【摘要】 丝网印刷工艺参数对丝网印刷的质量有着极为重要的影响,各工艺参数与印刷质量之间并不是简单的线性关系。深度神经网络是具有非常强学习能力的非线性系统,因此提出了一种利用深度神经网络预测丝网印刷质量的方法,建立全连接构造的回归模型,将浆料型号、印刷速度、印刷压力、离网距离等4项参数作为神经网络的输入,输出的预测值为印刷的质量。根据训练得到的预测模型调整印刷机参数的设置,有效提高了印刷机的印刷质量。更多还原

【Abstract】 Screen printing process parameters have a very important impact on the quality of screen printing,the relationship between process parameters and printing quality is not simple linear.Deep neural network is a nonlinear system with strong learning ability.Therefore,a method to predict the quality of screen printing using deep neural network is proposed,establishes a regression model of full connection structure is established.Four parameters such as solder paste type,printing speed,printing pressure and distance from the stencil are taken as the input of neural network,and the output prediction value is the quality of printing.According to the trained prediction model,the setting of printing machine parameters is adjusted,and the printing quality of the printing machine is effectively improved.更多还原