Dr Jacek Kozłowski (Warsaw University of Technology)

Virtual experts, as a support in the issues of predictive maintenance and predictive quality in casting processes

Abstract:

In the era of Industry 4.0 it seems to be significant to develop autonomous supporting tools on many levels in the company, i.e.: management, quality control, research and development, maintenance or operators (including operators 4.0). Therefore, it seems reasonable to use AI methods and tools as a source of knowledge extraction by applying special techniques of their interrogation. Such approach allows to create the so-called Virtual Experts, as the form of autonomous computer systems supporting the decision-making process or even directly controlling the process. Therefore, examples of such solutions will be presented, as a solutions for off-line support of detection the most important groups of process parameters, taking into account the phenomena of their synergistic or competitive action, as well as solutions supporting the on-line prediction process for time-varying phenomena, allowing the prediction of quality or breakdowns with a defined time advance, based on the behaviour of parameters in time (their behavioural patterns).