论文部分内容阅读
以现场收集的四钢轧 SS400热轧板的原始化学成分、终轧厚度、实测的力学性能数据为基础,通过回归模型和人工神经网络 BP 算法建模,确定其相互关系,并最终通过其化学成分和终轧厚度来预测产品力学性能。现场使用证明,在现有的条件下,回归模型比人工神经网络更适用。经测试,其抗拉强度预报值与实测值的相对误差有80%不超过5%,屈服强度预报值与实测值的相对误差有76%不超过10%,延伸率预报值与实测值的相对误差有77%不超过10%。
On the basis of the original chemical composition, the finish rolling thickness and the measured mechanical properties data of the SS400 hot-rolled sheet collected from the field, the regression model and the artificial neural network BP algorithm were used to establish the correlation between the chemical composition and the final rolling thickness through the chemical Composition and finish rolling thickness to predict product mechanical properties. Field use proved that under the existing conditions, the regression model is more suitable than artificial neural network. After testing, the relative error between the predicted value of the tensile strength and the measured value is less than 5%, the relative error between the predicted value of the yield strength and the measured value is not more than 10%, and the relative value between the predicted value of the elongation and the measured value is relatively The error is 77% less than 10%.