| 3D printing technology is widely used in various fields because it efficiently produces complex parts.With the continuous development of industrial design,higher requirements have been put forward in terms of the performance,function,and application of parts.Multi-color and multi-material 3D printing can add multiple attributes to the model,such as color,strength,modulus,etc.,thus creating more complex parts with various performance attributes and further enhancing product functionality.However,current 3D printers mainly manufacture single-color and single-material parts.Fused deposition modeling(FDM)3D printing is popular due to its low cost and ease of operation.Therefore,developing a consumer-grade FDM multi-color and multi-material 3D printer is of great significance.Multi-color 3D printing is an essential form of multi-material 3D printing.This study takes multi-color 3D printing as the starting point of research,establishes relevant theoretical foundations,and further explores the co-extrusion of multi-material 3D printing in depth.The main research content is as follows:(1)The mechanical structure design of the printer was carried out,including the printer frame,XYZ axis movement mechanism,and extruder feeding mechanism.The physical assembly of the printer was completed.A printhead that can mix three filaments was designed and processed on this basis.To solve the problem of difficult disassembly and assembly of the printhead,a universal quick-disassembly device was designed.The printer control system was designed by selecting a suitable control board and control strategy,calculating the pulse number of each motion mechanism motor,compiling and burning the firmware program for the control system,and realizing multi-material mixing printing.(2)To address the issue of printhead clogging,temperature field simulations were performed on the printhead using ANSYS Fluent.The influence of the throat material on the heat dissipation performance of the printhead was investigated,and the results showed that the throat inlet temperature was less than the polylactic acid(PLA)filament glass transition temperature(Tg)only with the polytetrafluoroethylene(PTFE)throat and under the use of the cooling fan.On this basis,the effect of cooling fan speed on the heat dissipation performance of the printhead was investigated,and simulation results indicated that when the fan speed exceeded 1.5 m/s,the temperature at the throat inlet was below Tg,and further increasing the fan speed did not significantly shift the Tg position.The optimal fan speed was 1.5 m/s.The method of multi-material mixing printing was studied.By setting up a virtual extruder in both firmware and software,full-color and gradient color printing were tested and achieved,demonstrating the feasibility of the multi-material mixing scheme.To solve the problem of uneven mixing of filaments,three different structures of mixers were designed and tested by printing.The results showed that a two-stage reverse static mixer could improve mixing uniformity.(3)A total of 67 different blends of PLA,TPU,and PETG were prepared using a mixing 3D printing method,and their tensile properties,flexural properties,and hardness were tested.An artificial neural network(ANN)optimization algorithm was used to build a component-performance prediction model.The data samples were normalized,and the neural layers were L2 regularized to solve the overfitting problem of the artificial neural network.Using Nadam,Adam,and Adamax optimization algorithms with different learning rates,the tensile strength,tensile modulus,flexural strength,flexural modulus,and hardness of the blend were trained to obtain the best model with R~2 values of 0.9263,0.9924,0.9899,0.9971,and 0.9888,respectively.A mechanical performance palette was established using the best prediction model,and the accuracy of the prediction model was verified through experiments.The results showed that using ANN to predict the mechanical properties of the blend is feasible,and a multi-material blending 3D printer can quickly manufacture printed parts that match the target properties under the guidance of the algorithm. |
