Multi objective optimization of FDM 3D printing parameters set via design of experiments and machine learning...

Area of Science:

• Materials Science and Engineering
• Additive Manufacturing
• Polymer Science

Background:

• Optimizing Fused Deposition Modeling (FDM) 3D printing for engineering applications is complex due to intricate process parameter-mechanical property relationships.
• Achieving maximum performance in FDM-printed parts requires a thorough understanding of how printing parameters influence material behavior.

Purpose of the Study:

• To investigate the influence of key FDM printing parameters (layer thickness, extrusion temperature, printing speed, deposition pattern) on the mechanical properties of ABS specimens.
• To develop predictive models for mechanical properties and identify optimal printing parameters for enhanced performance.

Main Methods:

• Utilized a full factorial Design of Experiments (DoE) approach to study main and interaction effects.
• Employed Analysis of Variance (ANOVA) for statistical analysis of tensile strength, elastic modulus, and strain at maximum stress.
• Applied Response Surface Method (RSM) quadratic regression and Random Forest (RF) for enhanced predictive modeling, alongside a Non-dominated Sorting Genetic Algorithm II (NSGA-II) for multi-objective optimization.

Main Results:

• Deposition strategy was identified as the most influential parameter on mechanical response.
• The "Lines" deposition pattern provided the best balanced results, maximizing elastic modulus to 1381 MPa and tensile strength to 33.3 MPa.
• Random Forest regressor improved predictive capability by over 40% compared to traditional methods.

Conclusions:

• The study provides valuable insights for optimizing FDM printing parameters at the design stage.
• The developed models and optimization approach can guide the selection of printing setups for demanding engineering applications.
• Experimental validation confirmed the accuracy of the optimized printing parameters.