In industrial manufacturing processes, the most important tasks are usually concerned with an improvement of the quality outcome, a reduction of waste and an improvement of the process stability. This applies specifically to injection molding and metal machining with a turning lathe. Injection molding is the state-of-the-art process to produce plastic parts in different shapes and sizes. Machining on the turning lathe allows to perform highest precision process steps on expensive raw materials using specialized cutting tools. In both applications, machine learning allows to harness the process information more effectively to improve the yield of in-specification parts and understanding the quality-determining process mechanisms. In this talk Kistler approaches to implement solutions to realize this in both applications will be presented. In both applications, the insights obtained by machine learning, though different in nature, allow a reduction of waste. In both fields, further investigations are aimed to generalize these models and fully integrate them into the corresponding process automation and control.

Industrial machine manufacturers have heavily invested in IoT infrastructures in the last decade, yet tools that extract actionable insights remain illusive. To drastically reduce the maintenance costs and data analysis efforts, Amplo developed a smart maintenance platform which gives service engineers themselves access to state of the art operational machine learning systems that are deployable without any code or machine learning expertise. Currently, the platform provides models that diagnose failures, analyse production quality and monitor machine performance. 

The benefits are clear. Tritium avoids hours of data analysis and enjoys automated root cause analysis, allowing them to instantly send out repair orders. Solarmanager is able to notify thousands of homeowners when their PV panels are underperforming. TB Safety now services their filters only when necessary, instead of inspecting them every year.