摘要
针对智能化焊接中的熔池质量控制环节,采用端到端的集成学习模型,提出一种更加有效的实时预测方法。基于集成学习算法GBDT建立起焊接参数、熔池尺寸与焊缝背部熔宽之间的预测模型。训练数据集由GTAW焊接仿真数据以及焊接实验数据组成。该模型有效提高了焊接过程中背面熔宽的预测准确性,进而为智能化焊接提供更有效的熔池质量控制。
Using the end-to-end ensemble learning model, a more effective real-time prediction method was proposed for the quality control of molten pool in intelligent welding. Based on the ensemble learning algorithm GBDT, the prediction model of welding parameters, molten pool size and weld backside molten pool width was established. The training data set consists of GTAW welding simulation data and welding experiment data. The model can effectively improve the prediction accuracy of backside molten pool width in the process of welding, and then provide more effective molten pool quality control for intelligent welding.
作者
石运良
罗宇
陈正科
SHI Yunliang;LUO Yu;CHEN Zhengke(School of Naval Architecture,Ocean and Civil Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《热加工工艺》
北大核心
2021年第17期110-114,共5页
Hot Working Technology
关键词
焊接智能制造
GBDT算法
焊缝背面熔宽预测
intelligent welding
GBDT algorithm
prediction of weld backside molten pool width
