In plastic injection molding, efficient clamping force plays a critical role in maintaining product quality, minimizing energy consumption, and extending the lifespan of molds. Traditional methods often rely on empirical rules or fixed maximum force settings, which can lead to inefficiencies and higher costs. Therefore, a reliable prediction model can be a solution to optimize the injection molding process. Such a model has the ability to balance defect reduction, operational efficiency, and equipment durability. It also can correlate the complex interactions between process parameters and their associated uncertainties. This study investigates the injection molding of plastic bushing (size 63 mm) as the case study. A Plackett-Burman Design of Experiments (DOE) with two levels for six key factors (melt temperature, cooling time, packing time, holding pressure, injection pressure, and injection time) is used as the screening step. This method offered 36 experiments and after the analysis, holding pressure showed almost no impact on the maximum clamping force, leading to its exclusion from further analysis. Then, a Central Composite Design (CCD) was employed for the remaining five factors, considering five levels for each, resulting in 32 experiments. Finally, a Naive Bayes algorithm was trained to predict the results of the process, and its uncertainty was evaluated using the bootstrap sampling method. The findings of this study provide a clear understanding of developing high-quality prediction models with low uncertainty in prediction which can lead to optimizing the injection molding parameters.

Uncertainty analysis on a Naïve Bayes prediction model for the plastic injection molding process / Ardestani, A.M., Genta, G., Calaon, M., Hattel, J.H., Kulahci, M., Tosello, G.. - ELETTRONICO. - (2025), pp. 246-247. (25th International Conference of the European Society for Precision Engineering and Nanotechnology, EUSPEN 2025 Zaragoza (Spain) 2025).

Uncertainty analysis on a Naïve Bayes prediction model for the plastic injection molding process

Genta G.;
2025

Abstract

In plastic injection molding, efficient clamping force plays a critical role in maintaining product quality, minimizing energy consumption, and extending the lifespan of molds. Traditional methods often rely on empirical rules or fixed maximum force settings, which can lead to inefficiencies and higher costs. Therefore, a reliable prediction model can be a solution to optimize the injection molding process. Such a model has the ability to balance defect reduction, operational efficiency, and equipment durability. It also can correlate the complex interactions between process parameters and their associated uncertainties. This study investigates the injection molding of plastic bushing (size 63 mm) as the case study. A Plackett-Burman Design of Experiments (DOE) with two levels for six key factors (melt temperature, cooling time, packing time, holding pressure, injection pressure, and injection time) is used as the screening step. This method offered 36 experiments and after the analysis, holding pressure showed almost no impact on the maximum clamping force, leading to its exclusion from further analysis. Then, a Central Composite Design (CCD) was employed for the remaining five factors, considering five levels for each, resulting in 32 experiments. Finally, a Naive Bayes algorithm was trained to predict the results of the process, and its uncertainty was evaluated using the bootstrap sampling method. The findings of this study provide a clear understanding of developing high-quality prediction models with low uncertainty in prediction which can lead to optimizing the injection molding parameters.
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11583/3017652
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo