The digitization and automation of plastic profile extrusion processes are the objective of the joint research project AutoRock carried out by the SKZ Plastics Center and the Fraunhofer Institute for Integrated Circuits IIS. The key premise of the project is to develop a solution that enables automated and reliable defect detection in plastic profiles directly during production. The project addresses one of the main challenges in plastics processing, namely the early detection of quality deviations before they lead to rejects, material losses or costly rework.
To build a reliable database for an AI-supported monitoring system, profiles with characteristic defects were deliberately produced within the project. In addition, manufacturing companies provided scrap products with typical defects encountered in practice. The collected material includes, among others, cavities, foreign material inclusions, pores and geometric errors, which makes it possible to reproduce repeatable and at the same time realistic operating scenarios. This forms the basis for a data set that reflects the actual operating conditions of extrusion lines, including the occurrence of complex and overlapping internal discontinuities in profiles.
Database preparation and non-destructive testing
The profile samples were then analyzed using modern X-ray computed tomography systems and established non-destructive testing methods. The use of advanced CT systems enabled detailed imaging of the internal structure of the products, including volumetric defects and material inhomogeneities. The data set obtained in this way now forms the basis for training an artificial intelligence system that is intended in future to automatically detect, classify and evaluate defects in terms of their relevance for product quality.
Thanks to the diversified range of training data, the AI system can also learn to recognize complex or overlapping defect structures. It is important not only to distinguish between defects that are critical and those that are irrelevant for the function of the profile, but also to link them to the extrusion process conditions. This allows the development of algorithms that in future may serve as a basis for adaptive control of the production line based on information from the inspection system.
Concept of an online measurement demonstrator
In parallel, the project has developed the concept for a specially adapted demonstrator that will in future enable online measurements during profile extrusion. The design requirements for such a system are demanding, because the device must ensure stable and repeatable measurements under continuous material movement, high line speeds and limited installation space in the production area. In addition, radiation protection requirements associated with the use of X-rays must be fulfilled.
The technology must also be designed so that it can be integrated into existing production lines without disrupting the process flow. This covers both the mechanical installation of the system and communication with the extruder control units and quality assurance systems. The concept developed in AutoRock takes these constraints into account and creates the basis for a new generation of CT systems adapted for inline operation in the plastics processing industry. Practical tests are scheduled to be carried out later this year under conditions close to industrial practice.
CAD model of the pure online X-ray CT measurement system without cladding and radiation protection. (Photo: Fraunhofer IIS)
Real-time monitoring and quality documentation
The combination of artificial intelligence algorithms with inline-capable X-ray CT technology opens up new possibilities for the plastics processing industry. Real-time monitoring systems will enable immediate detection of quality deviations and dynamic adjustment of process parameters, for example temperature, extrusion speed or haul-off speed. It is assumed that the implementation of such solutions will significantly reduce scrap rates, increase process stability and improve raw material efficiency.
At the same time, continuous acquisition of measurement data from the CT system enables detailed documentation of product quality. Such information can be used both for internal process optimization and to demonstrate compliance with customer quality control requirements. Integration of the data with higher-level systems, such as manufacturing execution systems, is in line with the broader trend of digitization and the development of Industry 4.0 solutions in the plastic profile extrusion sector.
Project funding and invitation to industry
The AutoRockproject (funding code 01IF23187N) is funded by the German Federal Ministry for Economic Affairs and Energy as part of the Industrial Collective Research program IGF, implemented via the German Aerospace Center DLR. Cooperation between research institutions and processing companies makes it possible to obtain representative product samples and to adapt the technologies under development to the needs of industrial practice.
Companies interested in solutions for non-destructive quality control of extruded profiles are invited to contact the project partners for further information. The SKZ Plastics Center is a climate protection oriented organization and a member of the Zuse Association, which brings together independent, industry-related research institutes. The aim of this association is to improve the performance and competitiveness of industry, in particular small and medium-sized enterprises, through innovation and networked cooperation.
