There were two exhausting but very interesting exhibition days at the Days of Digital Technologies 2024 in Berlin, where we looked at a project on sustainable paint development using probabilistic ML methods in collaboration with Mankiewicz, Füll-Lab, the Niederrhein University of Applied Sciences, the Fraunhofer Institute and of course AOM system could imagine.

Raw material manufacturers, paint manufacturers and paint processing companies continuously collect material and process data. However, this data is not currently shared and not correlated with each other. It is also unclear whether the data currently collected even enables the creation of valid prediction models. The broad positioning of the consortium in the Na, Logisch project means that a possible data gap can be closed and valid prediction models can be created. The partners PI Probaligence in cooperation with the Niederrhein University of Applied Sciences demonstrated the fundamental feasibility of this in March 2022. Sections of a high-throughput plant were optimized using AI.

As of today, there is no standardized data exchange along the value chain in paint technology, so that the basis for the use of digital technologies has so far been fundamentally non-existent or only partially present in this sector of the economy. A “trial and error” process is therefore state of the art in paint development and production as well as in paint processing. This process, the “error”, leads to comparatively high scrap rates in paint development and production. Paint manufacturers record 5 to 30 tons (depending on production volume) of paint waste per year. More complex, more than one raw material, substitutions in these formulations are almost impossible in these processes with the same performance of the systems. This creates an enormous formulation effort in the laboratory, which is often not started due to the scarcity of R&D hours.

It is similar in paint processing; painting processes have to be adjusted until the desired paint result is achieved. In complex painting systems, this process can take several months. The components manufactured up to that point are declared as scrap.

In the work package description, this value chain is divided into two sections. Value chain 1 relates to paint formulation and paint production. Value chain 2 relates to the area of paint application and the correlation with the appearance of the painted surface.

It must be noted that painting processes are particularly energy-intensive. For example, 40% of the total primary energy requirement in automobile production is accounted for by painting (baking). If the nature or composition of the paint changes over time, for example due to regulations (REACH), if the paint development process is followed in accordance with the state of the art, painting errors and complex adjustment work on a painting system can be expected. A digital paint twin is being created in the Na, Logisch project and will help to apply paints efficiently and with few adjustment attempts in the future. The digital technologies developed in the Na, Logisch project are to be made available to German paint manufacturers and the paint processing industry. This enables paint manufacturers throughout Germany to avoid up to 2,000 tons of paint waste and produce digital, smart paints in the sense of a digital twin. In addition, the use of renewable raw materials is made possible and development times are significantly shortened. In the future, the paint processing industry will use the prediction models to adjust the painting process and minimize scrap rates. The Na, Logisch project also lays the technical foundation for the use of sustainable paint raw materials, whose material properties tend to fluctuate greatly, and enables the rapid adaptation of existing recipes while maintaining benchmark quality. For example, through regulations, e.g. CMR substances through REACH.