基于机器学习的按需式电流体喷射打印微滴直径预测

Prediction of Droplet Diameter Based on Machine Learning for On-demand Current Jet Printing

为节省按需式电流体微滴喷射打印技术的微滴直径预测时间及解决众多工艺参数的合理选择问题,实现更高打印质量和效率,提出了数值仿真与机器学习算法相结合的方法。基于线性回归、支持向量回归、神经网络和随机森林算法建立8种参数与微滴直径的关系模型。算法结果表明:支持向量回归算法准确率最高、误差较小,随机森林算法和线性回归算法次之,神经网络算法准确率较低且误差较大。机器学习可以对按需式电流体微滴喷射打印微滴直径进行有效预测,此方法可以有效地提高设计效率。

In order to save the time of droplet diameter prediction of on-demand current droplet jet printing technology and solve the problem of reasonable selection of many process parameters,and achieve higher printing quality and efficiency,a method combining numerical simulation and machine learning algorithm is proposed.Based on linear regression,support vector regression,neural network and random forest algorithm,the relationship model between eight parameters and droplet diameter is established.The results show that the support vector regression algorithm has the highest accuracy and smaller error,followed by the random forest algorithm and linear regression algorithm,and the neural network algorithm has the lower accuracy and larger error.Machine learning can effectively predict the droplet diameter of on-demand current droplet jet printing.This method can effectively improve the design efficiency.