论文部分内容阅读
通过正交试验获取样本数据,利用MATLAB的工具箱函数建立RBF神经网络预测模型,研究了稀土氧化物对焊缝低温韧性的影响和作用程度。结果表明,通过正交试验样本训练的RBF神经网络较好地反映稀土氧化物与焊缝低温韧性之间的非线性关系,可用于焊缝低温韧性预测;在焊条药皮中的氧化镧和氧化铈添加量较高,而氧化钇添加量较低时其焊缝低温韧性相对较好,当添加0.7%氧化镧、1%氧化铈和0.3%氧化钇的焊条其焊缝低温韧度值达到98 J/cm~2。
The sample data were obtained through orthogonal test. The RBF neural network prediction model was established by MATLAB toolbox function. The influence of rare earth oxide on the low temperature toughness and the degree of action of the weld were also studied. The results show that the RBF neural network trained by orthogonal test samples can well reflect the nonlinear relationship between the low temperature toughness of rare earth oxides and the weld, which can be used to predict the low temperature toughness of the weld. The lanthanum oxide and oxidized When the addition amount of yttrium oxide is lower, the low temperature toughness of the weld is better. When the addition of 0.7% lanthanum oxide, 1% cerium oxide and 0.3% yttrium oxide, the low temperature toughness of the weld reaches 98 J / cm ~ 2.