Optimization of process variables for additive manufactured PLA based tensile specimen using taguchi design and artificial neural network (ANN) technique

Abstract Fused filament fabrication (FFF) is one of the additive manufacturing processes which is used widely due to its ease of usage and low cost. The final quality of parts in the FFF process is determined by a careful selection of process parameters, as it is required to understanding the physical phenomena of process variables and their impact on mechanical properties. This study involves the independent analysis of five process variables i.e. printing nozzle temperature, layer thickness, raster orientation, speed/feed rate and infill pattern. Taguchi L27 orthogonal array opted to investigate the tensile strength of Polylactic acid (PLA) specimens. By using this technique, the number of experimental reduced from 243 to 27 experiments. These specimens are fabricated based on ASTM D-638 tensile standard design. Artificial Neural Network tool has opted by using MATLAB software for training and testing of data. Some notable observations based on our experimental investigations carried out for process parameter optimization was that for the least change, layer thickness of 200 μm, nozzle temperature of 210° C, speed/feed rate of 50 mm/min, grid as a structure/infill pattern and raster orientation of 0° are required.