Levels of advancement in prescriptive maintenance (PxM)

Maintenance has traditionally been reactive, where equipment breaks down and gets fixed. While functional, this approach has limitations in efficiency, cost-effectiveness, and operational reliability. With the growth of advanced technologies and data analytics, maintenance strategies have evolved significantly. One of the most promising advancements is prescriptive maintenance (PxM) (opens in new tab) —a proactive approach that leverages data, analytics, and machine learning to predict, prevent, and prescribe remedies to failures before they occur.

In this blog, we’ll explore the levels of advancement in prescriptive maintenance (PxM) through detailed examples and highlight how Fiix products fit into the picture. Additionally, we will review the future states of prescriptive maintenance and what those states mean for the evolution of manufacturing and beyond.

Level zero: Having the right systems in place

Before you can advance to the first level of prescriptive maintenance, you need a system to track your maintenance, assets, and parts. Teams often leverage a CMMS (computerized maintenance management system) (opens in new tab) to track maintenance activities. It keeps teams more organized than traditional pen-and-paper methods. Also, paper and pen can rely heavily on manual processes where a CMMS is built for succession planning.

This is the first level before advancing because a CMMS will help teams get into the habit of tracking their maintenance activities and having histories of maintenance for their specific assets. Having an asset’s history, as well as baseline knowledge about assets, parts, etc., is a requirement to move into level one.

Level one: Basic prescriptive maintenance

You have an impending failure. Now you need to understand how long it will take to fix, the criticality level, the parts you have on hand, etc. Maintenance teams can do all of this within a CMMS, but you need expert and baseline knowledge about your assets to input into your CMMS. This includes any knowledge about the product the asset is producing (or is part of making), parts for the asset, and routine maintenance done for the asset. Let’s review some examples that illustrate level one:

• Example one: A failure occurs on an asset, and you’ve had this failure X number of times. We can estimate it will take X amount of time from your previous data to fix it.
• Example two: If one of your assets’ motors is expected to fail, you need to use your spare part to fix the problem. Generally, this is fixed every 12 months, but you could use software to predict instead of estimating the timeframe. For example, you can replace the motor when Fiix’s Asset Risk Predictor (ARP) score is at a particular risk level.

In example one, typically this knowledge is known by experience maintenance team members, a CMMS will help capture it in a way that is shared and known by more people. The prescriptions (or recommendations) are relatively simple and straightforward in level one. In each example, the prescriptions are generally based on specific histories and conditions related to the asset.

Level two: Intermediate prescriptive maintenance using AI algorithms

You have an impending failure and use AI (Artificial Intelligence) algorithms to predict and prescribe remedies. You already have data and other variables being read from sensors for your asset.

• Example one: Data is read on a machine, and temperature and vibration readings suggest that the machine will not operate optimally soon and can impact production quality. The team is notified of the reading and prescribed remedies to fix it.
• Example two: An asset that has yet to fail is going to. You are notified, and the system notes preventative actions to take as soon as possible. In Fiix, this is done with ARP and the Fiix Prescriptive Maintenance product.

In level two, the prescriptions use data and AI to understand the potential failure and when (approximately) it will occur. They are more in-depth and provide the maintenance team with real-time data from sensors.

Level three: Advanced principles of prescriptive maintenance using in-depth data

You can predict changes in your assets based on the available data. You have data and other variables that may affect your assets in question, such as the temperature within and outside of your facility, the number of employees working for the day, etc. You can use more advanced AI algorithms from these in-depth variables to predict what may or may not happen to your assets.

• Example one: When you enter a new variable on an asset, you are notified on your system (in this example, let’s say a CMMS) of the changes. You can see if this new variable can help or harm and effectively predict the changes that may or may not occur.
• Example two: An error is detected on an asset, and you are notified on your system (again, in this example, we will refer to a CMMS). A series of prescriptions (or recommendations) are made to remedy the error. For example, it suggests that you stop running the asset every hour and take a 10-minute break because it is overheating. If the asset continues to run without a break, it will result in machine damage and downtime. The system also reminds you that the asset is due for scheduled maintenance over the weekend when production is stopped. So, in the meantime, if the asset takes a 10-minute break every hour for the next two days until the weekend, production will not be impacted. The asset will need to run additional hours on overtime during the week. With all this data and information on hand, you can continue running the asset without worrying about downtime.

In level three, you have the most detailed prescriptive recommendations. At this level, assets are most likely integrated with data devices beyond simple thermal sensors. This involves IoT devices and more advanced AI to develop a future prediction of failures and test and optimize failure potentials based on data and variable additions and subtractions from an asset.

Level four and beyond: A theoretical future of prescriptive maintenance

The future of prescriptive maintenance is one of the most exciting and promising topics from a manufacturing perspective. Imagine seeing all of your production facilities from a bird’s eye view and then being able to zero in on specific assets. When you review these assets in detail, you can predict their future failures and the possible root causes of those failures in accurate detail. To take it a step further, you can also see what machines will be used to fix future failures, the overall cost, employees on hand, and scheduled maintenance. You can even see a more in-depth analysis of the root cause of these predicted failures and leverage digital twin technology to test the prescribed solutions.

This is the future of prescriptive maintenance; it’s already here. Recent research highlights that PPC activities (including line balancing, capacity planning, and job scheduling) are expected to become more autonomous and smarter thanks to the vertical integration of physical and cyber layers in the shop floor digital twin [1]. It’s important to note that this research also speaks to the fact that the majority of companies do not see it as deployable in the near future and short term, but in the next thirty or so years, many changes are going to happen.

Jason Afara, our Director of Solutions Consultants, shares his sentiment on this:

“We’re not yet at the spot where machines are fixing other machines and sending reports back to us on their progress and status. But that doesn’t mean it isn’t happening to some degree,” says Jason.

He adds that there are examples today of machines working autonomously with other machines, “For example, some robots can work on machines and then self-diagnose that their battery is low and go and charge themselves.”

It’s an exciting future to think about, but behind it all, maintenance teams will still have to oversee all of the machines… those fixing and those producing.

The future of predictive and prescriptive maintenance is only just beginning. It illustrates a world where machines are assisting in maintenance strategy alongside technicians. Robots can take on timely interventions and risky maintenance tasks to contribute to a safer working environment and ensure regulatory compliance. This is just one example of the benefits of this potential future for maintenance and manufacturing.

From the basics of prescriptive to advancing even beyond

By progressing through these levels of advancement, organizations can enhance their maintenance practices, reduce downtime, and improve the efficiency and longevity of their assets. As technology continues to evolve, prescriptive maintenance will play an increasingly pivotal role in shaping the future of maintenance and reliability management.

References

[1] Kurt Matyas et al. (2017) A procedural approach for realizing prescriptive maintenance planning in manufacturing industries. Available: https://www.sciencedirect.com/science/article/abs/pii/S0007850617300070 (opens in new tab) (Accessed: April 30, 2024).