AI at rework for 30% less error rate

Data driven solutions create maximum efficiency in the rework process. For Knorr-Bremse this is an essential quality control process at their plant on Aldersbach. Together with STIWA Group rework was now enhanced to be a self-learning system - with the goal to make rework altogether redundant.

 

Since 2011 Knorr Bremse builds upon the collaboration with STIWA Group for a structured data-based process analysis with professional data based rework.With AMS analysis-Cl rework another milestone was set in the last year: As s self-learning tool for professional data-based rework, AMS Analysis Cl Rework is integrated in the machine control center “AMS ZPoint-Cl”. “The machine operators thus have access to all error messages and the data from the process analysis directly at the machine.With the knowledge of how faults were solved correctly in the past, they can immediately and rapidly take action with the correct rework actions”, explains Thomas Langthaler, Team Leader Engineering and technical contact person for all Knorr-Bremse plants worldwide. Essential for implementation was a completely uniform structure with a common data base and a user-friendly look and feel, says Langthaler: “During refurbishment of necessary technical structures it became very quickly apparent that the calculation algorithm towards Machine Learning was necessary”, according to Langthaler.

 

Self-learning system

Until now specially trained machine operators have analyzed possible error sources at the rework station for defective components in the production process. With their years of experience they were able to evaluate whether a delivered component was defective, something was stuck or the lubrication insufficient. The success of this search for causes and the measures taken from it were always dependent on the knowledge of one person. The rework analysis tool automates a majority of these processes and expands the information base for the self-learning system continuously. Along with the consolidated data from the overall system ZPoint-Cl, more and more information is successively available in the future to detect and eliminate the causes for quality deficiencies.

“In the future”, explains Armin Einsenreich, Team Leader Quality Planning in the Unit mechatronic systems at the plant in Aldersbach, “the machine control center, based on all the data at its disposal, can predict when a certain process starts to produce deficient systems. Then measures can be taken in due course.”

 

Lasting decrease of production costs

The system is running very promisingly. “We expect an improvement of the functional-test-failure-rate by 30 percent from it”, says Gerhard Schwarz, Plant Manager at Knorr Bremse in Aldersbach. What has been successfully launched in Aldersbach will now be implemented throughout the entire corporation. From 2023 on, AMS Analysis-Cl Rework will be rolled out at the location in the Hungarian Kecskemét and at the North American subsidiary Bendix. The data bases of ZPoint-systems on all sites are transferred in a cloud-solution. This way the machine operators in Hungary and North America can draw from the comprehensive practical knowledge from Aldersbach from day one.