Have you ever heard engineers say, "that doesn't sound right; something's broken," based on how a piece of equipment sounds?
The first time an engineer said that was the first instance of "anticipating a fault based on current operating parameters of that equipment before it breaks down" or predictive analysis.
Predictive maintenance is not a new concept.
However, technological advancements have enabled various applications for predictive maintenance across several industries and use cases.
So, what exactly does predictive maintenance mean, and why is it important?
What is predictive maintenance? (PdM)
Predictive maintenance (PdM) uses condition monitoring tools and machine learning (ML) algorithms to predict potential failures, faults, and deterioration for assets and equipment.
Using historical asset performance and maintenance histories and predefined failure modes, PdM estimates the time of equipment failure, what exactly is failing, and recommendations to fix it.
This way, organizations are aware of potential problems and can line up resources to fix them just before equipment fails.
How does predictive maintenance work?
Predictive maintenance leverages combinations of condition-monitoring devices, hardware, and software to predict failures effectively and plan maintenance tasks before breakdown.
For instance, say the manufacturing guidelines recommend an oil change for your chiller after every 3000 operational hours.
As a preventive maintenance measure, a professional will dutifully change the oil on reaching 3000 operational hours–whether or not the oil change is actually necessary.
On the contrary, PdM uses Internet of Things (IoT), machine learning, and big data analytics to continually or periodically monitor equipment conditions to determine if and when an oil change is necessary.
Essentially, PdM uses real-time operating conditions data from the chiller to determine that an oil change is not necessary after 3000 hours but 5000 hours.
Further, it gives you a heads up when you have 500 miles left to go so you can line up resources and personnel for the upcoming change.
And voila, you just saved time and costs on unnecessary maintenance and also kept the chiller performing at its peak!
Here are some PdM technologies used for condition monitoring:
- Vibration analysis
- Acoustic analysis
- Ultrasonic analysis
- Thermal imaging
- Current and voltage sensors
- Oil analysis
PdM is used to schedule routine and predictable actions like inspecting parts, repairing or replacing parts, oil changes, and lubricating industrial equipment and machines.
According to a report by Deloitte, the use of IoT sensors in facilities can
- reduce maintenance costs by up to 25%,
- equipment downtime by 70%,
- improve productivity by 25%, and
- extend asset lifetime by several years.
However, choosing the right tools for each piece of equipment is critical to see the best RoI from PdM.
Which industries use predictive maintenance?
Predictive maintenance has applications in industries, including:
- Food & beverage industries
- Electric power
- Energy (including oil & gas, wind, and more)
This is not an exhaustive list; however, you can identify industries that can benefit from PdM if they have the following traits:
- They use sensitive and expensive equipment
- Have a lot of capital and operational expenses
- Suffer major business disruptions because of equipment downtimes
- Stand risk of worker safety
Wake up before break down: How to proactively manage facilities
A Quick Guide to Avoid Getting Pencil Whipped!
Why is predictive maintenance important?
PdM collects a wealth of data that provides insight into important questions like:
- What is the probability of an asset failing within a specific period?
- What is the remaining useful life (RUL) of an asset?
- What is the likely root cause of a particular fault?
- Which assets are at the highest risk of failure?
- What necessary maintenance should be performed to resolve the problem effectively?
These data points help businesses strategically plan resource utilization to achieve maximum uptime and productivity across portfolios.
The screenshot shows a detailed report for faults in a Fresh Air Handling Unit (FAHU) which was detected to be over-cooling.
Using Fault Detection & Diagnostics (FDD), the CMMS charts supply air temperature over a period of time, detects possible causes, and makes recommendations for fixing the over-cooling issue.
This way, PdM reduces instances of emergency repairs for unexpected equipment breakdowns, which are inherently more dangerous for maintenance personnel safety.
What’s the Difference Between Predictive Maintenance and Preventive Maintenance?
Benefits of predictive maintenance
The facility manager of a 29-storey office building said their company saved $16,742 in operating costs and another $32,300 in repair costs annually by deploying PdM for their heating, ventilation, and air conditioning systems (HVAC) alone!
Impressive, isn't it?
Early adopters of predictive maintenance have realized cost savings much more significant than their initial investments in more ways than one, as in the example above.
Some common benefits of predictive maintenance programs include:
- Improved asset reliability with real-time condition tracking
- Reduced maintenance waste by scheduling only the necessary tasks
- Efficient inventory of spare parts to reduce mean time to repair (MTTR)
- Perform maintenance on machines and equipment while they are operating to avoid disruptions
- Improved workplace safety for everybody
However, the use of PdM isn't widespread just yet.
This is because it involves substantial upfront costs for installing condition monitoring IoT sensors, developing predictive algorithms, and connecting them with computerized maintenance management software (CMMS) or specialized predictive maintenance software.
With wireless technology bringing implementation costs down, more and more businesses will harness predictive analytics in the future.
Learn how predictive maintenance can benefit your business
Implement predictive maintenance with a CMMS software
Implement predictive maintenance with a CMMS software
CMMS provides the most comprehensive and easily accessible source of historical information to get started with predictive maintenance.
It automatically creates and schedules maintenance work orders when sensors detect an asset operating outside its normal, predefined parameters.
These warnings prompt the maintenance team to take preventive measures before the equipment or asset breaks down.
CMMS facilitates the interpretation of data and serves as a central organizational tool.
Further, it leverages machine learning to:
- Build "normal" and expected event models
- Detect component problems in real-time and alert maintenance teams
- Predict component life and prioritize component replacement using predictive algorithms
- Optimize the frequency of scheduled maintenance activities and spare parts usage
Steps to implement a predictive maintenance program:
- Identify critical assets/equipment: Assets with high repair or replacement costs are the best candidates to show maximum RoI from PdM programs.
- Gather data from different sources: Collect historical data for chosen assets from your CMMS or hard and soft copies of maintenance records.
- Establish failure modes: Perform a Failure Mode & Effects Analysis (FEMA) to define failure modes for each asset.
- Implement condition monitoring sensors and tools: Deploy IoT sensors and condition monitoring techniques to collect data on the expected failure modes.
- Develop predictive algorithms: Create predictive models based on historical asset data and operating condition data gathered from sensors.
- Implement pilot and monitor continuously: Deploy PdM for your pilot assets and validate your PdM program based on outcomes such as increased asset reliability, decreased downtime, and reduced frequency of maintenance and equipment downtime.
Suggested read: How to measure asset reliability with a bathtub curve?
Stay ahead of the curve with predictive maintenance
The key to realizing the true value and cost savings from PdM depends on your ability to analyze available data efficiently and accurately.
Around 10% (or even less) of industrial equipment ever wears out, which means most mechanical failures are avoidable and can benefit from PdM to perform optimally.
Facilio's Connected CMMS platform leverages data from existing building automation systems (BAS) using IoT devices to provide facilities management teams actionable intelligence to fix assets before they break down
Further, it empowers you to optimize portfolio scale O&M in real-time with data-backed decisions.
Interested in learning how predictive maintenance management can benefit your business? Don't hesitate to get in touch!