What is predictive maintenance and why should I consider it?
What is predictive maintenance exactly?
Predictive maintenance is a bit of a buzzword at the moment. There’s more than enough talk of predictive maintenance as part of the next step in Industry 4.0, but few companies have gotten started with implementation. Why is that?
One of the main issues is that, to most people, the topic of predictive maintenance is still vague. If you ask any service engineer or system integrator what predictive maintenance is, they will likely know how to answer that question. You will likely hear some variation of: “Predictive maintenance is estimating ahead of time when maintenance on a system should be performed to avoid machine failure.” Most explanations will not say anything about how to achieve predictive maintenance.
In this article, we’ll explain exactly how to set up predictive maintenance. But first, we should take a look at why predictive maintenance is even a topic you should be interested in.
Why you should even consider implementing predictive maintenance
Maintenance has been a part of machine manufacturing and system engineers for decades. Over the last years, we’ve seen predictive maintenance become a time- and cost-effective alternative to reactive or preventative maintenance. So what’s the benefit?
When you start using predictive maintenance as part of your service department, you can expect some major advantages. After you have predicted the future failure point of your system, you can start planning the work for your service department ahead of time. Components can be replaced just before failure - all according to the predicted plan.
In the end, you will experience minimised equipment downtime and maximised component lifetime. Added value for both you and your customers. You can even use predictive maintenance to broaden your service. Find out how to use predictive maintenance as a new business model.
The key to make predictive maintenance a success: start small
As is the case with most Industry 4.0-related trends, the key to a successful implementation of predictive maintenance is to start small. Sure, there are many tempting solutions for fully automating every step (e.g. Artificial Intelligence), but if you don’t lay the right groundwork you’ll likely end up creating a big mess that’s too complicated for anyone to actually work with.
We’ve previously mentioned what predictive maintenance is. In essence, it’s more a technique than anything else. It is a way of working. And you don’t start changing the way you work by purchasing a shiny new tool.
Following these steps will maximise your chances of a successful predictive maintenance implementation:
- Set up a strategy
- Start collecting data
- Determine the future failure point
- Configure alarms
You’ll now have everything in place that you need to start providing predictive maintenance to your clients! So let’s explore these steps in more depth.
1. Setting up a predictive maintenance strategy
Before you go and purchase or arrange any necessary tools, it’s important to make sure that you have a solid strategy in place. First off, you need to know exactly what predictive maintenance is and can do for you. So by reading this article, you’re already off to a great start!
When setting up your strategy you need to consider which systems would value from predictive maintenance. You need to be sure that it’s feasible and advisable to conduct predictive maintenance for your machine. It will, after all, take some time and effort to set up.
Lastly, you also need to identify the variable you’re going to track. This should be a variable that’s indicative of the machine condition. This is the variable that will need to be monitored for the health status of the machine. Due to automation innovations, this is a step that can now easily be automated right away. Which brings us to the next step.
2. Data acquisition, processing and storage
Now that your strategy is in place, it’s time to really get started with the implementation. Acquiring, storing and monitoring your machine data is a crucial part of the predictive maintenance journey.
It’s important to first establish a historical correlation between the selected variable and the component life. This is done by collecting data at regular intervals up until the time the component fails. Check out the various methods for logging your data.
3. Determining the future point of failure
With all the historical data at your disposal, it’s time to determine the future point of failure. Plotting the variable against time using a data visualisation tool will show you a typical curve. This curve will subsequently suggest when a component should be replaced. For example, if a failure occurs after the temperature of an electrical connection has reached 40 degrees.
4. Configuring alarms and notifications
You currently know everything you need to know to start predicting maintenance. Let’s stick with the example used in our previous point. Use your collected data to figure out at which temperature degree you need to send out a notification so that you can stay ahead of the issue. Simply configure an alarm and warn your service engineers about the impending failure ahead of time.
Now you’re ready to get started with predictive maintenance!
With all the important information in your possession, you’re now ready to get started with your implementation. Have you got your strategy ready to go? Implementing the next steps is a piece of cake with IXON Cloud.
Ready to start collecting and analysing your data?