Some weeks ago we launched, Aion, our enhanced Machine Diagnostic tool (you can learn more about it here). Two of its strength points are the ease of application across many different machine types and its ability to conduct pre-processing of the data it collects on the edge.
Mr Horst Benderoth, for more than 40 years expert consultant for energy-efficient fan systems and innovation in fans construction, wrote an article on Aion and how its technology can analyze data and prevent failures in industrial fans.
Here is the translation of the original article (which can be found here).
Two important factors to take into consideration when designing a digitization strategy are that all processes can be digitized and that it can sometimes be easier than imagined to do so. AiSight, a young, Berlin-based startup that operates internationally, has continued to grow based on these principles. The latest version of their solution, Aion, which launched in September 2020, leverages artificial intelligence algorithms to predict errors, dynamically regulate the machine’s parameters based on sensor data, and determine the state of a machine in real-time. What differentiates Aion from other solutions on the market is the ease of its implementation and scalability, which makes these very advanced analytics easily accessible for SMEs.
Aion is a hardware and software solution that consists of a small, sensor-filled device that is optimized for plug and play installation without requiring any specialized knowledge. Aion’s software uses machine learning models to evaluate the condition of a machine, identify anomalies and determine causes of errors based on patterns and physical parameters in the sensor data. The key to this solution is "acoustic fingerprinting," in which physical parameters are extracted from the sensor data in order to identify patterns.
Detecting and identifying faults in industrial fans is also among Aion’s many applications. Aion’s thorough understanding of what causes industrial fan failures allows it to easily detect abnormalities based on a machine’s vibrations and link them to specific faults. The main failures that industrial fans typically face are bearing faults. The bearing is one of the most important components of a fan, so predicting its malfunction can be extremely important. The main causes of bearing faults are lubrication—the lubricant could be insufficient, incorrect or degraded due to high temperatures—and contamination, which occurs when extraneous materials and particles come into contact with the bearing. It is possible to spot a bearing fault by observing vibrations (Figure 1).
Avoiding bearing faults is also possible by monitoring key events that might increase the chances of bearing faults to manifest. These include, but are not limited to, the following examples:
Imbalances: These faults generally occur when material accumulates on the fan blades, when the blades chip off, or when the high temperatures reached by the fan during use cause the uneven expansion of some of the fan’s parts. Such imbalances exponentially increase the likelihood of a bearing fault. (Figure 2)
Misalignment: This fault is mainly related to the incorrect installation of new equipment or to improperly seated shafts and bearings (Figure 3).
In addition to detecting failures and/or key events that might lead to them, vibration monitoring can optimize current maintenance operations. There are two examples of faults that can result from improper maintenance operations.
First, misalignments often occur after maintenance sessions in which the bearing on a shaft has been changed. By monitoring industrial machines constantly, it is also possible to spot such mistakes right away. Second, in industrial fans, proper cleaning of the blades is essential. Carrying out this operation too often or too late can lead to a decrease in efficiency. AiSight is able to detect this kind of fault and facilitate the scheduling of maintenance operations that can help your industrial fans run at peak performance.
The way AiSight alerts operators of all these faults is very straightforward. Aion can be plugged into any industrial machine and, once plugged in, its sensors will monitor many parameters, such as temperature, magnetic field and, most importantly, vibration. Aion’s triaxial analysis allows high-resolution monitoring of any type of equipment. After sending Aion online, its firmware will conduct a Fast Fourier Transform (FFT) analysis and use a number of other signal processing techniques in real-time. FFT analyses can acquire time domain-data and convert it to frequency-domain, allowing operators to determine whether a machine’s vibrations are normal (as shown in the graphs above). Analysis of the frequency graphs facilitates detection of the root-cause behind each fault. The Machine Learning algorithms carrying out the analysis are already trained to interpret this data and channel it to the powerful microprocessors embedded in the sensor. This is possible through a combination of the latest in microprocessor technology and the optimization of algorithms to run in extremely light-weight environments. The edge-computing architecture ensures that only relevant data is sent to the cloud, significantly reducing data traffic. All this data is displayed in real-time on the dashboard and whenever a fault is detected, alerts are sent through the dashboard (Picture 2), as well as email and SMS, facilitating timely intervention.
Finally, in the case of rotating equipment, this analysis can lead to an average uptime increase of 6%, positively impacting the equipment’s OEE. On average, Aion improves maintenance productivity by 45% and, most importantly, will be scheduled and carried out according to equipment’s needs, thus eliminating the stress caused by unplanned interventions.
*Images were provided by AiSight based on real data, demonstrating how faults are recognized by our algorithms
If you're interested in seeing how the product works, schedule a demo by hitting the button below.Schedule a demo