Prediction of Mechanical Properties of 3D Printed Parts Using Machine Learning Techniques

Additive manufacturing (AM), specifically Fused Deposition Modelling (FDM), is one of the most widely adopted fabrication technologies, enabling layer-by-layer construction of complex geometries with minimal material waste. The mechanical performance of FDM-fabricated components — including tensile strength, hardness, surface roughness, and impact strength — is critically dependent on process parameters such as layer thickness, printing speed, and infill density. This study presents a comprehensive machine learning (ML) based comparative framework for predicting four mechanical properties across three thermoplastic materials: ABS, PLA, and PETG. A full factorial design (27 runs per material) was employed. Eight ML algorithms — ANN, KNN, LR, SVR, XG Boost, Bayesian Ridge, BPNN, and Random Forest — were implemented in Python using Google Colab and evaluated using R2, MAE, MSE, and RMSE. Results show ANN achieved peak R2=0.9894 for PETG tensile strength and R2=0.9822 for ABS hardness. SVR was best for ABS impact test (R2=0.8928). These findings provide a validated ML-based optimization framework for FDM process parameter selection.