Supervision of Quality in Process Chains by means of Learning Process Models

This paper deals with a new method for the supervision and optimization of process parameters, influencing and disturbance quantities and quality in the process chain of manufacturing with high-quality demands. This method, demonstrated using series production as an example, starts from the capability of modern information technologies to learn, simulate and optimize complex but reproducible processes in production. Its aim is to forecast the achievable quality at the beginning of manufacture, taking into account all influencing and disturbance quantities. The method permits learning step by step from deviations and continuous improvement of processes. It appears advisable to use modern high-performance parallel computers to solve this complex task.