baumhueter extrusion GmbH is a family-owned company from Rheda-Wiedenbrück that specialises in the production of strapping, construction fibres and technical fibres. With its specialisation in the extrusion of plastics, the company is one of the pioneers of this process technology in Germany.

The construction fibres from baumhueter, known as PB EUROFIBER, are specially designed for use in building materials and optimise the properties of concrete, mortar, adhesives and so-called DryMix products. The short cuts of the technical fibres are specially designed for technical applications and can be individually adapted. An experienced team of chemists, plastics engineers and machine operators regularly carry out manual dispersion tests for quality assurance purposes. The fibre cuts are evaluated in terms of their homogeneity, optimum mixing with the matrix materials and the expected performance.

Digitalisation and automated evaluation in quality assurance

The evaluation of these dispersion tests currently still consists of a manual visual test, relies heavily on the experience of the employees and is therefore also prone to errors.

Together with the Institute for Industrial IT (InIT), baumhueter extrusion GmbH would like to introduce machine learning processes such as image processing methods in order to support employees in quality assurance with automated test results.

To this end, the manually evaluated tests are first digitised and compiled in a database. To do this, the test results from the past are scanned, photographed and saved with relevant meta information. This data is used to train intelligent image processing models that will automatically analyse dispersion tests in the future. The test results are then validated and optimised so that the AI models can ultimately be introduced at baumhueter.

For baumhueter, a transfer project like this is a good opportunity to introduce data-driven methods into the company. In this case, we would like to relieve our experts in quality control through AI support.

– Dr. Stefan Barwich, baumhueter extrusion GmbH

Relieving staff and increasing efficiency with the help of AI

The project will result in a variety of added values for the company: the digitalisation of the tests will soon enable faster access to historical test results and greater transparency of the test data. At the same time, baumhueter can speed up the evaluation of new test results and improve production control.

In addition, the introduction of AI methods, especially automated image processing, should support and relieve employees, stabilise work processes and thus ensure even utilisation of the machines. This is particularly helpful for employees working night and late shifts. For baumhueter, the project also opens up opportunities to introduce further data-driven methods in the company in the medium term.