基于GA—NN人体心血管模型FDM制造的工艺参数优化

Research on Rapid Prototyping Technology of human cardiovascular model

基于熔融沉积成型(Fused Deposition Modeling,FDM)的3D打印技术而制作的人体心血管模型,是批量制作透明人体心血管模型的母模,其制作步骤是:利用三维激光扫描技术和逆向工程软件来反求重构模型;将模型切片后导入到桌面FDM设备采用默认参数打印出实体模型。针对默认工艺参数下模型出现的破损缺陷,采用参数优化的方式改善缺陷。以心血管模型破损面积(mm2)为目标,利用正交试验数据建立了反应其与分层厚度(d/mm),喷头温度(T/℃)和填充速率(C/mm/s)之间关系的神经网络(Neural Network,NN)模型。采用遗传算法(Genetic Algorithm,GA)方法对上述工艺参数,进行优化并得到相应的制造参数。通过实验验证,结果表明:采用优化后的工艺参数,模型的预期破损面积(0.021mm2)与实验结果(0.023mm2)相符合,并且小于之前的试验结果。故采用GA-NN对人体心血管模型FDM制造进行优化是可行的。

Human cardiovascular model produced by 3D printing technology based on FDM is the master die of the mass-produced transparent human cardiovascular model. It＇s production process is： to achieve reverse reconstruction of cardiovascular model by using 3D laser scanning technology and reverse engineering software; The model after being sliced is introduced into the desktop FDM device to print out the physical model by using the default parameters. To adjust the damage defects of the model under the default process parameters,printing parameters are optimized. Set the cardiovascular model damaged area（ mm^2） as the goal,neural network（ Neural Network,NN） model is set up which can reflect the relationships among damaged area,the layer thickness（ D/mm）,nozzle temperature（ T/℃） and filling rate（ C/mm/s） based on orthogonal experiment data. Genetic algorithm（ Genetic Algorithm,GA） method is used to optimize the parameters above,and the corresponding manufacturing parameters are obtained. Finally,the experimental results show that the expected rupture area（ 0.021mm^2） of the model is consistent with the experimental results（ 0.023mm^2）,and less than the previous test results. Therefore,it is feasible to optimize the FDM manufacturing of human cardiovascular model by GA-NN.