Probabilistic Property Modeling for Reliable Casting Design and Production
摘要
A new methodology leverages virtual casting process simulations to predict the statistical distribution of local mechanical properties, based on computed local microstructures and defects. The extensive result sets generated from each simulation can be statistically evaluated in a manner similar to numerous individual tensile tests. By incorporating process variability into the simulation, this innovative approach enables probabilistic modeling of expected local properties. Consequently, this method supports designers during the design phase for reliable component development and assists casters in achieving robust process designs.