Intelligent Modeling of Mechanical Behavior in FDM ‐Fabricated Graphene/ PLA Composites Using ANN

ABSTRACT

Polylactic acid (PLA) composites reinforced with graphene nanoplatelets (GNPs) offer potential for sustainable, high‐performance additive manufacturing. This study investigates the influence of graphene loading, layer height, infill density, and nozzle temperature on the mechanical behavior of fused deposition modeling (FDM)‐fabricated graphene/PLA composites. A Taguchi L27 orthogonal array was employed to design the experiments, and tensile strength, compressive strength, and Shore D hardness were evaluated experimentally. Scanning electron microscopy (SEM) was used to examine graphene distribution, matrix–reinforcement contact, interlayer continuity, void formation, and fracture characteristics. Increasing graphene content was associated with improved microstructural continuity and fewer visible interfacial defects. The composite containing 1.5 wt% graphene, printed at a layer height of 0.1 mm, 100% infill density, and a nozzle temperature of 220°C, exhibited the highest observed mechanical responses, with tensile strength of 67.42 MPa, compressive strength of 81.35 MPa, and Shore D hardness of 82.4. Artificial neural network (ANN), gene expression programming (GEP), and response surface methodology (RSM) models were developed and compared using the coefficient of determination ( R ² ), root mean square error (RMSE), and mean absolute error (MAE). ANN demonstrated the best predictive performance, achieving R ² values of 0.982, 0.978, and 0.971 for tensile strength, compressive strength, and hardness, respectively. Confirmation experiments yielded prediction errors below 1% for all three responses using ANN. The integrated experimental and computational framework supports mechanical performance prediction and process‐parameter assessment for graphene‐reinforced PLA composites within the investigated experimental domain.