摘要
目的节省电流体喷射打印精度预测的时间和解决电流体工艺参数的选择问题,达到提高电流体打印的质量和效率的目的。方法为了对电流体喷射打印精度进行预测,提出有限元模型与机器学习相结合的方法。基于线性回归、支持向量回归和神经网络等机器学习算法建立4种参数与射流直径的关系模型。结果算法结果表明:支持向量回归和神经网络预测模型的决定系数R2能达到0.9以上,表示模型可信度高;支持向量回归和神经网络预测模型指标都比线性回归预测模型的小。结论机器学习算法可对电喷印打印精度进行有效预测,预测效率提高了十几倍,节省了精度预测的时间。
The work aims to save time in predicting the accuracy of E-jet printing,solve the problems in selection of electrofluidic process parameters,and improve the design quality and efficiency of electrofluidic printing.A combination of finite element models and machine learning was proposed to predict the accuracy of E-jet printing.Based on machine learning algorithms such as linear regression,support vector regression and neural networks,a model on relationship be-tween four parameters and jet diameter was established.The algorithm results showed that the determination coefficient R2 of the support vector regression and neural network prediction models could reach above 0.9,indicating that the mod-els were highly credible;RMSE and MAE,which were indicators of model error,were both smaller than those of the lin-ear regression prediction models.Machine learning algorithms enable effective prediction of E-jet printing accuracy,in-creasing prediction efficiency by more than a factor of ten and saving time on accuracy prediction.
作者
杨静文
陈小勇
张军华
YANG Jing-wen;CHEN Xiao-yong;ZHANG Jun-hua(College of Mechanical&Electrical Engineering,Gulin University of Electronic Technology,Guangxi Guilin 541004,China)
出处
《包装工程》
CAS
北大核心
2022年第13期203-208,共6页
Packaging Engineering
基金
广西自然科学基金(22GXNSFAA035616)
广西制造系统和先进制造技术重点实验室基金(2006540007Z)。
关键词
机器学习
电流体微纳打印
射流精度
预测模型
machine learning
electrofluidic micro-nano printing
jet accuracy
prediction model