利用神经网络预测热轧板带力学性能  

PREDICTING MECHANICAL PROPERTIES OF HOT ROLLED STRIP BY USING NEURAL NETWORKS

作　　者：王殿辉[1] 刘振宇[1] 王国栋[1] 张强[1] 姜洪生[1] 王洪水[1] 

机构地区：[1]东北大学,宝山钢铁(集团)公司

出　　处：《钢铁》1995年第1期28-31,共4页Iron and Steel

摘　　要：用神经网络预测热轧板带产品的力学性能可显著地改进回归数学模型的不足，并可充分地反映热轧生产主要工艺参数对产品力学性能的影响规律。为预测和控制热轧板带产品的性能质量提供了行之有效的手段。The application of Neural Networks to predicting mechanical propries of hot - rolledstrip can not only improve the predicting precisions comparing with the traditional regressionmodels,but also reflect the regularities of the product mechanical properties with hot rollingparameters.The effective way for predicting and controlling the mechanical properties of hot rolledstrip has been suggested.

关 键 词：神经网络 力学性能 热轧板带 轧制 板材 带材 

分 类 号：TG335.5[金属学及工艺—金属压力加工]