Multi-objective optimization of injection molding process using interpretable extreme gradient boosting model based on improved north eagle optimization algorithm

During the injection molding of sensor housings, plastic parts often suffer from warpage deformation and volumetric shrinkage. To address this, Moldflow simulation and central composite face (CCF) design were used to generate a dataset. An improved northern goshawk optimization (INGO)-XGBoost model,...

Full description

Saved in:
Bibliographic Details
Published in:Journal of mechanical science and technology Vol. 39; no. 10; pp. 6171 - 6180
Main Authors: Zhu, Liuyu, Fan, Xiying, Guo, Yonghuan, Wang, Zhijiang, Hua, Junyi, Li, Lie
Format: Journal Article
Language:English
Published: Seoul Korean Society of Mechanical Engineers 01.10.2025
Springer Nature B.V
대한기계학회
Subjects:
Accuracy
Control
Cooling
Datasets
Deformation
Dimensional stability
Dynamical Systems
Engineering
Housings
Industrial and Production Engineering
Injection molding
Integrated approach
Mechanical Engineering
Multiple objective analysis
Optimization algorithms
Original Article
Pareto optimization
Process parameters
Science
Sensors
Simulation
Support vector machines
Vibration
Warpage
기계공학
SAHP
CRITIC-TOPSIS
MOMVO
INGO-XGBoost
Process parameters
CCF
ISSN:1738-494X, 1976-3824
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:During the injection molding of sensor housings, plastic parts often suffer from warpage deformation and volumetric shrinkage. To address this, Moldflow simulation and central composite face (CCF) design were used to generate a dataset. An improved northern goshawk optimization (INGO)-XGBoost model, enhanced by three strategies, was compared against NGO-XGBoost, AdaBoost-SVM, and AdaBoost-ELM models, demonstrating superior performance. SHAP analysis was applied to interpret the INGO-XGBoost model, and multi-objective multiverse optimization (MOMVO) was used to generate the Pareto front. The CRITIC-TOPSIS method was then employed to select the optimal process parameters. Results show that warpage deformation and volumetric shrinkage were reduced by 30.9 % and 8.7 %, respectively, compared to the initial settings. The proposed integrated prediction–optimization framework significantly improves the molding quality and dimensional stability of plastic parts, providing both theoretical support and a practical pathway for intelligent injection molding.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1738-494X
1976-3824
DOI:10.1007/s12206-025-0946-2