With the rise of technologies such as the Internet of Things, machine learning, etc., new concepts such as predictive maintenance and condition-based maintenance (PdM/CbM) have entered the field of vision. In simple terms, predictive maintenance is based on a large amount of real-time data collected from the production site, using a specially trained predictive model to predict the state of machinery and equipment, helping professional operation and maintenance personnel to identify existing problems and hidden dangers in time, and according to A proactive approach to maintenance is required. This undoubtedly helps to improve the reliability of the operation of machinery and equipment, reduce downtime and improve work efficiency.

Maintenance of machinery and equipment is an important issue in industrial production. Traditionally, it is up to the operator to rely on his own intuition and experience to determine whether the state of the machine needs maintenance, which is obviously not reliable; another method is to strictly follow the operation and maintenance manual to carry out regular maintenance and do preventive maintenance. Maintenance, this will do a lot of useless work, and it is still difficult to detect in time if there is an emergency; and if you wait until the machine has a problem, and then passively carry out corrective maintenance and repair, the loss caused by this It may be bigger… Is it possible to make more active, timely and accurate monitoring and evaluation of the state of the machine, so that the maintenance work can be done just right? This has always been the direction people are looking for.

With the rise of technologies such as the Internet of Things, machine learning, etc., new concepts such as predictive maintenance and condition-based maintenance (PdM/CbM) have entered the field of vision. In simple terms, predictive maintenance is based on a large amount of real-time data collected from the production site, using a specially trained predictive model to predict the state of machinery and equipment, helping professional operation and maintenance personnel to identify existing problems and hidden dangers in time, and according to A proactive approach to maintenance is required. This undoubtedly helps to improve the reliability of the operation of machinery and equipment, reduce downtime and improve work efficiency.

According to a research report by Markets and markets, the global machine condition monitoring market size is expected to grow from USD 2.6 billion in 2021 to USD 3.6 billion in 2026, with a compound annual growth rate of 7.1%. The interest in predictive maintenance in the industrial sector is evident here.

Accelerometers in Predictive Maintenance

To achieve predictive maintenance, an important foundational work is data collection, which is done by sensors that sense and collect valuable data including temperature, pressure, vibration, rotational speed, current, liquid properties, etc. , as the basis for prediction.

Among many data, vibration is the most critical one, because once a problem occurs in a rotating machine (such as ball bearing failure, shaft deviation, unbalance, excessive looseness, etc.), it will be reflected by the vibration spectrum, and it can be judged according to the characteristics of the spectrum. Where is the source of the problem. The sensing and acquisition of these vibration-related data requires acceleration sensors.

It’s Ta! An indispensable sensor “prophet” in industrial predictive maintenance
Figure 1: Spectrograms of mechanical shock with fracture (yellow) and without fracture (blue), the difference can be clearly observed (Image source: TE)

For effective and reliable vibration sensing, the traditional approach has been to mount an accelerometer on the machine and hardwire it back into a central machine protection system (such as a vibration monitor). This method, while reliable, is expensive, so it is usually only used on large, high-value rotating machines. However, with the implementation of Industry 4.0, the application of predictive maintenance has gradually penetrated into more small and medium-sized equipment and systems, which also promotes smaller size, lower comprehensive cost, greater ease of use and flexibility. Embedded accelerometer market.

To achieve long-term, reliable, stable and accurate sensing to provide accurate data for industrial condition monitoring and predictive maintenance, embedded accelerometers need to have the following characteristics:

wide frequency response

Since the vibration spectrum of different faults is widely distributed, in order to detect the faults of all mechanical equipment as much as possible, the accelerometer needs to have a wide frequency response range. Generally, for bearing monitoring, the frequency response of the accelerometer should be 40 to 50 times the shaft speed; for fans and gearboxes, the upper limit of the minimum frequency response of the accelerometer should be 4 to 5 times the pass frequency response of the blade.

Resolution and Dynamic Range

The measurement resolution of an accelerometer is a function of the ratio of the output signal amplitude to the broadband noise of the onboard electronics. The high resolution means that the accelerometer can measure vibrations of smaller amplitudes compared to sensors with lower dynamic range, and this ability to “capture” the lower amplitudes allows users to detect faults earlier. Noise affects measurement resolution, so in general, the output signal should be 10 times higher than the sensor noise level to ensure reliable measurement results.

long term stability

The performance of accelerometers drifts over time, and this change may trigger false alarms. Therefore, long-term reliability and stability of accelerometers are critical for continuously operating industrial equipment.

Ease of use

The ease of use is mainly reflected in two aspects: one is that the acceleration sensor needs to provide different packaging options according to different design needs, which is convenient for “embedding” into the final application scenario; the other is that the analog signal detected by the sensor is ultimately It needs to be converted into a digital signal and processed by the main control device, so providing digital output has become an important option to improve the usability of the product in addition to the traditional analog output.

How to choose between two embedded accelerometers?

There are two main types of embedded accelerometers currently used for predictive maintenance:

・One is a variable capacitance (VC) sensor based on MEMS technology. Its principle is to place a vibrating mass block between two parallel capacitive plates. When acceleration is sensed, the mass block will vibrate accordingly. cause a change in capacitance. This capacitance change is proportional to the applied acceleration. By converting the slight capacitance change into a voltage output, the corresponding vibration data can be measured.

・The other is a piezoelectric (PE) accelerometer, which contains a piezoelectric crystal that generates electrical signals when subjected to external mechanical vibration stress, thereby acquiring vibration data.

Both VC MEMS and PE accelerometer technologies are constantly evolving to accommodate the requirements of predictive maintenance applications. Compare the characteristic requirements for embedded accelerometers described above. We will find that, in comparison, the advantages of PE accelerometers in predictive maintenance are more obvious, which are mainly reflected in the following aspects-

First, most PE sensors are based on lead zirconate titanate ceramics (PZT), which are polarized to align dipoles and create a piezoelectric effect in the crystal. Because PZT crystals have a large frequency response range – which can support > 20KHz applications – they are ideal for condition monitoring applications. It can be seen from the figure that the VC MEMS sensor has better frequency response characteristics at the low frequency end, and there will be obvious deviation above 3kHz; while the performance of the PE sensor is always stable in the frequency range of 10kHz.

It’s Ta! An indispensable sensor “prophet” in industrial predictive maintenance
Figure 2: Frequency response characteristics of PE and VC MEMS accelerometers (Image source: TE)

Second, PE accelerometers have higher resolution and better dynamic range, allowing users to detect potential problems at an earlier stage. The figure below shows the results of measurements made on several PE and VC MEMS accelerometer products based on 0.03-10KHz bandwidth. It can be seen that the resolution of the PE sensor is about 9 times higher than that of the VC MEMS sensor.

It’s Ta! An indispensable sensor “prophet” in industrial predictive maintenance
Figure 3: PE has higher resolution and dynamic range than VC MEMS accelerometers (Image credit: TE)

Furthermore, in terms of long-term stability, PZT piezoelectric crystals have better stability, and PE accelerometers based on this material also show excellent stability over time. Although in principle, MEMS sensors using bulk silicon micromachining have the best long-term drift characteristics, their cost is relatively high, so they are usually used in inertial applications; for condition monitoring, surface micromachining is used. Processed VC MEMS sensors, which cost less, but compromise on resolution and long-term stability. Therefore, in a comprehensive comparison, PE accelerometers using PZT materials are the best choice for both performance and cost.

In addition, the PE sensing element itself does not require a power supply, and signal conditioning for charge-to-voltage conversion can be implemented in devices with extremely low current requirements; while VC MEMS sensors need to be continuously activated during use, power consumption is usually significantly higher than that of piezoelectric sensor.

To summarize, VC MEMS may be a good choice for vibration sensing applications that have lower frequency response bandwidth and are more cost-sensitive; while for frequency response range, resolution and dynamic range, long-term stability, power consumption, etc. For condition monitoring applications in the industrial field with higher requirements, PE accelerometers should be the preferred option.

It’s Ta! An indispensable sensor “prophet” in industrial predictive maintenance
Figure 4: Comparison of PE vs. VC MEMS accelerometer characteristics (Image source: TE)

TE’s Embedded Condition Monitoring Accelerometer Portfolio

In response to the needs of the industrial condition monitoring and predictive maintenance markets, TE Connectivity (hereafter referred to as TE) can provide a range of embedded PE accelerometers, including single-axis and triple-axis products, with different measurement ranges and packages, to Adapt to the design requirements of different end applications.

It’s Ta! An indispensable sensor “prophet” in industrial predictive maintenance
Figure 5: TE’s main embedded PE accelerometer (Image source: TE)

Among them, the 830M1 three-axis condition monitoring accelerometer is a low-cost, miniaturized, SMT-mounted accelerometer designed for embedded condition monitoring and preventive maintenance applications. It measures from ±25g to ±2000g, has a flat frequency response up to 15kHz in all three axes, and features low power consumption.

The 830M1 accelerometer is designed in shear mode with three independent and stable piezoelectric ceramic crystals. It is housed in a fully hermetically sealed leadless chip carrier (LCC) package that also includes a Ni1000 nickel thin film RTD temperature sensor for combined “vibration + temperature” sensing in a single package.

For cost-effective condition monitoring of those miniaturized devices, the 830M1 is perfect. If you need a single-axis version, you can also choose the 820M1 of the same series.

It’s Ta! An indispensable sensor “prophet” in industrial predictive maintenance
Figure 6: Block diagram of the 830M1 three-axis PE accelerometer (Source: TE)

If your design requires direct mounting of the PE sensor to the bearing housing, TE also offers the 805M1 single-axis 3-wire accelerometer with a stainless steel housing in a rugged TO-5 can package with a dynamic range of ± From 20g to ±500g, it can provide a flat frequency response of 12kHz, the sensitivity range is 4-100mV/g, and it supports adhesive or screw mounting. Since the 805M1 combines a stable piezoelectric ceramic crystal with a low-power processing circuit in a shielded enclosure, it is suitable for condition monitoring applications in many industrial sites.

It’s Ta! An indispensable sensor “prophet” in industrial predictive maintenance
Figure 7: 805M1 TO-5 single-axis 3-wire accelerometer (Image source: TE)

“Almighty” wireless accelerometer

Building on its high-performance, full line of PE condition monitoring accelerometers, TE has also taken a step forward by integrating the sensor, data collector, digital signal processor, and LoRaWAN™ radio into a compact, battery-operated, complete Wireless accelerometer system – this is the 8911 wireless accelerometer.

It’s Ta! An indispensable sensor “prophet” in industrial predictive maintenance
Figure 8: 8911 Wireless Accelerometer (Image Source: TE)

The 8911 wireless accelerometer is used to help customers quickly complete the proof of design concept and easily add remote wireless condition monitoring functions to machinery and equipment. The main advantages of this wireless accelerometer include:

Low-power wireless connectivity

The 8911 wireless accelerometer adopts LoRaWAN™, a low-power wide-area network protocol, which can realize wireless networking of battery-driven devices, and has the characteristics of long transmission distance and not easy to be interfered by external signals. Predictive maintenance is built into areas of the plant that are hard to reach with wired networks.

Enable edge computing

The 8911 wireless acceleration sensor has a built-in microprocessor that provides ambient temperature sensing and calculates FFT (Fast Fourier Transform) in the device to realize data processing at the edge.

Small form factor

Featuring a corrosion-resistant stainless steel housing and plastic cover design and compact form factor (1.25″ hex x 3.1″ height), the product’s mounting specs and stud sizes are standard so it can be quickly installed in existing and new mechanisms superior.

Reliable performance

On the one hand, the robust signal of the 8911 wireless accelerometer prevents interference and interruption; on the other hand, the rugged stainless steel sensor sheath and O-ring sealing ring make it IP66 rated, which ensures protection in harsh and harsh environments. operate effectively in the environment.

long working hours

The 8911 wireless accelerometer has extremely low sleep standby power consumption. If the sampling interval is set reasonably, the battery life of this sensor can be as long as 10 years. Due to the use of piezoelectric sensing elements, it has the advantages of high bandwidth, high measurement resolution and ultra-low power consumption compared to MEMS solutions, which is beneficial to improve the detection of machine faults while extending battery life.

It’s Ta! An indispensable sensor “prophet” in industrial predictive maintenance
Figure 8: Block diagram of the 8911 wireless accelerometer (Source: TE)

According to data, the industrial sensor market will grow from $18.2 billion in 2020 to $29 billion in 2025, with a compound annual growth rate of 9.8%. At the same time, industrial sensors are developing towards smarter, lower power consumption, smaller size, higher performance and easier to use. This trend is even more vividly reflected in accelerometers for condition monitoring and predictive maintenance.

In response to the development of the market, TE, relying on its technical advantages in the field of PE accelerometers, has extended a rich product series in breadth, and developed a complete wireless accelerometer solution such as 8911 in depth, which can be said to be the best solution for predictive maintenance applications. The landing provides a key ability of “foresight”, which is an indispensable existence in the future industrial field.

The Links:   LM80C312 SKK323-16E