MODELING THE STRENGTH-TO-MASS RATIO OF THE FUSED FILAMENT FABRICATED ABS POLYMER WITH ANN

The present study investigated the Strength-to-Mass ratio (StMr) yielded by Acrylonitrile Butadiene Styrene (ABS) specimens, fabricated with the Fused Filament Fabrication (FFF) method. A universal testing machine was utilized for the Ultimate-Tensile-Strength (UTS) measurement of the specimens, whereas a precision balance was used to measure their mass. The experiments were designed according to the Central-Composite-Design (CCD), by considering four process parameters: the infill, the layer thickness, the line direction of the top and bottom layers, as well as the pattern. In addition, a shallow Artificial Neural Network (ANN) was developed to predict the StMr, which was then compared to the empirical model generated by the CCD method. The analysis revealed a strong correlation between the two models, with the Mean Absolute Percentage Error (MAPE) being below 1%. Finally, verification testing was performed to evaluate the absolute error handling of the ANN model, which was found to be 6%.