The test series for the DarWIN research project have been running since the end of 2020. The aim of the AI project is to learn detailed behavior models of injection molding machines on high-frequency machine data. To do this, machines from different manufacturers that produce similar parts over time are used. In this way, the behavior models learned on one machine should also be transferable to other machines without having to learn the models again completely for each machine. The behavior models suggest optimized process parameters for the next shot in order to produce without scrap at the minimum possible cycle time.
Comparing injection molding machines and processes — independent of manufacturer
The experts from SKZ and plus10 are conducting application-oriented research into the latest machine learning models for describing the behavior of cyclical manufacturing processes using injection molding as an example. The focus is on online capability, i.e. the formation and extension of a model while the process is running. In addition, the investigation of the transferability of pre-trained machine learning models from one machine to similar, non-identical machines also plays a central role.
An “evolution learner” from the company plus10 generates optimization suggestions based on behavioral comparison with all identical or similar machines involved. SKZ provides a wide range of machines from the manufacturers ARBURG, ENGEL, KraussMaffei, Sumitomo (SHI) Demag and WITTMANN BATTENFELD for the test series.
plus10’s expertise in intelligent data processing and automated production optimization by means of continuously learning models will flow into the project. The injection molding specialists evaluate the transferred optimization proposals and check the component quality in the test laboratory.
The “DarWIN” research project (BMBF funding code 01IS20066) is funded by the German Federal Ministry of Education and Research (BMBF) and is expected to end in November 2021. Publication of the final results is planned for the end of the year.