Vibration Monitoring (VM) has been around for quite a long time and has been used to monitor the health of a machine, equipment, or a structure. The vibration data collected via dedicated sensors during the operation of a machine is monitored and analyzed in real-time.
The main goal of vibration monitoring is to reduce the risk of fatal damages and potential line-down situations which leads to final operational cost control and reduction.
The vibration data from a vibration sensor can be used as a stand-alone input or be combined with other sensor data depending on the operational requirements. For example, in a factory automation application, the vibration data can be combined with:
This combination generates a complete system that will provide a more robust and reliable solution.
In some other use cases such as structural monitoring, the vibration data can be combined with the tilt position data that is collected via an inclinometer to determine the health of the structure.
The collected data are fed into dedicated algorithms, including the emerging artificial intelligence (AI) algorithms, to develop a model that can predict potential future failure. The model prediction information can then be used to build knowledge for making decisions on whether any immediate actions need to be taken to avoid productivity loss.
A new trend in factory automation is the emergence of AI algorithms which can be trained based on sensor data to predict which tasks should be performed. This lessens the burden on individual operators who previously had to make critically difficult and time-consuming decisions. An autonomously automated factory takes away the responsibility of individual operators and reacts automatically to any changing operating conditions.
A key component in a vibration monitoring application is a vibration sensor. The latest vibration sensors are based on MEMS technology using the same concept of acceleration detection in an accelerometer. The main difference is in the bandwidth of the sensor. A MEMS accelerometer has a typical bandwidth of 3 kHz, however, a vibration sensor is capable of detecting the vibration at significantly higher bandwidth. The capability of a vibration sensor to capture high-frequency signals enables more accurate frequency analysis of the vibration. The latest MEMS vibration sensor offers a bandwidth of over 6 kHz which will be discussed later.
A MEMS-based vibration sensor has many use cases and Figure 1 provides a list of some major applications. Motor vibration monitoring is an essential building block of successful factory automation. Vibration monitoring in railways can help avoid catastrophic train accidents. Home appliances such as washing machines have been equipped with vibration monitoring since the inception of MEMS sensors in industrial applications. The structural monitoring application has gained momentum since the emergence of MEMS sensors at an affordable cost. For example, municipalities bear the responsibility to monitor bridge vibration to ensure that structures are in good health and sound condition. The bridge vibration data, particularly during peak traffic hours, can provide valuable information on any abnormality that may cause the collapse of the bridge.
Figure 1: Some MEMS sensor vibration sensor applications. (Image source: STMicroelectronics)
The technical specifications of a vibration sensor need to be carefully analyzed to ensure that the sensor can meet the requirements of the target application. Table 1 depicts the major parameters of one of the latest vibration sensors offered by STMicroelectronics. This device can capture the vibration in the 3-dimensional space (x, y, z). The three degrees of freedom offered by this device provide the flexibility to position the device in mounting orientation.
The full-scale range of up to 16 g of acceleration per axis is sufficient to cover the vibration amplitude range that is typically required to monitor the health of a machine.
This device offers an ultra-wide bandwidth, flat frequency response up to 6.3 kHz, and embedded filtering which eliminates frequency aliasing.
Another major characteristic of this device is the very low spectral noise density. This is a very important advantage when low-frequency vibration needs to be captured.
Compared to the existing vibration sensor, the operating temperature range is extended to +105°C to meet the requirement of a demanding operating environment.
The device can be operated either in a 3-axis mode or a single-axis mode which can be selected through dedicated registers. In the 3-axis mode, all three axes (x, y, z) are simultaneously active. In the single-axis mode only one axis is active. In single-axis mode, the resolution (noise density) of the active axis significantly improves.
Table 1: The major parameters of the latest vibration sensors offered by STMicroelectronics.
Vibration Monitoring Applications
Vibration monitoring usually refers to the analysis of the vibration of a machine, equipment, or an appliance as part of a comprehensive application that is known as Condition Monitoring (CM) or Condition-based Monitoring (CbM). The vibration analysis plays a significant role in monitoring the health of the machine over time. However, in addition to the vibration data collection, a complete condition monitoring solution incorporates multiple sensors to collect vital equipment parameters including temperature, noise, pressure, smoke, and humidity. Each of these sensors provides valuable information on a specific condition of the machine. These sensor data are fused, processed, and analyzed to build knowledge of the overall condition of the machine to make critical decisions on machine maintenance.
Figure 2 illustrates some of the major applications of vibration monitoring in various markets. The breakdown in this figure highlights the importance of vibration data collection and analysis as part of a comprehensive solution for CM. Additional sensors can be used to collect data that will be fused together for a reliable and effective outcome. In the latest solutions offered in the industry, intelligent algorithms using sensor data bring the capabilities and effectiveness of such solutions to a new level. These innovative and powerful solutions can help significantly reduce the cost and inefficiencies associated with equipment line-down situations that would otherwise be inevitable.
Figure 2: Various applications of Vibration Monitoring. (Image source: STMicroelectronics)
Cloud computing has become one of the critical parts of an extensive solution involving sensor data collected from multiple locations of an enterprise to ensure that there is no interruption at any level at any location. The central processing unit in the cloud is used to combine and analyze all the data and monitor the involved machines and equipment in real-time to ensure a smooth and uninterrupted operation.
Figure 3 provides a list of the essential building blocks of a vibration monitoring system. Depending on the needs and requirements of the system, a variety of sensors can be mounted on the equipment that must be monitored. The list of sensors includes:
- Inertial sensor module
- Ambient light sensor
A processing unit is required to analyze the collected data. Depending on the amount of the data, privacy, data security, latency, and power requirements, the analyses may be performed at the local processing unit or transmitted to a cloud processing center where all the data from multiple pieces of equipment is collected and analyzed.
Figure 3: Building blocks of a vibration monitoring system. (Image source: STMicroelectronics)
At some point after the installation and during the operation of the machine, the condition of the machine starts to change. It is critical to have all the required sensors installed to collect data on ultrasound and audible noise, vibration, power consumption, temperature, and any potential smoke. As time passes, the necessity of collecting machine parameters and sensor data becomes critical to monitor the health of the machine.
Figure 4 depicts the typical Installation and Point of Failure (IPF) curve of a machine that is being monitored. The time from the machine condition change to the final failure may take months or even years before it starts to show symptoms of failure. Early analysis of the sensor data can give an indication of the machine health and trained AI algorithms using sensor data as input can predict a failure and initiate the process of taking the necessary actions.
Figure 4: IPF-Curve. (Image source: STMicroelectronics)
Figure 5 provides an example for vibration monitoring of an electric pump. Different conditions, such as imbalance, looseness, output shaft, and the gearbox of the pump can be monitored using a vibration sensor. The vibration sensor data is then transmitted for further extensive analysis including a Fast Fourier Transfer (FFT) of the vibration data that can determine the individual frequency signature of these conditions.
Figure 5: Vibration monitoring of an electric pump in various conditions. (Image source: STMicroelectronics)
A condition monitoring system for an electric motor can have several components in addition to the electric motor. The solution can have multiple sensors including the ones for vibration, temperature, pressure and other sensors depending on the requirements of the operating environment. The connectivity option between the pump and the processing unit can be a wired or wireless one with dedicated communication protocols. The processing and analysis unit can provide pump diagnostics and visualization tools to help the operator proactively identify and tackle issues such as pump irregularities that could result in operational downtime and disruptions. This proactive engagement can increase a company’s profit by lowering operating and maintenance costs for the factory.
Many sensors are being deployed to implement a comprehensive solution for predictive maintenance. The latest MEMS-based vibration sensors have enabled efficient and cost-effective vibration monitoring solutions in factory automation, power utilities, home appliances, and structural health surveillance and supervision. Vibration monitoring can be deployed as a stand-alone solution or as part of condition-based monitoring which has been emerged as an integrated part of a comprehensive solution to monitor various machines by collecting and analyzing the data in real-time. This solution has empowered the factories of the 21st century to proactively monitor and address issues arising from machine productivity disruptions and line-downs. Vibration monitoring is a critical building block of a comprehensive solution in any factory automation.
- Ultra-wide bandwidth, low-noise, 3-axis digital vibration sensor: https://www.st.com/en/mems-and-sensors/iis3dwb.html
- Analog bottom port microphone with frequency response up to 80 kHz for Ultrasound analysis and Predictive Maintenance applications. https://www.st.com/en/mems-and-sensors/imp23absu.html
- Low-voltage, ultra-low-power, 0.5 °C accuracy I²C/SMBus 3.0 temperature sensor. https://www.st.com/en/mems-and-sensors/stts22h.html