摘要
在焊接工艺的研究过程中,对熔池的成型形态和熔池内部液态金属的瞬时运动状态进行分析是很有必要的,因此需要对抓拍焊接熔池的工业相机选择合适的视场。建立了3-5-1结构的BP神经网络的焊接熔池照度模型。以MAG焊接工艺参数保护气流量、焊接电流和接收弧光点到焊接电弧的距离作为网络输入,接收弧光点的光照强度作为网络的输出,优化后的GA+BP神经网络模型能够对焊接熔池照度进行准确的预测。
It is necessary to analyze the shape of molten pool and the instantaneous movement state of molten metal in molten pool during the research of welding technology,Therefore,it is necessary to select suitable view field of the welding pool industrial camera.The illuminance model of weld pool of GA+BP neural network with 3-5-1 structure was established.Shielding gas flow,welding current and the distance from receiving arc point to welding arc in MAG welding process were used as network input,the light intensity of receiving arc point was used as the output of the network,the optimized GA+BP neural network model can accurately predict the illumination of welding pool.
作者
关子奇
朱玉龙
刘晓光
刘丹
常云龙
GUAN Ziqi;ZHU Yulong;LIU Xiaoguang;LIU Dan;CHANG Yunlong(School of Material Science and Engineering,Shenyang University of Technology,Shenyang 110870,China;Guangdong institute of Intelligent Manufacturing,Guangzhou 510070,China;School of Automation,Guangdong University of Technology,Guangzhou 510006,China)
出处
《热加工工艺》
北大核心
2019年第7期216-220,223,共6页
Hot Working Technology
基金
国家自然科学基金项目(51575362)
沈阳市"双百"项目(Z17-5-070)
广东省高级人才引进项目资助(2016GDASRC-0106)
关键词
焊接熔池
遗传算法
BP神经网络
照度
优化预测
welding molten pool
genetic algorithm
BP neural network
illumination
optimal prediction
