difficult proposition.


Greg Purcell

Greg Purcell

Industrial IoT Lead Implementation Engineer

Greg’s role involves leading the implementation of Facteon’s Industrial IoT solutions for our software product, COSMOline. With his prior experience as a Facteon Control Systems Engineer, Greg brings a unique perspective and a strong understanding of PLC’s. He’s also a certified integrator of Ignition by Inductive Automation. Greg plays a key role in defining and implementing Facteon’s modern day SCADA approach that aims to prevent SCADA becoming a bottleneck for machine data.