利用性能预报模型对SS400热轧板力学性能实施预报

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％7g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％。

To forecast the mechanical property of the hot rolled SS400 strip with regression model and artificial neural network

In order to predict the mechanical properties of the hot rolled SS400 strip of No.4 Steelmaking Rolling Mill, the regression model and BP algorithm of ANN were used to reflect the influence of the chemical composition and the finish rolled thickness on the mechanical properties of product and the best predict model was selected according to the current condition. In actual use, It is found that regression model is the best .There are 80% percent of the tensile strength forecasting values was only 5% percent errors more or less than the actual ones ;and 76% percent of the yield strength forecasting numbers was 10% percent errors more or less than the actual ones; 77% percent of the elongation forecasting numbers was 10% percent errors more or less than the actual ones.