Predictive quality assessment for complex production processes

The aim of the project at Furtwangen University is to counteract the increasing complexity of manufacturing processes with methods of artificial intelligence (AI) and to realise a predictive quality assessment in manufacturing. The skilled personnel are to be relieved of this task by a digital master. Three AI methods will be used for the three most important quality assurance tasks: process planning, process observation and process optimisation. Firstly, an expert system is to be set up. For this purpose, a database will be stored that will be used to pre-select tool specifications and process parameters. Data-driven digital twins will enable AI simulations. By observing the digital twin, predictions can be made about the actual production process. Under the keyword "eXplainable AI", attempts are made to understand how the AI arrives at the proposed results and what it is likely to plan in the next step. This form of "root cause analysis" is to be incorporated into quality assurance.