Lara Maroto-Diaz, Oisin Boyle, Marcus Law, Joe Brindley, Benoit Daniel, Gencoa Ltd, Liverpool, United Kingdom
Vacuum processes can suffer from faults such as air leaks or organic contamination. With pressure gauges alone, it can be difficult to constantly monitor for such errors. Residual Gas Analysers (RGA) can facilitate this, giving partial pressure measurements. However, determining if a fault is present from vacuum dynamics and partial pressures can be non-trivial. Manufacturing wise, the more important question is if the fault is substantial enough to lead to a failed process. For roll-to-roll processes in particular, detection of detrimental outgassing from web substrates is a concern.
Why don’t we use AI to automatically identify these problems?
AI has been implemented in combination with vacuum gas analysis based upon RPEM (Remote Plasma Emission Monitoring). RPEM operates from near atmosphere to 10^-7 mbar hence this technology can be used to monitor the pump down of a process and determine vacuum quality such as air or water leaks and any form of contamination. The model is pretrained to monitor and allows users to obtain real time predictions of the success of their vacuum process. In this talk, some case studies of the benefits of AI in combination of vacuum gas analysis will be presented and its ability to predict vacuum process performance on industrial like processes.