Are you sold on the value of predictive maintenance (PdM) and ready to take it further? Case studies abound of vibration analysis, infrared (IR) thermography, ultrasound, and oil analysis helping to predict asset failure and prevent it via well-timed maintenance. It’s no wonder that PdM applications are surging.
There are many ways to go about scaling up a PdM program, starting with choosing which PdM tools to add and where, whether to rent or buy them, whether to staff or outsource services, and how to ensure management support and a budget.
Read on for scaling trends and best practices from three points of view: a reliability solutions provider, an automation solutions provider, and a management consultancy.
Common drivers, shared goals
Reasons to scale vary greatly. The most common are rolling out existing PdM technology to additional equipment or sites or adding additional PdM technologies to an existing PdM program.
“I’d say 70 to 75% of our business historically has been brownfield sites that were either highly reactive or very strong in PM and needed help on the journey to becoming more effective in leveraging PdM technologies,” says John Schultz, founding partner and executive vice president at Allied Reliability Group. Greenfield sites and corporate expansion sites also are candidates for scaling.
Allied Reliability works to baseline the health of assets before companies make an acquisition and help newly merged companies move production lines from one facility to another. In the latter case, Allied baselines the assets before they are taken apart, identifies defects that need to be eliminated before reinstallation, and then checks to make sure the assets were put in properly. “Any change in leadership can lead to the scaling or descaling of a PdM program, depending on the new management’s level of understanding and commitment to the program,” notes Schultz.
Today, scaling often takes place in tandem with leveraging emerging technologies and the industrial internet of things (IIoT) as well as employing newer practices such as prescriptive maintenance. “It’s more about how to add additional sensors and machine-learning algorithms, how many failure modes are covered, and whether the process is connected with ERP and historian data,” Schultz explains.
Chris Coleman and Ed Deuel, both specialist leaders at Deloitte Consulting, observe that scaling PdM allows for the realization of a true “smart factory,” which can then be connected to a wider digital supply network (DSN) for ecosystem-wide benefits. “Early detection of imminent component failures that can be automatically communicated through the DSN will allow companies to greatly reduce their on-hand spares inventories and move toward a just-in-time replenishment strategy,” says Coleman. Through vast network connectivity, the DSN will get smarter more quickly by analyzing component performance trends, adds Deuel.
Prior demonstrable success of PdM tools is a key factor in program expansion, says Bruce Hawkins, director of reliability excellence at Emerson. A track record of successful implementations, bolstered by early and frequent communication of these successes, especially is crucial, he says.
Rent or own diagnostic tools, insource or outsource analysis and maintenance services, conduct route-based or continuous monitoring – these are some of the considerations when scaling a PdM program.
Allied Reliability Group’s Schultz sees very little renting of tools. “Most companies are purchasing the tools or asking their service provider to bundle the cost of the tools into either their hourly or per-component rate,” says Schultz. He refers to options such as Flir mounting an IR camera on an iPhone or android device and GTI Spindle enabling iPads to collect good-quality condition monitoring data as examples of how prices have come down significantly and user interfaces have been simplified.
Whether to outsource or internalize is a cultural decision, he believes. “Some of our customers say outsourcing works perfectly well in their environment, so we’ll embed people in their plants to handle the program,” Schultz says. “Other customers say outsourcing doesn’t fit their culture, so we’ll teach and mentor their employees to be able to do what we do,” remarks Schultz. “We also have a talent acquisition business that fills gaps in a customer’s workforce until they are ready to build out their own internal team.”
A growing number of companies want to internalize the data collection and outsource the analytics rather than staff trained Category II or III analysts. “A qualified machinery lubrication technician can be taught that while they are at a machine providing lubrication and contamination control services, they can also collect route-based condition data,” adds Schultz. “Specialized analytics can be performed elsewhere.”
If the business case is there, Emerson’s Hawkins leans toward “owning the equipment for vibration, IR, and ultrasonics, because you then own the database of information contained in them, and it becomes easier to insource the PdM activities.” Using in-house resources allows greater flexibility in scheduling the routes and taking ad-hoc baseline readings following a repair, he comments.
However, the ability to make a successful call depends on training and expertise, which takes time to develop, so program effectiveness at the beginning may suffer. Also, tools and training are a significant investment that can strain budgets. “For these reasons, we often see organizations start with contractors and then gradually take over the program as the tools, training, and expertise are acquired,” says Hawkins. Third-party service providers can use their own tools or the client’s.
Managing challenges along the way
Numerous dynamics come into play with any scaling endeavor. Training is probably the biggest hurdle, Emerson’s Hawkins believes, because owning condition-based equipment does not translate to expertise, nor will sustainable results be delivered without the necessary and commensurate training. “Expertise can be internal and organic or third-party,” says Hawkins. “Many factors, such as local labor rates, demand for expertise, and working conditions need to be considered.”
Deloitte Consulting’s Deuel believes that perhaps the greatest challenge is making sure not to initially scale too fast. “Connecting machines to an IoT solution to create dashboard visualizations can happen fairly quickly, but it takes time to analyze both the real-time and historical data and develop the predictive algorithms,” explains Deuel. “This process also takes longer on assets that don’t tend to fail very often.” To mitigate this, he recommends bringing in professionals to help prioritize and implement the program and then staying the course and being patient.
As the volume of data generated by the PdM program increases, the DSN will need to be ready to handle it, so the synergies between these systems should always be considered as capabilities are scaled across the business, Deuel adds.
Another issue is the tremendous number of new startups to sift through – end users are having a hard time differentiating among these companies’ offerings, notes Allied Reliability Group’s Schultz. “A lot of people are sitting on the sidelines waiting to see who shakes out as winners and losers,” he says.
Many of the companies that are buying are buying big, and they want to understand how PdM plays into their overall corporate technology play, Schultz observes. “They ask, ‘How do I integrate all of this data and information into my supply chain, work processes, and ERP system and bring across all of my historian data from OSIsoft?’” he says. “They’re looking to large systems integrators to help with that journey, who will call in companies like ours for expertise on how assets perform, how they fail, what the failure modes are, what measurements are taken, and what is and is not a good measurement.” He adds, “These are larger engagements than I’ve ever seen before.”
Schultz also says he is seeing startup technology companies call on PdM domain experts when their own expertise is more about data science than maintenance or when their clients want to scale faster than the startup can manage following a successful pilot.
Words of advice
Digital solutions in this space are growing at a rate never seen before, and it’s easy to get caught up in the excitement of new technologies and the insights they can bring, cautions Deloitte Consulting’s Coleman. He believes it is vital to have a solid strategy and approach and not to overlook some of the basic tenets of maintenance.
“Technology doesn’t fix everything, but it is a fantastic enabler and accelerator,” says Coleman. Pure data that hasn’t been analyzed and transformed into information is useless, as is information without an associated plan of action. It is imperative to look at the process holistically to ensure maximum return on investment in a next-generation PdM program, he advises.
Emerson’s Hawkins urges confidence that when you begin looking for problems in your plant, you will find them. If you’re not prepared to address the problems quickly and before failure, then you not only risk your investment in PdM, but also you damage your reliability program’s credibility.
“PdM technologies in and of themselves will not advance you to top-quartile performance,” Hawkins says. “A parallel effort of implementing precision maintenance practices has to occur to fully realize the value of predictive technologies. The frequency of failures may well remain the same if you do not also address the underlying causes.”
Allied Reliability Group’s Schultz says his biggest piece of advice is to not underestimate the impact of culture. “The culture you build is at least as important as the measurements that you take,” he says. He favors PTC’s motto for the IoT: Think big, start small, and scale fast. “For PdM, I suggest thinking big about the level of coverage of the various technologies for your plant, starting small by establishing the processes and culture to handle PdM, and scale fast only after you meet your definition of success.”