Predicting the properties of extrusion-based additive manufacturing parts using classification algorithms and machine learning

Jabbari, Amin

doi:10.22034/ijme.2025.551862.2139

Predicting the properties of extrusion-based additive manufacturing parts using classification algorithms and machine learning

Articles in Press

Document Type : Original Article

Author

Amin Jabbari

Mechanics Research Institute of the Iranian Research Organization for Science and Technology

10.22034/ijme.2025.551862.2139

Abstract

Recent advancements in additive manufacturing (AM), artificial intelligence (AI), and machine learning (ML) have significantly influenced engineering practices, particularly in the design and fabrication of complex components. As the demand grows for predictive capabilities in assessing the mechanical properties of parts designed and produced via layer-by-layer 3D printing, machine learning algorithms offer promising solutions. This study examines the application of classification models to predict the tensile strength and Young’s modulus of components manufactured via fused deposition modeling (FDM) under various process parameters. Tensile test data were systematically analyzed to train decision tree, support vector machine (SVM), and k-nearest neighbor (KNN) models. Among these, the decision tree algorithm exhibited the highest predictive accuracy (90.9%), with an AUC of 0.92 and an F1 score of 0.88. These findings underscore the value of data-driven methodologies in enhancing engineering design processes and establishing reliable experimental databases, contributing to the evolution of intelligent digital manufacturing systems.

Keywords

3D Printing

Data Science

Artificial Intelligence

Fused Filament Fabrication

Material Properties

Subjects