Predictive maintenance is a maintenance strategy that helps to check the condition of equipment, to predict possible breakdowns and to determine maintenance operations that should be done in order to avoid breakdowns.
Predictive maintenance can be considered as a part of preventative maintenance strategy. Indeed, predictive maintenance provides a very important information about the equipment that helps companies to determine the right time and the right type of maintenance. These two strategies support each other aiming at proactively increasing the reliability of equipment and its lifetime.
It is important to know predictive maintenance techniques that can help you to detect any signs of future breakdowns.
Visual inspection is often performed by every technician or operator of machine even if predictive and preventative maintenance strategies are not implemented.
This technique is based on detecting visually seen defects and signs of future breakdowns.
The main drawback of this predictive maintenance technique is that it highly depends on personal senses. We all use sight, smell, sound, taste and touch senses, but we treat the same properties differently and usually we need some kind of a reference or a standard to notice small variations.
Vibration analysis is one of the most popular technique of predictive maintenance. The idea of vibration analysis is that all machines that have moving parts generate some amount of vibration.
By detecting deviation in vibration technicians can identify problems with equipment, predict breakdowns and determine the right time and type of maintenance.
While vibration analysis is most commonly used on rotating machines, it can be applied for all machines with moving parts generating vibration.
All equipment with moving parts.
Rotating: electric motors, generators, centrifuges, gearboxes, centrifugal pumps etc.
Reciprocating: cylinders, pumps, engines, compressors etc.
Continuous or cyclical process machines: printing, production lines, packaging lines, plating lines etc.
Vibration can be detected using devices called transducers. Usually these devices have accelerometers for detecting vibration and are placed on the machine or its housing.
Transducers transform the mechanical energy of vibration into electric signals and send it to a data-gathering device. This device stores the data which can be sent to servers for further analysis.
The data can be analyzed using two main approaches: trending and comparative.
With trending approach, you can identify overall trend of changing vibration. For example, higher overall vibration can be a sign of the machine wear and necessity of maintenance.
With comparative approach, you compare the machines vibration signature against some baseline data.
Baseline data is a vibration data that was recorded when equipment was working without any deviations. Usually this data is stored when the equipment is new or after maintenance. If the actual data is different from the baseline data, you can do a deeper analysis and identify what part of the machine needs maintenance or replacement.
Thermography is a predictive maintenance technique that is used to detect deviation in the work of equipment by measuring its temperature. Usually infrared technology is used for measuring the temperature of equipment, determining deviations from the standard temperature and understanding where these abnormalities are happening.
Engines, power electronics, boilers, steam system components, transformers, capacitors, fuses etc.
There are 3 main types of infrared instruments that can be used in predictive maintenance:
Infrared thermometers. They can measure the actual temperature at one relatively small point of a machine. That’s why the application of infrared thermometers is limited and ideally they should be used in conjunction with other techniques.
Line scanners. They are very similar to thermometers and have the same limitations. The only difference between them is that line scanners can scan temperature along a line instead of in just one point.
Infrared imaging. This instrument is much more expensive than the previous two, but it also gives much more information about the equipment. Infrared imaging devices can scan the temperature of the whole equipment and show it in the real time or store it for the future analysis.
You should create a benchmark for each part of the equipment that will be analyzed with infrared thermography as a part of your predictive maintenance strategy. Then, systematically measure the temperature of this part and determine any deviations from the baseline temperature.
This is a predictive maintenance technique that uses fluids for analysis of the condition of machines. Most often these fluids are water and lubricating oil. If you regularly take samples and analyze the condition of lubricating oils, you will get two benefits.
First, you will determine the condition of the oils and understand when it should be changed. At big plants, even a 10% increase of oil lifetime can save huge amount of money. At the same time, you will be able to notice signs of early wear of oil and schedule next oil change in shorter time.
Second, you will be able to determine the condition of the machine itself. Here you can search for any traits of tiny particles in the oil. By analyzing the particles, you can understand which part of the machine wears and how to solve this issue.
Tests of oil sample include: Viscosity, contamination, fuel dilution, solids content, fuel soot, oxidation, nitration, total acid number, total base number, particle count, spectrographic analysis etc.
Each oil test gives you important information about the condition of the oil and the machine. For example:
Viscosity. Compare the viscosity of actual oil against the viscosity of a new one. Low viscosity means that the oil doesn’t create a film on the parts of the machine and doesn’t protect them. High viscosity means that the oil doesn’t oil all parts of the machine.
Contamination. Check if the oil has traits of coolant or water. If yes, the oil must be changed and the leakage must be prevented.
Fuel dilution. Higher than normal fuel dilution weakens the oil lubricating abilities. You should check fuel leakages, ignition, timing and other systems of the engine.
Wear particle analysis gives you even more information about the condition of the machine. In this test particles that are present in the oil sample are collected and analyzed. The analysis can give you important information about the type and degree of the wear.
With wear particle analysis you can identify the following types of wear: rolling fatigue wear, cutting wear, rubbing wear, severe sliding wear, and combined rolling and sliding wear.
The equipment for oil testing is quite expensive and usually only large companies can afford to buy it. Smaller companies that decided to include fluid tests as a part of their predictive maintenance strategy can do oil analysis in a laboratory.
Ultrasonic is a predictive maintenance techniques that relies on detecting changes in noise generated by machines.
The idea of this technique is that machines with signs of wear or breakdowns generate sound that is different from the sound generated by machines in good condition. For measuring the sound special ultrasonic sensors are used.
Ultrasonic as a predictive maintenance technique is used for the following purposes:
Leak detection. Ultrasonic is very powerful for leak detection. Flow of fluids and gases through a tiny whole produces a unique signature of sound that can be easily detected with this technique.
Materials testing. Ultrasonic is also the primary choice for material testing. Cracks in materials of which parts of the machine are made cause sound reflections. These high frequency reflections are easily identified with ultrasonic devices.
Airborne noise analysis. All plants are required to meet noise level requirements. Ultrasonic is a reliable device for measuring the noise produced by the plant. However, in this case it will be hard to determine the source of the high sound level.
This is a predictive maintenance technique that relies on measuring the level of wear to predict when next maintenance will be required. It is used for sliding, rotating and reciprocating elements.
For example, brake pads must be measured regularly in order to predict when they need to be changed.
This is a predictive maintenance technique that relies on the principle that the equipment performance drops if it wears or there are faulty components.
While this technique is very popular it is not efficient in all cases.
You should record the baseline of the level of performance when the equipment is in good condition to be able to compare it with the actual level of performance.
With this predictive maintenance technique, you will identify the faulty machine, but you will not understand what caused the problem and which part of the machine must be serviced.