KIStE: AI failure analysis of technical elastomers

Example of fatigue cracks

Technical elastomers in the form of seals or vibration control applications are indispensable today. These high-tech products are used in cars, aircraft or hydraulic and pneumatic systems, and are exposed to the highest requirements in terms of temperature, media and mechanical loads. In the event of damage, it generally means entire machines or systems come to a standstill. Economic losses due to production downtime or the endangerment of users’ lives or limbs are possible consequences. AI failure analyses from Fraunhofer LBF can prevent these scenarios.

If damage has occurred, a failure analysis is required to remedy it and prevent future damage. Damage to elastomer products can have many causes, including aging, manufacturing defects, and mechanical, thermal or climatic stresses. Among others, one popular method for analysis can be seen in VDI 3822. However, performing an extensive analysis using this method requires experience and expertise. The situation is further complicated by the fact that similar damage patterns can result from different causes or complex stresses.

Wouldn’t it be simple to have artificial intelligence perform this analysis?

Teams of experts from Fraunhofer LBF are developing a model trained with machine learning that automatically performs the failure analysis based on imaging, physical and chemical methods as well as other provided information, and independently evaluates the causes of the damage. In order to successfully train the AI, a very large amount of data is required. Together, we look at your existing damage and evaluate it according to our taxonomy.

With you as a partner, we can make failure analysis more efficient!