基于BP神经网络的锆合金电子束焊接熔深控制模型  

Depth control model of molten zirconium alloy electron beam welding based on BP neural network

作　　者：李博[1] 陈路[1] 朱天军[1] 王世忠[1] 

机构地区：[1]中国核动力研究设计院 反应堆燃料与陈路国家重点实验室,四川成都610041

出　　处：《电焊机》2013年第12期64-67,共4页Electric Welding Machine

摘　　要：以锆4合金电子束焊接熔深为研究对象,用BP神经网络的方法建立电子束焊接工艺参数与熔深之间的数学模型;本研究采用MATLAB语言实现BP算法,其程序设计主要是调用MATLAB神经网络工具箱的相关函数来完成。结果表明,所建BP神经网络模型具有较好的自适应和预测能力。This paper presents the application of software package MATLAB to establish model of the welding bead depth of electron beam welding Zirconium alloy based on BP neural networks.The model was trained with actual samples obtained from high power electron beam welding Zirconium alloy.The simulation results indicated that the model has strong abilities on nonlinear adaptivity and predicting.The model not only possesses the ability to estimate the welding bead depth by the parameters,but also can optimize the welding technics according to predicted results.

关 键 词：锆4合金 电子束焊接 BP神经网络 

分 类 号：TG456.3[金属学及工艺—焊接]