For starters, it’s a highly accurate way of using data to cut maintenance costs.
Hardware advancements have helped Computerized Maintenance Management Software (CMMS) collect a lot of insightful data. With that data, businesses can predict when problems will occur, such as a piece of equipment breaking.
Here’s how we define predictive maintenance:
Predictive maintenance relies on conducting maintenance based on trends within equipment data. This technology is tied to condition-based monitoring systems for reading the output of an asset’s variables.
Whether a business needs to track a piece of equipment’s condition through temperature, oil viscosity or vibration frequencies, CMMS can connect equipment to a database to predict when it’ll require maintenance or a replacement.
But predictive maintenance isn’t just a CMMS feature.
The process of “predicting” when an asset will need maintenance ties into age-old reliability standards. What’s unique is how CMMS connects cloud technology, condition-based monitoring and predictive algorithms for giving a wide range of users a precise method of conducting maintenance.
CMMS has allowed more maintenance facilities access to this tech than ever before.
Both big and small maintenance facilities benefit from investing in a CMMS solution, but until now, only the big ones have cared about predictive technology.
For help understanding the applications of predictive maintenance, we’ve asked Fred Schenkelberg, Reliability Engineering and Managing Consultant with FMS Reliability and lecturer with the University of Maryland, to give us insight:
Predictive maintenance and condition-based maintenance are two buzzwords that overlap.
In general, there are variations of these definitions, and they all are applicable to some degree.
A Refined View of Your Assets
The benefit of predictive maintenance applies in cases where it’s indicated that failure is about to occur. When you observe variables indicating that the asset is degrading you can track these indications so that, at set intervals, you maintain the equipment. Then you can run a condition-based diagnostic on it to measure temperature, vibrations, current, chain stretching, or other variables that occur with a predictable pattern.
If you measure the asset regularly, you can plot an indicator variable over time and have sufficient lead time to order the parts and prepare for maintenance on the asset. Instead of vaguely understanding that the equipment is about to fail, you get a refined picture and reduced amount of uncertainty on why and when.
Condition-based maintenance concerns reading sensors, but it doesn’t include a formula for interpreting trends.
Predictive maintenance has an advantage over condition-based maintenance in that it uses both a degradation pattern and a formula to describe how the asset is wearing out.
Plotting Patterns of Wear
If the pattern of wear is something that can be plotted and tracked, it falls into the realm of predictive maintenance.
Then the solution relies on either automating the measurement with an electronic system or implementing a routine inspection method where you watch for the variables to reach a critical level and plan in the lead time to replace it or order spare parts for it.
The main idea is to get all the life out of the product before it fails, then replace it on a schedule so there’s no risk of downtime. Predictive methods give an improvement on when to schedule maintenance, but they require a bit more math than most people like to use.
You can see where the two overlap:
If you’ve got a good maintenance guy who intuitively knows the pattern, he’ll tell you when the asset needs maintenance based on an expert observation, but you otherwise need the right predictive formula for minimizing the risk of unexpected failure.
Predictive and condition-based maintenance are very similar and revolve around methods of preventive maintenance, but relying on just a fixed time for maintenance doesn’t take into account equipment wear.
How Does CMMS Tie Into This?
One of the best CMMS solutions I’ve seen had data-entry fields for technicians to input degradation values manually.
The system would then provide a graph indicating that there were so many months left until failure and give a plan for replacement on a set date if the equipment continued being used excessively.
This is the type of CMMS with a data-based formula that reports when an asset is expected to fail after the last 10 readings or so. CMMS systems often support these methods, but in my experience it has been pretty rare when they actually provide the useful information needed to the team through a formula.
The formula comes from reliability-centered maintenance techniques. RCM can be used at any level; just plotting patterns of measurements or comparing readings to predictive models; it’s more of a mindset though, that we can forecast accurately when a failure will occur.
CMMS serves to store data taken from the condition of assets, and becomes predictive when connected to the algorithms and analytic features for interpreting data.
Predictive algorithms are still relatively new to the CMMS market, but those standards are rapidly changing.
The capabilities of most CMMS solutions depend on target industries, and they have a variety of features along a spectrum of maintenance operations to choose from.