Towards an accurate pressure estimation in injection molding simulation using surrogate modeling

The computational cost of high-fidelity injection molding simulations has been growing in the past years making it more and challenging to use them for performing analyses such as optimizations or uncertainty quantification. Surrogate modeling offers a cheaper way realize studies gaining attention field simulation. In this work, we propose compare three surrogate techniques along with two design experiment methods their ability predict pressure signal at surface node Moldflow simulation by varying process parameters. A Sobol sensitivity analysis is performed study contribution varied parameters on results. addition, one generated models used experimental sensor data improve estimation calibrating heat transfer coefficients during filling packing well pressure-dependency coefficient Cross-WLF viscosity model. This resulted major improvements predictions all 27 considered cases comparison using default model parameter.