Towards Industry 4.0 in plastic injection molding: prediction and optimization of injection molding cavity filling using multivariable regression and genetic algorithms

Patricia Frontini
(Consejo Nacional de Investigaciones Científicas y Técnicas, CONICET, Argentina)

Hosted by Claudia Marano

Abstract:

Determining optimal filling conditions in injection molding is essential to ensure part quality and process stability; however, it remains challenging when process information is limited or difficult to obtain. To address this, a simple data-driven optimization strategy combining multivariable regression (MVR) and a genetic algorithm (GA) is proposed. Second-order MVR models were trained using reduced short-shot experimental data to quantify the influence of key processing variables on part weight and projected area. The fitted models were integrated into a GA-based inverse optimization scheme to identify parameter combinations that achieve complete cavity filling. The approach was experimentally validated, resulting in fully filled parts with average deviations of 2% in weight and 0.6% in projected area. Overall, the MVR–GA scheme provides a data-efficient alternative to trial-and-error, complex simulations, or highly data-intensive machine learning approaches.

P. M. Frontini established the first research group on deformation and fracture of polymeric materials in Argentina. Her expertise includes polymer mechanics, fracture behavior, and the performance of polymers and composites under complex loading.

> Flyer
> Link Teams meeting