Process chain simulation is a rapidly growing field of research. On one hand, the allowed tolerance bands for the properties of semi-finished and final products are becoming continually narrower, thereby demanding increased accuracy of the respective prediction models; on the other hand increasing computational capacities and improved simulation methods allow to include the complete chain of production processes – if necessary from primary shaping to surface finishing – into the models for predicting the final properties.
For enabling an end-to-end process chain simulation for property prediction, one needs detailed knowledge and simulation models for each individual manufacturing process as well as appropriate interfaces between the individual simulations. Some examples for process chains investigated at MCL are
- casting and heat treatment of aluminium components,
- rolling, accelerated cooling and straightening of steel sheets,
- forming, inductive hardening and tempering of steel bars,
- forging and direct aging of nickel-base superalloys.
For implementing such process models in model based plant control systems, the computing time must be reduced from many hours, as it is typical for finite element simulations, to seconds or milliseconds for real-time applications. To this purpose, MCL has been developing effective methods for model reduction in the course of previous and current research projects.