“Factory automation and overall efficiency have naturally received great attention, not only because productivity increases (even if a little bit) can bring positive benefits, but equally importantly, it can reduce or eliminate serious losses caused by equipment downtime. Now, instead of relying on advances in analysis technology to gain insight into available statistics to predict maintenance needs, or simply relying on strengthening the training of technicians, we can achieve true real-time analysis and control through advances in detection and wireless transmission technologies.

“

Factory automation and overall efficiency have naturally received great attention, not only because productivity increases (even if a little bit) can bring positive benefits, but equally importantly, it can reduce or eliminate serious losses caused by equipment downtime. Now, instead of relying on advances in analysis technology to gain insight into available statistics to predict maintenance needs, or simply relying on strengthening the training of technicians, we can achieve true real-time analysis and control through advances in detection and wireless transmission technologies.

Precision industrial production processes increasingly rely on the efficient, reliable and consistent operation of motors and related machinery and equipment. Unbalances, defects, loose fasteners and other abnormal phenomena of machinery and equipment often turn into vibrations, resulting in decreased accuracy and safety issues. If ignored, in addition to performance and safety issues, if the equipment is shut down for repairs, it will inevitably lead to loss of productivity. Even small changes in equipment performance are often difficult to predict in time and quickly translate into significant productivity losses.

As we all know, process monitoring and condition-based predictive maintenance are effective methods to avoid productivity loss, but the complexity of this method is comparable to its value. Existing methods have limitations, especially when it comes to analyzing vibration data (regardless of how it is obtained) and determining the source of error.

Typical data acquisition methods include simple piezoelectric sensors installed on the machine and handheld data acquisition tools. These methods have many limitations, especially when compared with the ideal comprehensive detection and analysis system solution, which can be embedded on or in the machine and can work autonomously. The following is an in-depth discussion of these limitations and their comparison with the ideal solution-autonomous wireless embedded sensors. The analysis of options for complex system targets with fully embedded autonomous detection elements can be divided into ten different aspects, including achieving high repeatability measurement, accurate evaluation of collected data, appropriate documentation and traceability, etc. The following will Explain all aspects and discuss available and ideal methods.

Accurate and repeatable measurement

The existing handheld vibration probe has some advantages in the implementation method, including no need to make any modification to the terminal device, and its integration is relatively high, the size is large, and it can provide sufficient processing capacity and storage space. However, one of its main limitations is that the measurement results are not repeatable. A slight change in the position or angle of the probe will result in an inconsistent vibration profile, making it difficult to make accurate time comparisons. Therefore, maintenance technicians first need to figure out whether the observed vibration shift is caused by actual changes inside the machine or simply because of changes in measurement technology. Ideally, the sensor should be compact and fully integrated, and be able to be directly and permanently embedded in the target device, thereby eliminating the problem of measurement position offset, and allowing complete flexibility in scheduling the measurement time.

Frequency and timing of measurements

In high-value equipment production facilities, such as when manufacturing sensitive Electronic devices, process monitoring is extremely beneficial. In this case, a small deviation of the assembly line may not only cause a decrease in factory productivity, but may also shift the key specifications of the final equipment. Another obvious limitation of the handheld probe method is the inability to point out problematic vibration excursions in real time. The same is true for most piezoelectric sensors, whose integration level is generally very low (in some cases there is only one sensor), and data needs to be transmitted to other places for analysis. These devices require external intervention, so some events and vibration excursions may be missed. This is not the case with the autonomous sensor processing system. It has built-in sensor, analysis, storage and alarm functions, while still small enough to be embedded in the device, able to notify vibration excursions at the first time, and best Display time-based status trends.

Understand the data

The above-mentioned concept of embedded sensors sending out real-time notifications can only be realized by frequency domain analysis. Generally, any equipment has a variety of vibration sources, such as bearing defects, unbalance and gear meshing, etc., in addition to vibration sources caused by design, such as vibration generated by a drilling machine or a press during normal operation. Time-based analysis produces a complex waveform that integrates all these vibration sources, and the information it provides is difficult to distinguish before FFT analysis. Most piezoelectric sensor solutions rely on external FFT calculation and analysis. This not only makes real-time notification impossible, but also pushes most of the extra design work to device developers. However, if the FFT analysis function is embedded in the sensor, the specific source of the vibration offset can be determined in real time. Such a fully integrated sensor element can also shorten the development time of equipment developers by 6 to 12 months, because it is fully functional, simple and effective, and works autonomously.

Data access and transmission

Embedded detection can perfectly provide accurate and real-time trend data, but it will not increase the complexity of data transmission to remote process controllers or operators. The premise of embedded FFT analysis is obviously that the analog sensor data has been conditioned and converted into digital data in order to simplify data transmission. In fact, most vibration sensor solutions currently in use only provide analog output, resulting in a reduction in signal quality during transmission, not to mention the complexity of offline data analysis (discussed above). Considering that most industrial equipment that requires vibration monitoring often exists in high-noise, moving, inaccessible, or even dangerous environments, the industry is eager to reduce the complexity of interface cables and also perform as much data analysis as possible at the source. Work in order to capture the most accurate information about the vibration state of the equipment. Sensor nodes with wireless transmission capabilities not only facilitate immediate access, but also greatly simplify the deployment of sensor networks and significantly reduce costs.

Data directionality

Many existing sensor solutions are single-axis piezoelectric sensors. These sensors do not provide directional information, thus limiting our understanding of the device’s vibration profile. The lack of directivity leads to the need for very low-noise sensors in order to provide the required resolution, which in turn affects cost. Multi-axis MEMS sensors are different. If each axis is precisely aligned, the ability to determine the source of vibration will be greatly improved, and it will also help reduce costs.

Location and distribution of sensors

The vibration profile of the equipment is very complex, changes over time, and also varies with equipment materials and locations. It is of course very important to determine where to place the sensor. The main determinants are the type of equipment, the environment, and the life cycle of the equipment. When using existing high-cost sensor elements, the detection point is limited to only a few or one, and this problem becomes even more important. This will lead to a significant increase in the pre-development time, because repeated experiments are required to determine the best location. But in most cases, the consequence is that the amount of data collected and the quality of the data will be affected. Fortunately, sensor probes with higher integration and greatly reduced costs are now available. Multiple probes can be placed in each system, thus shortening the initial development time and cost, or using fewer, lower-cost probes. fulfil requirements.

Adapt to life cycle changes

The handheld monitoring system method can be adjusted according to time changes (period, data volume, etc.), and to provide the same life cycle-based adjustment in the embedded sensor, attention must be paid in the early design and deployment stages to achieve the required Adjustable function. No matter what technology is used, the sensor element is very important, but more important is the signal conditioning and processing circuit around the sensor. Signal/sensor conditioning and processing not only depends on the specific equipment, but also on the life cycle of the equipment. This involves many important considerations in sensor design. First of all, the analog-to-digital conversion process is best performed as early as possible (in the sensor head, not outside the device) in order to support in-system configuration and adjustment. The ideal sensor should provide a simple programmable interface to simplify equipment setup, filtering operations, alarm programming, and testing of different sensor positions through fast baseline data collection. For the existing simple sensors, even if they can be configured when the device is set up, some sacrifices must still be made in the sensor settings to adapt to changes in the maintenance focus of the device throughout its life. For example, should the sensor be configured for the early stage when the equipment is less likely to fail, or for the late stage where the failure is more likely and more harmful, it is best to use a sensor that can be programmed in the system to follow the life cycle Adjust the configuration for changes. For example, early monitoring is relatively sparse and has the lowest power consumption; after observing the change (warning threshold), reconfigure to frequent monitoring mode (the monitoring period is set by the user); in addition to continuous monitoring, it also provides interrupts based on the alarm threshold set by the user Driven notifications.

The adaptation of sensors to changes in the equipment life cycle depends to a certain extent on the understanding of the baseline equipment response. A simple analog sensor can be used to obtain the baseline device response, that is, allowing the operator to take measurements, perform offline analysis, and store this data offline together with appropriate markers on specific equipment and probe locations. A better and less error-prone method is to store the baseline FFT in the sensor head so that the data will never be misplaced. Baseline data can also help determine the alarm level. The value is preferably programmed directly on the sensor so that in the subsequent data analysis and collection, if a warning or fault condition is detected, a real-time interrupt can be generated.

Data traceability and documentation

In a factory environment, a suitable vibration analysis program may monitor dozens or even hundreds of locations, whether through handheld probes or embedded sensors. In the entire life cycle of a device, thousands of records may need to be obtained. The integrity of the predictive maintenance program depends on the proper mapping of the location and time of sensor collection points. In order to minimize the risk and obtain the most valuable data, the sensor should have a unique serial number and embedded memory, and be able to add a time stamp to the data.

reliability

The above focuses on the improvement of existing sensor vibration monitoring methods for process control and predictive maintenance. In terms of fault tolerance and monitoring, the sensor itself should also be carefully reviewed. What if the sensor fails (change in performance) instead of the device? Or, if a sensor that works completely autonomously (the ideal method described above), how confident can we be that the sensor will continue to work normally? For many existing sensors, such as piezoelectric sensors, this situation does cause serious limitations because they cannot provide any in-system self-test functions. Over time, there is bound to be a lack of confidence in the consistency of the recorded data. In the critical monitoring stage at the end of the equipment life, real-time fault notification is of great significance in terms of time, cost, and safety, and whether the sensor is still working properly must be the focus of attention. The basic requirement of a high-confidence process control program is the ability to remotely self-test the sensor. Fortunately, some MEMS-based sensors can do this. The embedded digital self-test capability thus fills the last gap in a reliable vibration monitoring system.