基于BP神经网络遗传算法的薄板CMT点焊变形

焊接是汽车车身制造的一个关键环节，焊接质量的好坏严重影响汽车车身质量，所以焊接参数的选择至关重要. 针对薄板焊接质量控制问题，论文利用BP神经网络解决非线性问题的优势，建立焊接变形量与工艺参数之间映射关系模型；结合遗传算法构建基于遗传神经网络焊接的工艺参数优化系统；同时设计正交试验，将该方法与正交试验法相对比. 结果表明，该方法可以有效地实现CMT(cold metal transfer)点焊焊接变形预测与工艺参数优化. 通过预测模型给出合理参数，指导钢薄板和铝合金薄板的CMT点焊变形试验，提高焊接的效率.

CMT spot welding deformation of sheet metal based on BP neural network and genetic algorithm

Welding was a key link in automobile body manufacturing. The quality of welding seriously affected the quality of automobile body, so the selection of welding parameters was very important. Aiming at the quality control of thin plate welding, the advantage of BP neural network was used to solve the non-linear problem, and established the mapping model between welding deformation and process parameters. Combining with genetic algorithm, the optimization system of welding process parameters was constructed based on genetic neural network. Then the orthogonal test was designed and compared with the proposed model. The results showed that the method could effectively achieve welding deformation prediction and optimization of process parameter on CMT (cold metal transfer) spot. The reasonable parameters were given by the prediction model to guide the CMT spot welding deformation test of steel sheet and aluminium alloy sheet, and to improve the welding efficiency.