Artificial intelligence aims to improve quality and efficiency

© Fraunhofer IPA

With Artificial Intelligence (AI), a research team from Fraunhofer IPA aims to make the painting of plastic parts in automotive and commercial vehicle manufacturing more efficient. To do this, intelligent algorithms will analyze all data generated during the painting process and provide early warnings of potential errors.

Painting is still considered a process that cannot be fully controlled today. There is a risk of rejects, equipment failures, and rework, because, for example, the specified paint layer thickness cannot be maintained everywhere. The ambitions of a research team from the Fraunhofer Institute for Manufacturing Engineering and Automation IPA are therefore quite high: they want to reduce the number of errors by 30 percent and downtime by 20 percent. Additionally, the annual paint consumption and the ramp-up time for new colors should each be reduced by ten percent.

The scientists aim to achieve this by first combining quality data, such as visible painting defects or measurements of the paint layer thickness, with process data from the plant control system. From this data, a detailed behavioral model is to be developed, which will then be evaluated using a machine learning method. The algorithms should detect impending quality deviations early and also indicate their causes immediately.

High degree of automation and digitalization enables the use of AI

The researchers will focus on the painting of bumpers, mirrors, door handles, and other plastic components in the automotive and commercial vehicle sectors. "This industry has a large product volume and, consequently, a lively interest in increasing efficiency," says Oliver Tiedje, head of the Wet Application and Simulation Technology group at Fraunhofer IPA. "Furthermore, the painting process systems have a very high level of automation and digitalization, making the use of AI promising."

The research project "Efficiency increase of painting processes through multilayer networking of process and quality data using self-learning behavior modules" (pAInt-Behaviour) is ongoing until May 31, 2024, and is supported by the Federal Ministry of Education and Research with nearly 1.3 million euros. Project partners include, alongside Fraunhofer IPA, b+m surface systems GmbH, AOM Systems GmbH, HELMUT FISCHER GMBH INSTITUTE FOR ELECTRONICS AND MEASUREMENT TECHNOLOGY, and SMP Automotive GmbH.