论文部分内容阅读
以合金型号、预热温度、链条速度、链条角度、焊接温度和焊接时间为输入参数,直通率为输出参数,构建6×18×9×1四层拓扑结构的神经网络优化模型,可以实现无铅焊接工艺参数的优化设计,且预测精度较高、预测能力较强。现场验证表明,与产线上原用无铅焊接工艺参数相比,采用神经网络优化模型设计的无铅焊接工艺参数,获得的无铅焊接焊点直通率从98.12%提高到100%;焊点内部均匀、致密,无气孔等缺陷。
Taking the alloy type, preheating temperature, chain speed, chain angle, welding temperature and welding time as the input parameters and the pass-through rate as the output parameters, a neural network optimization model of 6 × 18 × 9 × 1 four-layer topology is constructed, Lead welding process parameters of the optimal design, and the prediction accuracy, prediction ability. Field verification shows that compared with the original lead-free soldering process parameters in the production line, the lead-free soldering process parameters designed by the neural network optimization model improve the pass-through rate of lead-free solder soldering points from 98.12% to 100% Uniform, dense, non-porosity and other defects.