Improving maintenance by adopting a P-F curve methodology
The P-F curve: One of the first, yet hardest, things to learn
By Kevin Clark, CMRP
The P-F curve is a way of representing an asset's behavior or condition before it has reached a failed state. It illustrates an asset's progression toward failure. On the chart, the x-axis represents the time to failure, starting with an asset's installation, and the y-axis represents an asset or component's resistance to failure.
Potential failure (PF) indicates a detectable state of failure, or the point at which degradation begins. Functional failure (FF) is when the asset or component has reached a failed state or no longer performs satisfactorily. The most important part of the P-F curve is the P-F interval, which represents the time between when potential failure is detected in an asset and when it reaches the failed state. The length of the P-F interval is largely determined by the technology used to detect failure.
Highly dependent on patterns
"Most assets follow a curve in their life cycle, from good to progressively worse, until they completely fail," says John Bernet, a product application specialist at Fluke with decades of experience in vibration monitoring. "The P-F curve is a conceptual representation only and does not have any units or scale."
The P-F curve is highly dependent on patterns. For example, if one motor ramps up to a specific rpm during operation, then an identical motor used on another asset should behave in the same way. The P-F curve is something that we-those of us in the maintenance and reliability space-depend on from a mature standpoint. As reliability engineers using the P-F curve, we have gotten to the point of trying to design out failure, of trying to indicate failure, rather than simply waiting for it to happen.
The way motors are built today, they have onboard diagnostics indicating how they perform-but not how they feel. Similarly, our human bodies are built to take care of themselves-but when things go wrong, we need doctors to help us determine what is wrong and what needs to be done to fix it. A motor takes care of itself. It knows how to run, and it knows how to deliver power and energy to whatever it is attached to. But when something starts to go wrong, it needs a diagnosis to determine the repair required.
The method and frequency of detection essentially determine the length of the P-F interval. The more often assets are inspected and the more sensitive the method of inspection, the more time there will be between detection of potential failure and when failure actually takes place.
Modalities of the P-F curve
Technologies and tools used to detect failure can include (earliest to latest):
- oil analysis
- motor testing
- physical inspection
Each of these methods of testing has something specific to say about an asset's operation, and the timing information has even more to say about the future of the asset. Using these modes of inspection and methods of detection can provide an early warning of decreased performance.
"Early indicators, like oil analysis and ultrasound, may just be a signal for additional maintenance actions, such as lubricating bearings, or beginning to plan and schedule maintenance to avoid surprises," Bernet says. "Late indicators, like thermography, may not be soon enough to prevent damage to the shaft, bearings, and components of the rotating machine."
The cost of maintenance commonly increases the closer you get to the failed state, as there is less time to mitigate or eliminate failure. And, at a certain point, there is no potential failure anymore-the asset has reached failure and must be repaired or replaced. It is also true that the further away from the failed state you detect potential failure, the more sophisticated (and thus expensive, considering both equipment and training) the detection technology.
Importance of timing
Sometimes taking corrective action too soon can have bigger consequences, such as higher costs or more downtime, than not acting. If you are repairing things too quickly, you are going to be spending money on changing out components more often than necessary. In such cases, you either have inaccurately identified the actual point of failure or have your parameters set incorrectly.
By detecting failures when they are actionable but still early, you are able to plan the most advantageous time to take corrective action. When used in conjunction with condition monitoring, the P-F curve improves maintenance by allowing you to do more than just react.
Oil analysis is one of the first indicators of potential failure. It tells you a lot about what is going on with a piece of equipment by indicating what kind of particles are in the oil. But not all assets have oil, so you cannot use oil analysis for everything. Vibration is typically the next earliest indicator. The most common thing to use the P-F curve on is equipment with rotating assets, which makes vibration a perfect technology to use. Rotation is something that is fairly consistent. When you have a consistently behaving piece of equipment, using the P-F curve makes a lot of sense.
P-F curve and vibration monitoring
All machinery vibrates, but excess vibration in rotating equipment can make potential issues known early on. Vibration monitoring can measure changes in the amplitude, frequency, and intensity of forces that can cause damage to rotating equipment.
"Compared to other technologies-like ultrasound, oil analysis, and thermography-vibration is kind of like Goldilocks," says Bernet. "Thermography can be too late, while ultrasound and oil analysis can be too early, but vibration is just right. With vibration, we can see indications of faults 12 to 18 months in advance-when there is still life left in the components-and not react too soon."
The four most common vibration faults are imbalance, looseness, misalignment, and bearing wear. Rare machine faults do happen, but almost all vibration faults fall into these four categories. "Vibration is especially effective in diagnosing mechanical faults," Bernet continues. "Even a healthy machine is going to have vibration, so it's easy to identify what is normal for a good rotating machine and then to be able to look for patterns in the change in the vibration."
Why condition monitoring pairs well with the P-F curve
Condition monitoring is the use of continual screening technologies to detect changes in the operation of assets, which means you are alerted to potential issues well before failure or downtime occur. It gives you real-time situational awareness of your operation and enables you and your team to schedule maintenance and corrective actions in advance. Furthermore, condition monitoring can provide a longer P-F interval than other maintenance methods.
"Reactive maintenance is all based on failures that have already occurred. Planned or calendar-based maintenance is based on taking some kind of corrective action globally but not having any indication of which machines actually need it," Bernet said. "Condition monitoring is based on knowing that a fault is coming-but has not occurred yet-and gives you some prewarning and the ability to schedule corrective repairs before the efficiency or capability of a machine is reduced."
Identifying anomalies before they result in damage, downtime, or disruptive repairs decreases costs, unexpected downtime, and production loss. Other benefits include extended equipment life and smaller spare parts inventories. And, when you have sufficient lead time to order parts on an "as needed" basis, you can even eliminate expedited costs.
Other maintenance methods, such as calendar-based maintenance, simply do not fit as well with the P-F curve. The real-time data provided by condition monitoring lets you stay on top of where each critical asset is on the P-F curve.
Tools to optimize
Maintenance is neither about fixing everything the moment potential failure is detected nor about waiting until everything fails before replacing it. It is about finding the right balance between taking action at the most opportune time for each asset and allocating your resources in the right way.
One way to manage the trove of data and insights you will get from condition monitoring is with a computerized maintenance management system (CMMS). This software maintains a database of information on an organization's assets and maintenance operations.
In addition to software like a CMMS, each organization should go through the process of performing a criticality analysis. This crucial assessment evaluates and classifies all of an operation's assets to provide a clear understanding of which require immediate attention, which can wait, and even which can run to failure. This evaluation will help clarify and inform your maintenance decisions and efforts. From a dollars and cents standpoint, the P-F curve applies more to critical assets.
A criticality analysis helps you identify how pivotal each asset is to the business. Think about it this way: You have two identical motors on different pieces of equipment, and they have two totally different rankings. That is because a criticality analysis considers the impact an asset has on product quantity and quality, on the environment, on safety, and on how that particular machine affects everything else.
When you do a criticality analysis, you are looking at every factor surrounding each asset and how integral they are to the process. Any time you start to analyze, you are likely to hear from your team that, "It's all critical. If I don't have any of these assets, then I can't produce products." It is not until you start breaking it down in detail that they start seeing how assets truly differ in importance. In general, assets are broken down into three different categories: A, high criticality; B, medium criticality; and C, low criticality.
Typically, 10 to 20 percent of assets are highly critical. Between 20 to 30 percent of assets are in the medium range. And the rest, up to 40 to 60 percent, are low criticality. The criticality analysis helps you understand how each asset affects your operation, and that helps you see where you need to spend your time, money, and energy.
The future of maintenance
Emerging technologies, namely artificial intelligence and machine learning, promise to take that kind of focus to the next level. With pattern recognition capabilities beyond human capacity, they will advance and enhance your maintenance operations.
Data is one of the most valuable things in the world-if you know what to do with it. One of the biggest benefits of condition monitoring is that, as you learn exactly how your assets perform and fail over time, you will be able to sharpen and refine your maintenance program. And by taking advantage of emerging technologies, you will have extremely powerful software watching for patterns and learning more every day.
For a long time, maintenance was seen as just a cost to doing business. Now, with the concept of reliability engineering gaining traction and new technologies becoming more widely used, companies are recognizing ways to achieve operational excellence and gain strategic advantage. Maintenance can help companies carve out a competitive edge, especially once they have started using the P-F curve.