머신러닝 기반 고강도 아연도금강판 CMT 용접부의 품질기준 예측을 위한 공정변수 최적화

Abstract

Recently, the adaption of the GA steel which has high corrosion resistance is increasing to the automobile. But, in case of th GA steel, the differences between melting point of the steel and boiling point of the zinc leads to defect such as the porosity. It causes not only the decreasing the strength of the joint, also decreasing the productivity. Thus, it must be solved to improve the productivity.

This study is to investigate the effect of process parameters on possibility to effectively reduce porosity in welded joints under different heat inputs ranging from 155 to 458J /mm by CMT process. Cold metal transfer welding that is automated GMAW process related to short circuit transfer has been used to join galvanized steel sheets in a lap joint without a gap. Moreover, machine learning has been performed to predict the quality factor and this result is compared with that of shallow neural network.

Therefore, this study involves an investigation of the optimization of process parameters based on machine learning for predicting quality factor in high strength GA steel welded joints by CMT.