National Control Devices Logo
Ultrasound Vibration Sensor for Preventive Maintenance

Ultrasound Sensors for Vibration Condition Monitoring

Predictive maintenance vibration sensors are a proactive approach to maintenance that aims to predict when equipment failure might occur and to prevent it before it does. Ultrasonic sensors are a key tool in predictive maintenance, as they can be used to detect early-stage failures for example in slow rotating machinery which may not be easily detectable with vibration analysis alone. This article will explore the use of ultrasonic sensors in predictive maintenance, including what they are, how they are used to detect equipment failure, their main features and use cases, their advantages over other vibration analysis techniques, and also what are the benefits in integrating them into a wireless IoT measurement network of sensors.

Introduction to Ultrasound Vibration Sensor

Ultrasound vibration sensors are devices that use high-frequency sound waves to detect the condition of mechanical assets, including bearings and other rotating equipment. They work by detecting the sound produced by the equipment, and analyzing it to determine its condition. This can be particularly useful for detecting early-stage bearing failures, as the sound produced by a failing bearing will be different from that produced by a healthy one. Ultrasonic sensors can be used in conjunction with vibration analysis to provide a more comprehensive understanding of the condition of mechanical assets, and can be particularly useful when monitoring slow speed bearings.

Ultrasonic sensors have a wide range of features and use cases. They can be used for a wide range of applications, including compressed air leak detection, condition-based monitoring of bearings, acoustic lubrication of bearings, and detecting arcing in electrical equipment. They can also be used for remote monitoring of equipment, allowing maintenance personnel to access real-time data and historical trends from anywhere. This capability enhances the early detection of bearing failure and other issues, leading to more proactive maintenance strategies. Ultrasonic sensors can be integrated with other IoT devices, such as vibration sensors, temperature sensors, and oil analysis sensors, to provide a comprehensive view of equipment health. This integration allows for a more holistic approach to predictive maintenance.

2 Channel Industrial IoT Long Range Wireless Ultrasound Vibration Sensor

$319.95 - $339.95

Video source: SDT Ultrasound Solutions

Principles of Operation

The idea is that certain sound frequencies over the 20kHz range (over the human hearing limit) are produced when rotating machinery starts experiencing issues. These soundwaves called ultrasounds have short wavelengths and low energy, which does not allow them to travel long distances (materials attenuate them heavily). On the other hand, this short wavelength results in a high frequency (they are inversely proportional), which makes them very directional and it is easy to detect and filter them out from other acoustic signals, allowing them to pinpoint their source, thus the source of the machine fault.

There Are Two Main Types of Ultrasound Sensors:

Airborne Ultrasound Sensors – these are utilized to detect leaks in pressure and vacuum systems, steam traps, valves, etc. These utilize a microphone.

Structural Born Ultrasound Sensors – early-stage bearing faults, lubrication issues, etc. These utilize either a piezo element or a MEMS sensor.

This article will focus on the latter and their use in early-earing fault detection in slowly rotating machinery. This choice is due to the fact that ultrasound sensors are a very go-no-go tool for predictive maintenance. They give a very early indication that a fault is impeding and can be the trigger for a more detailed vibrational analysis in the 0-10kHz range.

As can be seen in the Pi chart below the predominant fault in rotary machinery (contains a motor of some type) is due to issues with the bearing. Thus, we focus on this particular fault, complemented by lubrication issues (over or under) as one is related to the other.

Image source: BISinfotech

Methods for Analysis

There are two methods one can use:

  • Trending overall measurement – this is a static method and is relatively easy to implement
  • Waveform analysis – a dynamic method that can be performed either in time or in the frequency domain and is more complex and demanding (both on equipment and staff) than the static one.

Both methods have merits and we are going to look into each one in detail.

Overall Measurements

These are easy to obtain and are available no matter what particular sensor you pick. In most cases the following two measurements are considered:

  • RMS – root mean square value, represents energy in the measurement band
  • Peak – the highest registered value, the indication of very early-stage faults
  • Crest factor – the ratio of the Peak and RMS value, indicates the severity of the fault

In the case of structural ultrasound sensors these metrics are measured in dB𝜇𝑉, which means that the output is in decibel (nonlinear) concerning 1𝜇𝑉 (as we mentioned the vibrations are low energy, thus the magnitude of the registered Voltage is low)

A good sensor would measure all of the aforementioned, but more importantly, it does so with a good resolution and accuracy over a wide measurement range. For example, the NCD 2 Channel Industrial IoT Long Range Wireless Ultrasound Vibration Sensor is a high-quality sensor with industrial-level parameters that have the following values:

  • Range – 20 to 100 [dBµV]
  • Resolution – 5 [dB𝜇𝑉]

An important fact is that for this method it is vital to set a good baseline. You need to have a solid foundation to compare as the measurement is not absolute and only gives you an indication of to what degree the measured quantity has increased. Just as an example, you can consider the following as a rule of thumb:

Change over baseline

8 to 10dB          – Poor lubrication or early bearing fault

10 to 15dB        – Stage 2 bearing fault

15 to 30dB        – Late-stage bearing fault

> 30dB               – Bearing failure (catastrophic)

A thing to note is that these ranges are quite wide and depending on the calibration (if it has been done properly) and the established baseline, one might misinterpret these readings, thus the second form of analysis.

Waveform Analysis

Similarly, to vibrational analysis ultrasound measurements can be analyzed either as a function of time or frequency, creating two possible domains to examine the waveform in.

Time Waveform Analysis

This is the preferred method for ultrasound measurements (in contrast vibration analysis relies more on the frequency domain method). Early-stage defects exhibit visible changes in the time domain, where they become periodic in later fault stages.

What is characteristic of bearing faults in the early stage is that they exhibit very strong peaks, where the frequency of its repetition corresponds to the particular fault. The higher the amplitude the more severe the issue.

Frequency Vibrations Spectra Analysis

This type of ultrasound bearing analysis is not widely used and is limited to particular use-cases. For example, it can be used when the peaks are relatively small and hard to detect and the signal is periodic (this is a rare occurrence). Spectral analysis requires FFT (built into the sensor or post-processing) of the time domain signal, however, early-stage faults are not periodic with high peaks that produce a difficult-to-analyze waveform (in the frequency vibrations domain). Thus, not a method that is used a lot, best to stick to the time domain for bearing fault investigation.

Advantages of Ultrasound over Vibration Analysis

There is no shortage of industrial applications where low RPM machinery requires maintenance. Vibration monitoring systems for Pumps, fans, entire HVAC systems, etc. have system components that rotate at low speeds. These are more difficult to monitor than ones that operate at higher speeds as several factors influence measurements in low frequencies:

  • Amplitudes (g levels) of vibrations for low RPS machines are very small, which makes them hard to detect unless a very sensitive accelerometer is utilized, however, these are expensive (500mV/g or better).
  • Vibrations are in the 0 to 50 Hz range, which complicates detection as well and there is more noise in this band.

One could use a vibration sensor, however for the aforementioned reasons, it would have to be very sensitive and noise resistant, which would greatly increase its cost. To put it bluntly, the cost is simply not worth it when one could use a way simpler and cheaper ultrasound sensor.

In bearings ultrasound is produced by friction between the bearings and the raceway, the impacts produce acoustics in the ultrasound band (which has no direct relation to the RPM), thus it can be easily measured with an ultrasound sensor. As a rule of thumb bearing defects are in the 35 to 45 kHz band, thus it is suitable for sensors to work with a 37 kHz resonant frequency.

Enhancements Introduced by IoT

There are different methods to perform the actual measurements and gather the data. Traditionally ultrasound maintenance is performed hands-on, one uses handheld ultrasound sensors that convert the ultrasound to audible noise. This method while intuitive and uncomplicated is prone to subjective errors, requires experienced personnel and consumes a lot of time.

IoT is a big remote maintenance and management enabler and the Vibration based condition monitoring field is one that has seen a huge adoption. Utilizing smart sensors that require no human interaction once deployed can have a big impact on how efficient Condition Based Monitoring is.

What is an IoT-enabled ultrasound sensor, though? What makes it different to traditional solutions?

To answer these questions, we would have to take a look at what a sensor like the NCD one is comprised of. One could logically divide it into the following:

  • Sensor probe – The actual ultrasound sensor that converts the acoustics into electrical current/voltage
  • MCU – A microcontroller unit on a PCB that interfaces with the output of the sensor probe and converts the signal into bits of data. The brains of the device where processing is done and the logic behind the system is executed.
  • Communication module – Governed by the MCU this converts the bitstream to Radio Frequency (RF) wave to be transmitted over the air to a central hub for ingesting the data.
  • Power – An interface that powers the entirety of the system, this could be mains or more often than not in the case of IoT batteries.

Knowing what an IoT Ultrasound sensor is comprised of gives us sufficient insight into its operation to be able to discuss the main benefits it has over traditional ones.

Remote Operation

IoT enabled Ultrasound vibration sensors have the advantage of being able to wirelessly connect to a central hub to aggregate their data. This means you can perform measurements without a technician being on-site. Furthermore, numerous sensors can be deployed over several locations or even on the same machine and measured at the same time. This allows for data to be correlated and a better overall picture of the state of the machine/s to be obtained. Good IoT ultrasound sensors can also be configured remotely if the need arises to adjust the measurement parameters, for example, the reporting frequency.

Real-time Data

Deploy once, and measure many times. Once installed and connected to the wireless network these sensors can be configured to report their measurements over regular intervals. This gives you access to multiple data points at multiple locations at the same time. This type of real-time data can allow you to quickly detect any potential issues before they arise. This is the core of any predictive maintenance system that wants to be even remotely efficient.

Reduced Costs

This benefit is twofold. As the sensors themselves do not require human interaction there is no need for physical buttons, displays, etc. This reduces cost per unit or can result in a higher-grade sensor at the same cost. Additionally, it saves time, which is more important in many cases. You don’t want to waste your engineers/technician’s time, you want to send them on-site only and when there is a need to (to perform the actual maintenance not to check if there is a need for one).

Reduced Downtime

This is a direct consequence of the real-time data gathering. You can set alarms, look at trends, etc. in order to immediately respond to an issue. Long-term this ability to make decisions instantaneously and plan maintenance cycles leads to a big reduction of downtime of your machinery (stop production only when you need to and not before).

Conclusion

Predictive maintenance is becoming the norm for CBM and it is important to be done the right way. There are clear benefits to using ultrasound sensors, mainly the ability to detect faults early and not require expensive (sensitive) probes. The industry is no stranger to such methods, however, what comes as an improvement are the features IoT brings to the table.

The Industrial Internet of Things (IIoT) allows for data to be gathered more efficiently, in real-time and at multiple points at the same time. It does not have the limitations of on-hand measurements. Utilizing Wireless Ultrasound sensors allows for a lot better scaling where the data from thousands of measurement points can easily be gathered for post-processing and analysis.

At the same time, this allows for instantaneous decisions better maintenance scheduling and lower downtime.

Overall IIoT Ultrasound sensor networks bring greater benefits at a lower cost at scale. They are a true improvement to CBM and one of the biggest enablers of Predictive maintenance.

References

  1. Case Study: Detecting Low- or Slow-Speed Bearing Failure with Ultrasonic Tools

      https://www.machinerylubrication.com/Read/31954/detecting-low-or-slow-speed-bearing-failure-ultrasonics

  1. Ultrasound Monitoring for Predictive Maintenance

     https://www.sensegrow.com/blog/iot-solutions/ultrasound-monitoring-for-predictive-maintenance

  1. Choosing the Most Suitable Predictive Maintenance Sensor

     https://www.analog.com/en/technical-articles/choosing-the-most-suitable-predictive-maintenance-sensor.html

  1. SDT Ultrasound solutions

     https://sdtultrasound.com/using-ultrasound/

  1. Sensegrow

     https://www.sensegrow.com/blog/iot-solutions/condition-based-monitoring-predictive-maintenance-ultra-low-rpm-machines

  1. BISinfotech

     https://www.bisinfotech.com/choosing-the-most-suitable-predictive-maintenance-sensor/

Share this on:
Facebook
Twitter
LinkedIn
Pinterest
Reddit
WhatsApp
Email
YOU MIGHT ALSO LIKE...​
Industrial IoT Wireless Transmitter

IntroductionOverview of NCD Long Range Wireless Sensor TransmittersPresenting a Comprehensive List of Essential Features and Compatible Probes.List of the Supported Probes by NCD Industrial IoT Transmitter Device3-Axis Vibration And Temperature ProbeAC Current Sensor ProbeHigh-Temperature Thermocouple ProbeTemperature / Humidity ProbeAir Quality Sensor ProbeDoor Open/Close ProbeWater Detection ProbePressure Sensor ProbeUltrasound Tank Level ProbeUltrasonic Vibration Sensor ProbeOil Quality Analysis Sensor ProbeCO2 Temperature Humidity ProbeLinear Displacement Sensor LVDTWire Draw ProbeDust Sensor with Temperature Humidity

Read More »
Machine Up Time Monitoring Product Manual

IntroductionSpecificationsApplicationsHardwareSensor HardwareA. Main Electronics BoxB. The ProbeMechanical DrawingInternal ComponentsSensor DiagramDigital InputsAccelerometer Sensor Magnetometer Sensor Current Sensor Sensor PowerGatewaySensor Radio TechnologySensor Internal OverviewSoftware API StructureMachine Uptime Monitoring Sensor Payload StructureSensor Configuration StructureEnter Configuration Mode Read Sleep Duration Set Sensor Node ID and Sleep DurationRead Sensor Network IDSet Wireless Sensor Network IDRead Sensor Destination AddressSet Sensor Destination AddressSet Sensor Destination to Broadcast Read Wireless Sensor Transmission Power LevelRead Wireless Sensor RetriesSet Wireless

Read More »
Ultrasound Vibration Sensor for Preventive Maintenance
Ultrasound Sensors for Vibration Condition Monitoring

Predictive maintenance vibration sensors are a proactive approach to maintenance that aims to predict when equipment failure might occur and to prevent it before it does. Ultrasonic sensors are a key tool in predictive maintenance, as they can be used to detect early-stage failures for example in slow rotating machinery which may not be easily detectable with vibration analysis alone. This article will explore the use of ultrasonic sensors in

Read More »
Load More