杂质元素对熔敷金属低温韧性的影响

用人工神经网络建立了熔敷金属性能的预测模型。该模型的学习结果同试验值之间有很好的对应关系。用该模型研究了杂质元素S、P、O、N对熔敷金属低温韧性的影响规律。预测结果表明：随着O、N含量的增加，熔敷金属的低温韧性显著下降；S对低温韧性基本无影响；随着P含量的增加低温韧性有所增加。在不同的S、P含量水平下，低O低N时的熔敷金属具有很高的低温韧性，高N低O时的次之，高N高O时的低温韧性最差。

EFFECTS OF IMPURITY ELEMENTS ON THE LOW TEMPERATURE TOUGHNESS OF DEPOSITED METALS

A mechanical property prediction model, in form of artificial neural network for the diposited metals was sat up on the experimental data. There are good correlations between the learning results and the experimental data. The effects of impurity elements(S, P, O, N) on the low temperature toughness of deposited metals were studied with this model. The predicted resultes show that the toughness decreased with the increasing O, N contents, and increased with P, and the S does not apparently affect the toughness. With differents S and P contents, the deposited metals with low O and Iow N has the highest toughness, higher with high N and low O contents, when high N and high O contents, the deposited metals has the lowest toughness.