Kistler Introduces Machine Learning to Cavity Sensors
Düsseldorf, Germany — Known for its mold cavity sensors, Kistler Instrumente AG took that up a notch at K 2019, ushering in artificial intelligence and machine learning for injection molding with the new ComoNeoPREDICT.
“I believe we are one of the first ones to develop machine learning and introduce it into the marketplace,” Oliver Schnerr, Kistler’s head of sales, said at the company’s K show news conference in Düsseldorf.
ComoNeoPREDICT is based on Kistler’s ComoNeo process monitoring and control system for injection molding based on mold cavity pressure, enabling switchover from the injection phase to the packing stage and balancing hot runner control systems. That data can be transmitted to a molder’s plantwide systems.
ComoNeoPREDICT uses what Schnerr called “machine learning algorithms” to calculate in advance the properties of manufactured parts. The software makes real-time changes to the molding process and tells the molding press to kick out bad parts.
The predictive system can eliminate the need for a separate quality control step, he said.
Schnerr compared the software and sensors to sensors in a car that warn when another car is passing or getting too close.
Kistler is also meeting another trend in injection molding: very small parts for medical, which cannot have any vestige from the pressure sensor physically touching the part. Kistler has developed tiny “contactless” sensors that measure defection of the steel to gauge mold cavity pressure.