In our previous blogs, we discussed the basics of the P-F (Potential – Functional Failure) curve and the cost-benefit tradeoffs of various maintenance approaches. We’ll now describe the measures that can be taken to discover failure indicators along the P-F curve.
The basic concept of the P-F curve is that as a machine or asset deteriorates, various symptoms/indicators emerge. The early-stage indicators may be harder to detect and may require more sophisticated and expensive systems to analyze, but they give you more time to take action to prevent a catastrophic failure. They allow users to choose times to service a machine when it’s less disruptive to the manufacturing process and when only minor maintenance actions, such as changing lubricant, replacing a filter or balancing a fan, are needed rather than major parts repair/replacement. The later-stage indicators may be more obvious and simpler to notice, but they may require extensive and expensive maintenance since greater deterioration has taken place.
Some monitoring methods can be done on a continual basis by using a permanently mounted sensor that takes samples at intervals of once an hour or more often. Others can only be done on a one-time or periodic basis, as when a sensor is brought in for special analysis, perhaps once a month or less often.
Common indicators and detection methods
This version of the P-F curve lists several common indicators and detection methods, in rough order of when they might start to reveal deterioration in an asset:
-
- Ultrasonic Spike Energy. Ultrasonic condition monitoring sensors are often expensive and used in portable systems to take one-time readings, but they can provide very early potential failure detection.
- Vibration Analysis. Sensors and evaluation tools can range from very simple and low cost to sophisticated and expensive. The vibration analysis is done on either a one-time, periodic, or continual basis and often gives an early insight into emerging problems.
- Oil Analysis. An oil analysis may signal the need for additional, relatively simple maintenance actions, such as lubricating bearings, changing lubricant, or scheduling maintenance. This can usually be done on a one-time basis, but perhaps periodically, such as monthly or annually.
- Temperature Analysis. This analysis can indicate emerging “hot spots” on a machine, such as bad bearings or excessive friction that signal a future failure. Depending on the measurement system and asset, it can be an early or a late indicator of impending failure.
- Pressure & Flow. These indicators can fall into either the predictive or the fault domain, depending on implementation. If a proactive approach is taken, they might be condition indicators that can provide an early indication of potential failure; if a reactive approach is taken, they might be indicators of a functional failure (failure already occurring).
- Audible Noise. Noise is often an indicator of deterioration moving into the fault domain, and requiring more immediate action than vibration, temperature, or ultrasonic indicators.
- Hot to Touch. Generally, once bearings, motors, or shafts become hot to the touch, failure is imminent and quick action is needed to avoid catastrophic failure.
- Mechanically Loose. This indicator may fall into preventative maintenance (maintenance performed at time-based intervals rather than based on need) and may not catch impending failures until it is too late. Parts, which are obviously loose, can indicate a deeper problem, often close to failure.
- Ancillary Damage. This detects when other parts of the machine/assets are being damaged prior to a catastrophic failure (for example, a damaged belt due to belt misalignment caused by a failing bearing). Generally, when this is found, it is too late to prevent the failure of the asset.
This list does not cover all possible indicators. Machine users and builders may have others depending on their unique application – other potential methods to detect asset deterioration include monitoring of current, corrosion, or leaks.
The “best” indicator and approach will depend on each user’s and each asset’s unique risk/cost/benefit profile. Machine builders and users should work closely with an experienced condition monitoring solution provider who provides multiple solutions to help consider and assess the tradeoffs associated with various approaches.
