Moving away from break-fix maintenance or a regular maintenance schedule to a predictive maintenance model can look challenging and expensive at first. But it doesn’t have to be that way. You can create your new maintenance program one low-risk, controlled step at a time.
#1: Pinpoint losses for key machines
You should not aim to convert your entire maintenance operation in the first step. That would be overwhelming. Instead, build a strong business case for predictive maintenance by focusing first on business-critical machinery and equipment with the most significant losses.
It would be unusual for all of your machines to break down or produce substandard output at the same time. More than likely, the downtime or poor performance of a handful of machines has downstream and upstream consequences. Eliminating them could make a dramatic difference in your throughput.
Pick one of those machines for your proof of concept.
#2: Gather real-time data from the internet of things
Maintenance records and machine histories tell you what happened with your machine in the past, but do you know how well it’s working right now? You need to have reliable, current information to take the next step.
One easy way to gather meaningful evidence from a machine is using the sensors that may already be on your machines to collect and send data online. Connected to the internet of things (IoT), these sensors can highlight almost any machine condition – throughout speed, output quality, power use, idle times, vibration, wear and tear, and much more.
You now have real-time condition data in addition to your machine histories.
#3: Analyze, correlate, and understand your machine performance
Next, turn that machine sensor data into meaningful information. Data analysis at the level of the machine itself can show you how exactly it works and how certain parts and components contribute to its uptime, performance, and output quality.
You can use that information to replace parts before they break, or to make corrections before the machine’s output departs from acceptable tolerances. That’s how condition monitoring enables you to perform predictive maintenance.
What’s more, you can correlate your machine data to other business information. You can then assess, for instance, what the true costs of machine downtimes are, how a production slowdown delays shipments and forces warehouse workers to wait, or how a drop-off in customer orders corresponds to machine downtime or quality-related events.
Your data evidence may point outside the company as well. Maybe a supplier has been increasing your overhead by delivering materials that caused the machine to go out of spec.
You can take all those findings to get buy-in from company leadership for the new predictive maintenance program, which will involve advanced, accurate planning of skilled technicians, parts, and materials.
#4: Act on your evidence and observe the outcome
Before your machine fails or throughput and quality decline, perform the right preventive maintenance actions. Then watch the results. This may mean you observe nothing much happening – the machine runs as it should, with the uptime, performance, and output quality you aim for.
By taking the right step in predictive maintenance, you avoid the financial and operational consequences of your machine’s breakdown or poor performance. Be sure to document these successes, because they’re important in establishing the ROI and building the business case for predictive maintenance.
#5: Extend the predictive maintenance program and refine your approach
If your proof of concept worked out, be sure to share the story with your teams and executives. Then grow the program. Connect another machine or a group of machines, gather and analyze the data from condition monitoring, and use the data in your predictive maintenance planning.
As you refine your approach, unexpected findings may help you improve operations in the context of the machinery. Often companies realize that such factors as temperature, humidity, and dust can affect your machines. It may be outside of the realm of maintenance to improve these conditions, but the payoff can be significant.
You have also effectively opened a door for exploring IoT opportunities in your company. That can have far-reaching consequences that place you even further ahead of the competition than predictive maintenance does.
Learn how condition-based monitoring works