Artificial Intelligence can help improve sealing performance

Trelleborg Sealing Solutions is making use of artificial intelligence (AI) and neural networks (NN) to boost the performance of its seals and prevent failures. Utilising the concept of ‘Cognitive Sealing’, the sealing specialist is offering a webinar introducing the topic and demonstrating its use in smart manufacturing, smart engineering, smart services and smart products. 

The webinar, entitled "Artificial intelligence – Applications in Daily Life, Industry & Sealing Technology", aims to explain a slection of applications in smart manufacturing. Data regarding various factors that influence product quality, such as material choice and the methods used for processing or post-processing, is acquired by sensors, processed by AI-assisted software and then brought together with relevant manufacturing data. The results can be compared with setpoints, after which a neural network identifies any regularities in these relationships which the system can then use to propose improvements to the manufacturing process. In quality control, advanced image recognition techniques mean artificial intelligence can replace manual testing even for complex product geometries.

Dr. Johannes Kunze von Bischhoffshausen of Trelleborg Sealing Solutions is one of the webinar's two presenters. He offers one example which explains how AI can simplify industrial processes: "Variations in material quality can have an impact on product function. Using artificial intelligence we can make use of the results from our raw materials testing to optimise processing and minimize this impact."

The webinar will also explain how AI systems are used at Trelleborg Sealing Solutions for managing the maintenance of complex sealing systems. An AI system can use data acquired from the sealing system to predict its impending failure in good time and so help to optimize the maintenance cycle. Using such systems is particularly beneficial in machinery which has expensive spare parts and whose failure results in high consequential costs.

Dr. Niklas Kühl, Head of the Applied AI Lab at Karlsruhe Institute of Technology (KIT) and the webinar's second presenter, explains, an AI system, which monitors the state of sealing systems, for example in hydraulics, has been jointly developed with KIT. "One particular feature of this application is that, for technical reasons, no sensors could be placed in the relevant seals. The AI system was thus fed with data from surrounding sensors and was able to use it to draw conclusions about the surroundings of the seal to be monitored and so display its state without being directly connected to it – a virtual sensor, so to speak. 

"The algorithm is capable of setting the various items of data received into an overall context. The combination of temperature, vibration, pressure and other indicators is interpreted by the system so as, ultimately, to obtain a picture of the state of the sealing system."

Trelleborg Sealing Solutions claims to be "a pioneering industrial user of artificial intelligence" and that it is constantly developing its skills in the field. At present, over 50 developers are working on numerous digitalisation and AI projects. The webinar is available to attend free of charge and takes place on  Tuesday October 13th at 11.00 CET.