Abstract The complexity of the injection moulding process, combined with the number of injection variables and parameters that are manipulated, influences both the interpretation of the results obtained and the process of determining the injection parameters. This study proposes a sequential optimization methodology using statistical methods, such as Design of Experiments (DOE), Principal Component Analysis (PCA), and Analysis of Variance (ANOVA), to explore the solution space within the parameter search space that traditional trial-and-error methods tend to overlook. In this context, the research utilized isothermal and transient numerical simulations, first defining the filling and back-pressure parameters and subsequently the cooling parameters. The results demonstrated significant effectiveness, achieving cumulative reductions of 75.04% in warpage under ideal conditions and 62.67% in transient simulations, which provide a more realistic representation of the process dynamics compared to steady-state simulations. In addition to ensuring that the parts remained within the dimensional tolerance range of 1.6 mm (DIN 16742 standard), the methodology enabled cycle-time optimization by reducing the cooling time from 10 seconds to 1 second, resulting in a 90% reduction without compromising dimensional quality. It can be concluded that integrating statistical tools with numerical simulation provides greater scientific predictability, improved process stability, reduced operating costs, and applicability to different polymers and complex geometries.
Amarante et al. (Wed,) studied this question.