Advantech, Actility unveil AI-based predictive maintenance IoT Solution for rotating machinery

Maintenance companies have been facing many difficulties in fulfilling their duties lately due to the pandemic or conflicts and wars. Traveling to customer facilities can be a challenge, and many service providers cannot offer their services or products as usual. This situation has resulted in an exponential slowdown across the industry with a high downtime cost.

“Implementing efficient maintenance processes is crucial to operators of industrial facilities, but it’s not an easy task,” stated Actility in a statement. “In traditional methods, still used by most factory operators, the on-site inspection is done manually, regularly but without stable inspection quality, and only passively reacting and informing the maintenance crew when the problem occurs.”

In response, Actility and Advantech have developed their Predictive Maintenance Solution for Rotating Machinery, which uses LoRaWAN-connected vibration sensors to monitor machine health. This solution helps industrial customers reduce unexpected shutdowns and accidents and improve performance and safety by detecting abnormal behavior in the data from the sensors measuring vibrations frequencies. Easy implementation of the infrastructure makes the solution accessible and reliable, the companies say.

According to the companies, one of its most notable features is that it acts proactively based on the future state of the machine. Instead of carrying out routine inspections, it prevents machine faults in advance.

“Using this solution means joining the Industrial Revolution 4.0, and therefore ensuring longer lifespan of equipment, lower spare parts inventory, and having more cost-effective than run-to-failure or planned maintenance,” explained Actility.

The companies say that the Predictive Maintenance Solution is highly versatile and manufacturers can use it in many situations. Examples of its application include petrochemical plants, PCB plants and semiconductor manufacturing plants. One particularly successful case is the use of the solution for managing rotating machinery in steel plants.

“In this specific industry, the failure of the manufacturing machine will seriously affect the product line, that is the reason why it is extremely important to prevent unexpected breakdowns,” stated Actility.

Concerning how the technology works, Advantech’s Predictive Maintenance Solution uses a pre-trained model to analyze and process data. This model has been trained on a large benchmark dataset to address customer problems. It automatically performs preprocessing, feature value selection and deep learning model processes.

Meanwhile, the degrading level indicates the status of the machine’s rotation compared to when it’s functioning optimally. To ensure adequate maintenance strategies, operators typically set a threshold index according to expert advice.

Additionally, ISO 10816 outlines general guidelines for evaluating machine vibration through measurements taken from non-rotating parts. It provides special procedures and directions for assessing the mechanical vibration of reciprocating compressors. Utilizing multi-indexes can help keep track of machinery and production line operations.

The Trial Kit contains the Advantech WISE-2410 and WISE-6610 devices for reliable and secure industrial edge intelligence. The WISE-2410 device is certified to ISO 10816 standards for measuring machine vibration on non-rotating parts. It balances wireless bandwidth between it and the gateway, helping to reduce data transmission fail rates.

The WISE-6610 is a high-performance LoRaWAN gateway that supports the building of private and public networks. It also supports VPN tunneling and various protocols, such as MQTT.

Advantech and Actility also conducted a Joint Webinar called “Intelligent Motor Predictive Health Monitoring & Management” where experts from both companies explained the solution in detail.

Article Topics

Actility  |  Advantech  |  AI/ML  |  Industry 4.0  |  IoT  |  predictive maintenance  |  sensor