【摘要】 针对LD封装激光焊接后变形产生的耦合偏差,本文首先设计了不同参数对焊后偏移(PWS)影响程度的实验,通过参数间排列组合的变化观测其对PWS的影响程度;进而设计出神经网络焊后偏移预测模型,并将预偏移大小、焊点位置、激光脉冲能量以及焊接前初始位置等可以实时调节的焊接参数作为神经网络输入,模型的PWS预测值与实际值吻合效果良好。实验结果表明,利用神经网络预测PWS的方法是可行的,并且在实际焊接过程中,可以借助神经网络预测模型选择最佳的焊接初始条件使PWS降至最低,大大提高了LD的封装效率和产品可靠性。更多还原

【Abstract】 Post welding shift(PWS) is a kind of welding deformation caused by residual stress,which will induce the coupled fiber out of the original position.How to minimize and compensate the PWS is always a bottleneck for the automatic laser diode(LD) packaging.In accordance with the experimental results and the previous researches,this paper analyzes the influence of different welding parameters.There are four parameters that can be controlled during the whole welding process,namely the pre-welding shift,position of welding spots,laser power,and position of the ferrule.Using these adjustable parameters as the input,a neural network based PWS prediction model is constructed.By neural network training,the predicted PWS shows good results in accordance with the experiments.The research achievement of this paper will help to minimize and compensate the PWS and improve the efficiency of automatic laser diode packaging.更多还原