Blake Griffin, senior analyst, Interact Analysis presenting on predictive maintenance strategies at PMMI's Executive Leadership Conference.
Although the ability to perform predictive maintenance analysis may not be everywhere today, packaging and processing equipment will standardize it in the future.
“If an OEM opts not to provide predictive maintenance, they risk obsolescence in the long term,” said Blake Griffin, senior analyst at Interact Analysis. “[Predictive Maintenance] is a trend that is unstoppable.”
Griffin presented the findings of a new PMMI white paper, “Packaging and Predictive Maintenance,” during the virtual Executive Leadership Conference today, April 19, 2021. Interact Analysis defines predictive maintenance in the following way. Monitoring a machine or a component on a machine to determine when it is likely to fail and take action to stop it. The goal is to avoid downtime.
“It is a technology built around data architectures and networking. But the definition we gave it [includes the] use of technology to gather data on assets, like temperature or vibration levels. And perform an analysis on that data to predict when the asset needs to be repaired to eliminate the risk of failure,” Griffin said. “It’s about the bottom line from the user perspective. If you can improve equipment effectiveness, you can improve profitability as a company. It’s about risk avoidance and optimization, those are the two key value propositions for predictive maintenance.”
It is a way for OEMs to reconnect with manufacturers to boost customer satisfaction in regards to service, Griffin said. He cited a previous PMMI report which found customer satisfaction with OEM service declined from 42% to 21% between 2015 and 2019. The decline was due to an aging workforce carrying a lot of native knowledge retiring and new, less experienced talent coming in. And predictive maintenance is essential in machines where manual labor can't replace the operation. This would include tasks such as filling and dosing, form, fill, and seal, and labeling, decorating, and coding.
With 45% of CPGs piloting or using predictive maintenance technology and 29.4% evaluating it, now is the time to offer it as a service.
Source: PMMI's Packaging and Predictive Maintenance report.
Business Models
There are two different approaches to predictive maintenance: sensor-based and software-based. The first version uses wireless smart sensors to communicate asset health. After deploying sensors, edge computing devices can tie into the system to transmit useful data to the cloud for analytics. The software-based approach, on the other hand, taps into the smart devices already on the machine, like the drives and PLC.
“If you have a lot of sensing components around the machine it is giving good information on how the machine is behaving,” Griffin said. “Also safety stops can be indicative of stress levels the machine is facing.” The software-based option must be flexible to adapt to emerging Internet of Things (IoT) platforms, as many automation suppliers are rolling out proprietary formats. “Don’t lock yourself out of a solution by being overly dependent on one. Think about the context of the entire operation, not just the context of your machine.”
Ultimately, OEMs will want to make money on a predictive maintenance analysis, but there are often obstacles in the way of that. For example, their customers may restrict remote access. There is also a fear of cannibalizing revenue associated with service level agreements (SLAs) or the fear of cannibalizing replacement revenue. “SLAs just need to be redesigned around a factor of uptime. If you guarantee 95% uptime, you price it based on how much downtime affected the operation. Anything above 95% is charged a premium,” said Griffin.
It is also a way of differentiating business, and, as end users see the value, they will lift those remote access restrictions. In addition, as predictive maintenance becomes common, OEMs can develop more complementary services or offer completely new business models such as machine-as-service.
