基于人工神经网络含稀土元素熔敷金属力学性能预测

发布时间：2018-02-11 19:23

本文关键词： La元素焊接速度 BP神经网络 RBF神经网络预测模型熔敷金属力学性能焊条　出处：《沈阳工业大学学报》2017年03期 　论文类型：期刊论文

【摘要】：为了提高焊条的力学性能并缩短焊条研发周期,在E4301型焊条药皮配方基础上加入了CeO_2和稀土元素La,并对焊条进行了力学性能试验.对试验数据进行分析后发现,加入适量的稀土元素可以改善焊条的力学性能.利用典型BP和RBF神经网络分别建立力学性能预测模型.将焊条中的CeO_2、La、Si、Mn含量与焊接速度作为预测模型的输入变量,将熔敷金属的抗拉强度、下屈服强度、断后伸长率与热影响区平均硬度作为输出变量.结果表明,将BP和RBF神经网络用于对含稀土焊条力学性能的预测是可行的,且RBF神经网络模型的预测精度和效率要高于BP神经网络模型.
[Abstract]:In order to improve the mechanical properties of the electrode and shorten the research and development cycle of the electrode, CeO_2 and rare earth element La were added on the basis of the flux coating formula of the E4301 electrode, and the mechanical properties of the electrode were tested. After the analysis of the test data, it was found that, The mechanical properties of the electrode can be improved by adding proper amount of rare earth elements. The prediction models of mechanical properties are established by using typical BP and RBF neural networks. The content of CEO _ 2O _ 2La _ 2O _ Si _ 2O _ mn and the welding speed in the electrode are taken as the input variables of the prediction model. The tensile strength, lower yield strength, elongation after break and average hardness of heat-affected zone of deposited metal are taken as output variables. The results show that BP and RBF neural networks are feasible to predict the mechanical properties of rare earth electrode. The prediction accuracy and efficiency of RBF neural network model are higher than that of BP neural network model.
【作者单位】：江苏师范大学机电工程学院;大庆油田管理局大庆钻探工程公司钻井生产技术服务二公司;
【基金】：国家自然科学基金资助项目(51475220) 江苏省产学研前瞻性联合研究项目(BY2016028-02) 徐州市科技计划资助项目(KC15SM031)
【分类号】：TG407

【参考文献】