摘要
应用基于BP算法的神经网络能较好处理非线性问题的特点,建立起能描述工艺参数(化学成分、板材尺寸和相关生产工艺参数)和热轧带钢最终力学性能的映射关系的神经网络模型。该模型均具有较好的预报精度,预测值与实测值之间的相对误差小于±10%。
Taking the characteristics that BP neural network can solve non-liner problem well, a neural network which describes the mapping relation between processing parameters （chemical ingredient, sheet dimension and relevant production processing parameter） and final mechanical property of hot strip rolling steel. The model has good forecasting precision. The relative error of prediction value and actual value is less than±10%.
出处
《模具技术》
2005年第6期52-54,共3页
Die and Mould Technology
关键词
神经网络
热轧带钢
力学性能
BP neural network
hot strip rolling steel
mechanical properties
