For manufacturers looking to take the next step in streamlining their operations, predictive maintenance deserves a high place in their priority list...
With a changing global marketplace, disrupted supply chains due to the COVID-19 pandemic, and increased reliance on data and analytics in business, manufacturers are rethinking their digital priorities to reduce unplanned downtime, prevent equipment failure, and reduce maintenance costs.
Today, manufacturers can be more proactive in their maintenance thanks to more accurate sensor technology, innovative analytics solutions, and increased integration between operational technology (OT) and information technology (IT).
As a result, maintenance engineers also see a shift in their roles, with expectations that they become less reactive and more proactive about equipment health and performance.
Predictive maintenance technology can prepare them to meet the challenges, and forums on Schneider Electric Exchange like the Machine Automation Community can help facilitate the types of partnerships that empower manufacturers to take their maintenance operations from reactive to predictive.
What is predictive maintenance?
Predictive maintenance, PdM, for short, is a type of maintenance strategy centered on monitoring equipment during normal operations to improve performance and reduce the risk of failure.
A well-planned predictive maintenance strategy combines data, technology, processes, and people to help organizations predict when equipment failure could occur. With a tool like EcoStruxure™Machine Advisor, for example, machine builders can access the equipment information from wherever their machines are operating.
In setting up a predictive maintenance regimen, an organization can set up and track conditions that trigger alerts on equipment performance. This is also known as condition based maintenance, which enables engineers to follow through on performing tasks to perform corrective maintenance.
This set up differs from preventive maintenance, which involves manufacturers scheduling regular check-ins to ensure equipment is operating as it should. However, it is common for manufacturers to combine predictive and preventive maintenance measures to help ensure overall equipment effectiveness, minimize or eliminate unplanned downtime, and optimize equipment performance.
Moving from reactive to proactive
Predictive maintenance machine learning
According to Klaus Kruppel, Vice President of Alliances at Senseye, maintenance has historically been a reactive activity triggered by an event, such as equipment breaking down.
In a reactive maintenance approach, an engineer identifies the cause of a problem and makes repairs to restore the equipment to normal operating conditions.
"In the old days of manufacturing, changes took place every ten years," he says. "Now you see changes several times a year, and these changes have a big impact on maintenance operations."
Kruppel explains that amid the changes and disruption in manufacturing and supply chains, organizations looking to gain a competitive edge need to embrace a more predictive approach to maintenance.
Predictive maintenance is a data driven maintenance approach that combines the power of asset information, engineering technology, the industrial internet of things (IIoT), artificial intelligence (AI), and machine learning (ML) with human insight to predict impending equipment failure.
Predictive maintenance approaches, such as IoT predictive maintenance and predictive maintenance machine learning, are a leap forward from more traditional forms of maintenance. Process analytical technologies used in integrated and flexible predictive maintenance systems leverage data from connected production systems and timely input from people to help ensure assets are always available for production when needed.
Predictive vs. preventive maintenance
An organization that uses a periodic or preventive maintenance approach will check in on the condition of their equipment on a scheduled basis, and as needed, and make repairs or replace parts to minimize the risk of equipment failure.
A downside to these maintenance approaches is that they rely more on dated “best practices” and broad recommendations that aren’t based in the latest IIoT advancements, and less on data-driven insights. Lack of data about equipment creates a blind spot for engineers, hampering their ability to see potential issues that could occur in between scheduled maintenance activities.
On the other hand, manufacturers that leverage predictive maintenance approaches, including the use of predictive maintenance machine learning, use real-time data to provide insights that can help an engineer determine when a problem may occur at any time – even when they are not on site and working remotely.
The result is a more agile workforce, less production downtime (up to 50% reduction in some cases), improved sustainability and a greater ability to adjust to changes and schedule critical maintenance work well in advance of a failure.
Solutions for predictive maintenance
Predictive maintenance approaches can cut down on the need for manual inspections and the travel they require. Senseye PdM’s advanced automated analysis can give a warning up to six months in advance of component failure, ensuring that parts don’t need to be expensively sent across the world, reducing environmental impact.
Manufacturers looking for predictive maintenance machine learning technology can learn about it in communities such as Schneider Electric Exchange. Senseye's predictive maintenance solution is integrated with EcoStruxure, Schneider Electric's IoT-enabled, plug-and-play, open, interoperable architecture and platform.
Kruppel explains that predictive maintenance can help manufacturers reduce unplanned downtime ensure their assets are always available for production when needed. In addition to ensuring overall equipment effectiveness, it can also enhance productivity and efficiency, maximize cost savings, extend machine lifetime, and reduce downtime. All of these factors together can yield significant business gains.
Working from trustworthy data
Getting to one version of the truth
As manufacturers evolve to adopt IoT predictive maintenance, predictive maintenance machine learning, and other predictive technologies, knowing about the vital role of accurate and accessible data is essential.
A key challenge for building an intelligent maintenance platform is disparate data sources and several systems of record. For example, information about equipment or machinery can live across 12-15 different data sources, according to Kruppel.
In this scenario, not even a well-presented dashboard can fix the problem of inconsistent or inaccurate data and datasets. The result is the user not trusting the data.
A manufacturer that lays the proper technical foundation to analyze its data and turn it into insightful reports properly empowers its leaders to make informed strategic decisions.
So, when organizations discuss digital transformation, they start realizing that they need one version of the truth. Technology is useful in this aim.
In fact, when people think of digital transformation, often technology is the first thing that comes to mind. But digital transformation is also about people and processes.
Kruppel explains that technology is an enabler, but that digital transformation starts as an idea. Therefore, the journey to digital transformation begins in people's minds.
Joining a community for innovation
How people drive digital transformation and adoption
People are responsible for championing and winning early technology adoption that will drive change in maintenance operations, improving processes from the shop floor to the C-level.
From a process standpoint, changes in people's behavior in the plant and on the production side are essential to take an organization's maintenance approach from reactive to predictive, according to Kruppel.
Additionally, through change management, people are introduced to the benefits of predictive maintenance, encouraging use and adoption.
Kruppel explains that people need to see the benefits of predictive maintenance, such as how it can help solve unique problems difficult to solve without data.
Take, for example, two production lines: one in Europe and the other in North America. These two sites may have the same equipment, but somehow, they behave differently across sites.
"You have to understand what it means, and that's where technology, such as EcoStruxure Machine Advisor, is useful in acquiring the digital fingerprint of the asset or the machine," says Kruppel. He explains that this data enables people to lay the foundation for intelligent processes which provide outcomes for the business.
Kruppel explains that technology like Senseye's solution leverages AI and machine learning to translate the data into a language that people can understand.
It is the only platform that builds models of machine as well as user behavior. The platform automatically learns the behavior of any industrial machinery, meaning that the maintenance team can easily understand why they may behave differently across different sites, while ensuring the platform only shows insights of particular interest to that user.
Data is collected from different sensor types (for example, through Harmony ZigBee sensors or hardwired to a controller). Then, the data is transmitted to Harmony Edge Box in the cloud, where the data is stored. From there, Senseye predictive maintenance can collect the data and start processing it, learning the asset’s behavior from the inputs, and start monitoring any changes.
In addition to Senseye solutions, our community for manufacturers on Exchange can help people find answers to address the significant challenges they face and collaborate on building innovative solutions for predictive maintenance.
So, are you ready to help collaborate to help solve some of the most pressing challenges in the manufacturing sector?
A multi-lingual Global Marketing & Digital Communication/Community Management specialist, mostly involved in Sales enablement & Digital transformation. Born in Morocco, grown up in France, studied in England, worked in the US, I like to consider myself as a world citizen who treasures relationship building, intellectual curiosity & learning agility.