Multi objective optimization of FDM 3D printing parameters set via design of experiments and machine learning algorithms
Abstract The choice of the optimal printing setup for Fused Deposition Modeling (FDM) 3D-printing technology is challenging due to complex interactions between process parameters and mechanical properties. This especially affects engineering applications where the maximum performance is required. To...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-05-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-01016-z |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850273031831683072 |
|---|---|
| author | Antonio Panico Alberto Corvi Luca Collini Corrado Sciancalepore |
| author_facet | Antonio Panico Alberto Corvi Luca Collini Corrado Sciancalepore |
| author_sort | Antonio Panico |
| collection | DOAJ |
| description | Abstract The choice of the optimal printing setup for Fused Deposition Modeling (FDM) 3D-printing technology is challenging due to complex interactions between process parameters and mechanical properties. This especially affects engineering applications where the maximum performance is required. To address this challenge, this study explores the influence of main controllable printing parameters including layer thickness, extrusion temperature, printing speed and deposition patterns, on the mechanical properties of FDM-printed ABS specimens using the Design-of-Experiments (DoE) approach by a $$3^4$$ 3 4 full factorial design. Main-effects and Interaction-effects on tensile strength, elastic modulus, and strain at maximum stress are investigated via ANOVA analysis, providing interesting hints to evaluate at the design stage. Given the complexity of these effects, a deeper investigation is conducted with a quadratic regression model of the Response Surface Method and the Random Forest regressor, with the latter enhancing the predictive capability ( $$R^2$$ R 2 ) on test data by more than 40% for all the mechanical properties. Eventually, a Genetic Algorithm (NSGA-II) is integrated to estimate the optimal parameter set for multiple responses. Overall results indicate that the deposition strategy is the parameter affecting the most the overall mechanical response, with “Lines” pattern providing the best balanced results in maximizing the elastic modulus and the tensile strength, respectively 1381 MPa and 33.3 MPa. Testing of a set of specimens printed with the found optimal parameters confirm the model’s prediction. |
| format | Article |
| id | doaj-art-89b43e95544b410e91e1121dc61eda38 |
| institution | OA Journals |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-05-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-89b43e95544b410e91e1121dc61eda382025-08-20T01:51:38ZengNature PortfolioScientific Reports2045-23222025-05-0115111610.1038/s41598-025-01016-zMulti objective optimization of FDM 3D printing parameters set via design of experiments and machine learning algorithmsAntonio Panico0Alberto Corvi1Luca Collini2Corrado Sciancalepore3Department of Engineering for Industrial Systems and Technologies, University of ParmaDepartment of Engineering for Industrial Systems and Technologies, University of ParmaDepartment of Engineering for Industrial Systems and Technologies, University of ParmaDepartment of Engineering for Industrial Systems and Technologies, University of ParmaAbstract The choice of the optimal printing setup for Fused Deposition Modeling (FDM) 3D-printing technology is challenging due to complex interactions between process parameters and mechanical properties. This especially affects engineering applications where the maximum performance is required. To address this challenge, this study explores the influence of main controllable printing parameters including layer thickness, extrusion temperature, printing speed and deposition patterns, on the mechanical properties of FDM-printed ABS specimens using the Design-of-Experiments (DoE) approach by a $$3^4$$ 3 4 full factorial design. Main-effects and Interaction-effects on tensile strength, elastic modulus, and strain at maximum stress are investigated via ANOVA analysis, providing interesting hints to evaluate at the design stage. Given the complexity of these effects, a deeper investigation is conducted with a quadratic regression model of the Response Surface Method and the Random Forest regressor, with the latter enhancing the predictive capability ( $$R^2$$ R 2 ) on test data by more than 40% for all the mechanical properties. Eventually, a Genetic Algorithm (NSGA-II) is integrated to estimate the optimal parameter set for multiple responses. Overall results indicate that the deposition strategy is the parameter affecting the most the overall mechanical response, with “Lines” pattern providing the best balanced results in maximizing the elastic modulus and the tensile strength, respectively 1381 MPa and 33.3 MPa. Testing of a set of specimens printed with the found optimal parameters confirm the model’s prediction.https://doi.org/10.1038/s41598-025-01016-zAdditive ManufacturingDesign-of-Experiments (DoE)Random ForestNSGA-II |
| spellingShingle | Antonio Panico Alberto Corvi Luca Collini Corrado Sciancalepore Multi objective optimization of FDM 3D printing parameters set via design of experiments and machine learning algorithms Scientific Reports Additive Manufacturing Design-of-Experiments (DoE) Random Forest NSGA-II |
| title | Multi objective optimization of FDM 3D printing parameters set via design of experiments and machine learning algorithms |
| title_full | Multi objective optimization of FDM 3D printing parameters set via design of experiments and machine learning algorithms |
| title_fullStr | Multi objective optimization of FDM 3D printing parameters set via design of experiments and machine learning algorithms |
| title_full_unstemmed | Multi objective optimization of FDM 3D printing parameters set via design of experiments and machine learning algorithms |
| title_short | Multi objective optimization of FDM 3D printing parameters set via design of experiments and machine learning algorithms |
| title_sort | multi objective optimization of fdm 3d printing parameters set via design of experiments and machine learning algorithms |
| topic | Additive Manufacturing Design-of-Experiments (DoE) Random Forest NSGA-II |
| url | https://doi.org/10.1038/s41598-025-01016-z |
| work_keys_str_mv | AT antoniopanico multiobjectiveoptimizationoffdm3dprintingparameterssetviadesignofexperimentsandmachinelearningalgorithms AT albertocorvi multiobjectiveoptimizationoffdm3dprintingparameterssetviadesignofexperimentsandmachinelearningalgorithms AT lucacollini multiobjectiveoptimizationoffdm3dprintingparameterssetviadesignofexperimentsandmachinelearningalgorithms AT corradosciancalepore multiobjectiveoptimizationoffdm3dprintingparameterssetviadesignofexperimentsandmachinelearningalgorithms |