We at SPM consider high-definition monitoring to be a modern way of monitoring bearing conditions. There is no need for special hardware other than good, clean ADD processors and standard, good-quality ICP-compatible transducers.
There is, of course, also a proprietary SPM HD, or high-definition, in which we use a Shock Pulse transducer, which produces signals that are sensitive to elastic waves.
There are many factors that make up HD technologies. HD technology is neither magic hardware nor a single technique; it is a combination of techniques. There are 37 patents around this, but many of the principles apply whether you are using PeakVue, Spike Energy, Shock Pulse, or demodulated vibration data.
We will focus on four characteristics of HD technologies: pure digital filters with high signal-to-noise ratio, HD Order Tracking, symptom enhancer, and HD Real Peak.
Pure Digital Filters
We have made a significant effort to replace the analog components in the vibration hardware interface with modern technology. Analogs generate a lot of noise that can mask the distinct signals we are looking for. This core technology involves using high sampling rates, steep digital filters, and smart algorithms with a good signal path from the transducer.
As a component degrades, the signal moves from a higher frequency to a lower frequency. We track that transition using different filters of the envelope. There are several predefined filters, including Filter 4, which is optimized for earlier fault detection, and Filter 3, which is used to detect more mature faults.
Figure 1: Pure Digital Filters
Filter 4 detects energy from 5 to 40 kHz, characteristic of early-stage faults. If we see energy between 500 and 10,000 Hz, the failure has almost certainly reached Stage 3. If we see velocity, or damage with no filters, the failure has reached Stage 4.
Figure 2 shows the amplitude and signal detected by Filter 4 versus the stages of failure. A typical trend observed with Filter 4 is an increase in energy in the early stages of failure, followed by a leveling out and a decrease in later stages.
Filter 2: Filters versus stages of failure
As the bearing damage starts to drop in frequency, the higher signals can be observed in Stages 3 and 4 using Filter 3. We are thus able to determine the stage of failure based on the frequency of the bearing damage.
It normally takes a long time to progress between stages of failure. For example, a bearing in a mine hoist took a little over a year to progress from a Stage 1 to Stage 2 fault.
Once the velocity starts to react, the component has reached a very late stage of failure.
HD Order Tracking
HD Order Tracking is one of the more significant aspects of HD technology. The main purpose of order tracking is to keep the number of samples consistent per revolution. Consistency reduces smearing, thereby adding more clarity to the spectrums.
The bearing in Figure 3 is segmented into the number of samples per revolution. The distance between each sample should be identical, even if there is a time difference in between.
Figure 3: A visualization of number of samples per revolution
The HD Order Tracking algorithm adjusts the sampling rate in real time according to the rotational speed. It can handle up to a 50% fluctuation in speed change and still maintain
the same number of samples per revolution, providing crisp and clear results even with varying RPM.
To understand smearing, see the example in Figure 4 from a test rig with unbalance. The speed was deliberately changed during the sampling pe iod, creating the smearing effect.
Figure 4: A spectrum with HD Order Tracking turned off, showing smearing
It is difficult to see, but the overall amplitude is around 4.8 mm/s. But the spectrum energy only shows 0.55 mm/s because that energy has been smeared over several Hz due to the variable speed.
By enabling HD Order Tracking, we are able to maintain the same number of samples per revolution, reducing the smearing. The amplitude in the spectrum is now shown as 4.87 mm/s, matching the overall reading.
Figure 5: A spectrum with HD Order Tracking turned on
Figure 6 shows another example. There is a fluctuation of 19 RPM. When HD Order Tracking is turned off, the time signal decreases after a couple of turns because the RPM varied.
Figure 6: HD Order Tracking turned off
With HD Order Tracking is turned on, the time signal is maintained and provides crisp, clear data, even though there is a 62 RPM fluctuation rather than a 19 RPM fluctuation.
Figure 7: HD Order Tracking turned on
Symptom Enhancement
Symptom enhancement is a little-discussed algorithm that is specific to SPM technology. It searches for and identifies repetitive signals, and rejects “burst.”
Figure 8 shows two spectrums from a damaged bearing. The bottom spectrum is much more crisp and clear because symptom enhancement was turned on, thereby removing the noise.
Figure 8: Spectrums with symptom enhancement turned off and on, 990 RPM
This feature is equally helpful in clarifying readings from low-speed machines.
Figure 9: Spectrums with symptom enhancement turned off and on, 15 RPM
Which of these spectrums would you rather use when setting alarms or performing analysis?
HD Real Peak
HD Real Peak is a unit of measurement expressed in decibels. It is a true representation of the strongest peak.
In the typical envelope signal shown in Figure 10, there are four peaks of interest. We score the value based on the number of peaks for amplitude and envelope units to create a histogram. By collecting all of the peaks found during one measurement and extrapolating by using a profile of the histogram, we are able to calculate the HD Real Peak.
Figure 10: Time signal and histogram
This is all done in the time signal. HD Real Peak is the scalar value that we will use to determine the severity of damage and trigger alerts.
For anything that goes through the FFP process, the amplitude is considered not true.
