Mechanical Behavior Prediction of 3D-Printed PLA/Wood Composites Using Artificial Neural Network and Fuzzy Logic

Abstract

This study presents a novel approach to optimize and predict the mechanical properties of 3D-printed polylactic acid (PLA)/wood composites through artificial neural network (ANN) and fuzzy logic (FL) modeling. The research addresses the critical challenge of determining optimal process parameters in fused deposition modeling (FDM) of natural fiber composites. Using Taguchi's L27 orthogonal array, experiments were conducted with five key printing parameters: layer thickness (100-200-300 mu m), printing speed (PS) (40-60-90 mm/s), raster angle (RA) (0 degrees-45 degrees-90 degrees), infill density (ID) (30%-60%-90%), and nozzle temperature (NT) (190 degrees C-200 degrees C-210 degrees C). Analysis revealed that RA and PS were the most influential parameters, contributing 41.86% and 40.92% to tensile and compressive strengths, respectively. The developed ANN model demonstrated exceptional prediction accuracy with R2 values of 99.94% for both tensile and compressive strengths, surpassing the FL model's performance (R2 = 97.16%). The development of these models is crucial for accurately predicting mechanical behavior, allowing for efficient process optimization without extensive physical testing. Both methods demonstrated high prediction accuracy. Validation tests revealed that maximum errors of 1.95% and 2.81% for ANN and FL, respectively. The findings contribute valuable insights for the development of high-performance natural fiber composites and establish a foundation for future advanced manufacturing processes.