FOR HIGH speed bearings, different established techniques are commonly integrated in a PdM program: vibration, temperature measurement, wear debris analysis. Low speed bearing monitoring is a different story. These “normal” technologies are most of time blind until it is too late when speaking about less than 250 rpm. So in this slow speed application early bearing failure remains a notorious problem – for those who do not use ultrasound.

Why Ultrasound?

For mechanical inspection, ultrasound is detecting and measuring the acoustic waves generated by the impact in bearings, for example friction or rubbing from poor lubrication. Based on sound, the technology is called Ultrasound because it uses high frequencies, above the 20 kHz high frequency limit of human hearing.The answer lies in two key elements: high frequencies and shocks which are the foundations of ultrasound. The two properties are particularly useful for low speed machinery for which, by definition, signals coming from bearings are typically weak.First, high frequencies are sensitive to impact. Further the ultrasound range is naturally unresponsive to low frequency phenomena such as the machines´ running speed, which will tend to mask bearing failure in vibration.Second, ultrasound detects the shock waves generated by bearing failure, more precisely, the localized energy released by impact. In case of low speed, these impacts are too weak to cause much structural vibration.Combining these two advantages, ultrasound is a selective method, highlighting impacts and friction from the background noise produced by low frequency phenomena. The user benefits are:

• Ultrasound is simple to use. Any technician can successfully monitor low speed machinery, not only a 10 year experience engineer. Even a beginner is able to easily detect when a bearing has failed, just by listening to the specific repetitive crackling sound.
• The result is immediate. You don’t need a long and painful acquisition time of several minutes just to conclude the bearing is in a good condition. 5 to 10 seconds are enough with your ultrasound device. Just think how many bearings you need to inspect.
• No complicated signal treatment is needed to perform both detection and diagnosis.

Sensor and instrument: the key points

Ultrasound contact sensors are resonant sensors. Their sensitivity is an important point, but not the only one. The basis of Condition Monitoring is to trend data and to perform a corrective action when an alarm is triggered. If you have no guarantee that two sensors provide a similar result for the same signal, then change supplier. Your historical data and alarm thresholds will become useless when replacing a defective sensor! Therefore, as for all technologies, the transducer characteristics: the sensitivity and the resonant frequency must be clearly specified and certified by the manufacturer.The same remark is valid for the measurement device. The instruments must be calibrated and interchangeable without inducing measurement variation. Ideally, the input, receiving the sensor signal, has to be under control, not necessarily the heterodyne audio output.The weakness of the signal provided by a low speed bearing requires a suited dynamic range and noise floor. An amplification of 90 dB is frequently necessary. Above all, an excellent signal-to-noise ratio (SNR) is essential. You will commonly have a reading of -6 dBμV for a healthy bearing, getting closer to 0 dBμV when failed. -6 dBμV means a signal of 0.5 μV and 0 dBμV corresponds to 1 μV. Your instrument must be able to extract this low signal from the background noise. You understand now why SNR is so important.

Graph 1. Specifications of SDT RS1T Threaded sensor.

Graph 2. The duration between adjacent impacts provides the bearing defect frequency.

Graph 3. Time signal for a healthy bearing.

Graph 4. Time signal for a defective bearing.

Graph 5. The duration indicates the bearing defect frequency. The diagnosis is outer race damage on this bearing.

Listening is the primary defense line

The first functionality of an ultrasound device is to transform high frequency to audible sound. The operation is called heterodyning. People not familiar with ultrasound instruments think this an old-fashioned method which is outdated. In fact it is not and this is especially the case when inspecting low speed machinery. An operator without deep knowledge can distinguish a healthy bearing, producing a quiet steady signal from a defective bearing, causing an intermittent or specific repetitive ringing or crackling sound.However, listening is not enough. Reliable measurements are required to build a solid PdM program. Otherwise, your instrument is no more than a stethoscope.

Static measurements for failure detection

Static measurements (overall values) are simple to implement as the result is just numbers. They are easy to manage for storing, trending curves and triggering alarms. For low speed bearings, two indicators provided by static measurements are used: the RMS, which characterizes the signal energy and the Peak, which characterizes the signal amplitude. Combining these 2 indicators, early failure detection is easy to find. Table 1 shows the data comparison coming from machinery running at 50 rpm, using an acquisition time of 20 seconds.In some cases, for example a complicated machine or repetitive unexpected failures, the user wishes to go beyond the detection process objective of determining the presence of a fault. The need is to determine which component is defective. This is the diagnosis process, performed using Dynamic measurements.A Dynamic measurement is data acquisition over a selected duration. It is used to process time domain and frequency domain (FFT or spectrum) representations. For ultrasound technology, time domain representation is the preferred tool because the time signal works with intermittent signals and provides powerful information on the nature of the failure and the failure severity.First, the shape of the time signal indicates the failure presence with the visual interpretation of the time signal, showing, for example, repetitive impacts. The time signal can be analyzed in an identical way to vibration analysis to provide an indication of the cause the failure.Like vibration, the rotation speed of the machine and a detailed machinery schematic are required. In the case of a later stage defect in a bearing, the operator can measure the duration between impacts. Comparison with defect calculations derived from the schematics, the defective bearing and therefore the bearing failure origin. Finally, the impact amplitudes are often a good indicator of the failure severity.

Example of Dynamic data results

Here is the time domain comparison coming from machinery running at 25 rpm, using an acquisition time of 5 seconds.Note: the same vertical scale has been used for the two time signals above. The duration between impacts corresponds to the Ball Pass Frequency Outer race (BPFO). The diagnosis is that the bearing will have some damage on the outer race.

Conclusion

Definitively ultrasound is an excellent method for dealing with low rpm machinery. The advantages provided by the technology are clear: the efficiency, the simplicity, the data collection speed, the possibility to combine easy detection with more elaborated diagnosis.I keep in mind the pertinent remark made by a Maintenance Manager:– My department does not have the budget to cover the expenses of a team of 5 vibration experts. In addition, with 5 guys, we would monitor only the critical machines. With our ultrasound instruments, the first line is operated by basic technicians, even the lubrication technicians, and most machines are regularly inspected. 80% of the problems found are solved using ultrasound. My single vibration expert is used to solve the remaining 20%.