Prediction of the mechanical properties of heat-treated fused filament fabrication thermoplastics using adaptive neuro-fuzzy inference system

Abstract

The requirement for post-processing methods in 3D printing has increased due to its major limitations such as poor mechanical and surface properties. Thermal annealing, a heat-treatment procedure, has proven to be an excellent approach for increasing the mechanical strength of additive manufactured thermoplastics. The optimized thermal annealing parameters for maximum tensile strength are identified. The Adaptive Neuro-fuzzy Inference System (ANFIS) methodology is employed in this study to forecast the tensile strength of thermal annealed fused filament fabricated polylactic acid (PLA), carbon fiber filled polylactic acid (PLA-CF), and polycarbonate acrylonitrile butadiene styrene (PC-ABS). The root mean square error (RMSE) of the predicted tensile strength of PLA, PLA-CF, and PC-ABS are obtained as 1.012, 0.5, and 0.835 respectively. The coefficient of determination for the predicted model of thermoplastics PLA, PLA-CF, and PC-ABS is determined as 0.98, 0.99, and 0.89 respectively. The ANFIS model developed is successful in determining the tensile strength of annealed 3D printed thermoplastics at various annealing conditions of temperature and duration. Further, the study investigates the effect of heat treatment on the compressive, impact, and flexural properties of PLA prints.