论文部分内容阅读
采用物理冶金模型结合二维温度场对ASP(Angang Strip Production)热轧X70管线钢再结晶、相变等物理冶金过程进行了模拟,并结合BP神经网络对最终的力学性能进行了预测。研究表明,实验钢在层流冷却前的奥氏体晶粒尺寸为10～25μm,板带横断面奥氏体晶粒尺寸分布不均匀,心部的奥氏体晶粒尺寸比角部大15μm左右;在给定冷却速率的情况下采用前段冷却方式得到的铁素体分数比后段冷却方式大2%～5%;采用BP神经网络可以把伸长率预测结果相对误差标准差提高1.8%;Si含量0.2%～0.3%成为其对力学性能影响的转折点。
Physical metallurgy model and two-dimensional temperature field were used to simulate the metallurgical processes such as recrystallization and transformation of hot rolled X70 pipeline steel produced by ASP (Angang Strip Production). The final mechanical properties were predicted by BP neural network. The results show that the austenite grain size of the experimental steel before laminar cooling is 10 ~ 25μm, the distribution of austenite grain size in the cross section of the strip is not uniform, and the austenite grain size in the core is 15μm larger than the corner ; When the cooling rate is given, the ferrite fraction obtained by the former cooling method is 2% ~ 5% larger than the latter cooling method; the relative standard deviation of the prediction of elongation can be increased by 1.8% by BP neural network; ; Si content of 0.2% to 0.3% as a turning point on its mechanical properties.