High demands, high quality, high risk: even when producing sophisticated parts, manufacturers in the automotive, metalworking and medical technology industries have to meet increasing demands on the quality of their products. In the event of an emergency, they must be able to provide complete proof of compliance with the quality parameters in order to protect themselves from recourse claims. The entire quality control process must therefore be designed to be as precise and reproducible as possible. The quality of manual inspections depends on the skills and knowledge of the employees and the time available. Documentation is also time-consuming. Automating the process minimizes these variances. Exemplary solutions from Kistler for a research project show just how far it can go: together with the OST - Ostschweizer Fachhochschule in Rapperswil (CH) - the measurement technology expert is researching the new possibilities that automation opens up in terms of data quality and improving quality predictions using AI.

Automated, reproducible random sample inspection

Manufacturing companies usually rely on the statistical process control method to check the manufactured quality. This method determines the frequency and scope of the random samples so that users can monitor the production process based on previously defined relevant quality parameters. As these samples were previously taken, transported and checked manually, they tie up a great deal of time and personnel resources, depending on the scope and throughput of production. In addition, the quality of the data collected depends on the skills of the inspectors and can vary from person to person. The more safety-relevant the produced parts are, the lower the error tolerances and the higher the frequency of random samples - and therefore the more error-prone the inspection. At the same time, the costs of possible recourse claims increase. Automated, reproducible random sample inspection is therefore an attractive alternative for minimizing the costs of defective parts.

Consistently high test process reliability

Productions with a high throughput volume and parts that are subject to similar inspection requirements in particular benefit from a holistic, automated inspection concept: in mass production, cycle times and the number of parts produced cause enormous effort in manual random sample inspection. By automating the entire quality assurance process, manufacturers benefit from reproducibility and cost reduction. An interdisciplinary approach, such as that pursued by Kistler, offers additional advantages: The areas of sensors, process monitoring, automation, optical image processing, software for data analysis and mechanical engineering work together under one roof and cooperate efficiently thanks to many years of experience - also in the injection molding sector.

Interdisciplinary approach

Together with the Institute of Materials Engineering and Plastics Processing at the University of Applied Sciences in Rapperswil, Eastern Switzerland, Kistler is setting up a fully automated manufacturing and testing process.

© Kistler

At the beginning of such a test concept creation, the focus is on the requirements of the test part: Together with the manufacturers, the commissioned team works out the necessary quality-relevant test parameters - usually from the area of surface defects and dimensional accuracy - and selects the appropriate test methods such as incident or transmitted light, 2.5D or 3D testing. Mechanical tests of pressure, force and torque can also be integrated into the concept with the help of Kistler sensors. Experts from the competence center design the test cell accordingly: in addition to the number and positioning of the camera stations with lighting elements, the path of the test part plays a particularly important role; the aim is to achieve consistently smooth, efficient part handling throughout the entire test process. The integrated safety concepts monitor the systematic sequence of each step as well as the handshakes and ensure process reliability. At the same time, they prevent data loss. The machine sends the collected data via the OPC UA interface to the operator's higher-level quality assurance system and to databases for analysis.