Pressure control in material extrusion additive manufacturing

This paper discusses the effect called undesirable extrusion dynamics that occur during material printing or fused deposition modeling. is revealed high acceleration and deceleration of head results in over-extrusion on corners parts printed layer unevenness. To minimize this effect, modern control systems use Advance algorithms for control. However, such disadvantages as inertia, reduced speed, need manual calibration, well uncertainty influence material, nozzle geometry, process parameters algorithm performance do not allow group solutions to be applied an industrial scale. The presents a study characteristics, type, extruder temperature, through series experiments. experiments were carried out Creality Ender 3 printer. obtained experimental data will us deepen understanding FDM printing; they are also used present modify existing algorithms. proposes modification based machine learning. It proposed implement containing two trained neural network models. One model predicts changes motion residual defects increase average print speed. second compensation parameter specific conditions without calibration. A was determine depending type thickness, geometry temperature. data. developed introduced into Linear Marlin firmware tested 3. showed can successfully predict required process. helps facilitate automate which expand possibilities printers’ application conditions. improve accuracy speed subsequently help economic independence competitiveness small design organizations enterprises Russia 3D printers. research expands rapid prototyping may ensure creation pilot batches.