Advancing the structural performance of 3D-printed ABS materials for enhanced load-bearing capacity using TiO₂–ATO nanoparticle reinforcement and SSGCNN-based predictive modelling

Fused filament fabrication (FFF) is a widely used 3D printing technique, particularly for thermoplastics like acrylonitrile butadiene styrene (ABS). However, the mechanical performance of printed ABS components often falls short for advanced engineering applications. To overcome this constraint, this research suggests a targeted nanocomposite-enhancement approach to enhance the load-bearing and structural qualitiesof 3D-printed ABS materials. Novel nanocomposite filaments were developed by incorporating titanium dioxide (TiO₂) and antimony tin oxide (ATO) nanoparticles into ABS via melt-mixing. These filaments were then characterized and used for 3D printing. Experimental results showed a 27.4% increase in tensile strength, a 22.8% improvement in flexural strength, and a 19.6% enhancement in micro-hardness, in contrast to prints made entirely of ABS. Mechanical performance metrics were further predicted using a sparse spectra graph convolutional neural network (SSGCNN), trained on the 3D printer dataset for mechanical engineers (3D PDME), achieving a prediction error of just 0.035%, significantly outperforming benchmark models such as CNN (0.41%) and ANN (0.58%). These results highlight that integrating nanoparticle reinforcement with advanced deep learning prediction frameworks enhances the mechanical strength of ABS products that are FFF produced and makes precise performance forecasts possible, marking a substantial advancement over prior studies that focused solely on either material modification or predictive modelling.