摘要
采用BP神经网络与数学模型相结合的方法对热带精轧机组机架间水冷区带钢热流密度进行预测,进而优化了机架间冷却的数学模型。结果表明,利用BP神经网络得出的带钢热流密度计算的终轧温度与实测值的标准差比原来仅用数学模型的传统算法减少了14 08%,故该方法具有较好的在线应用前景。
The hot strip heat-flow density in interstand water cooling zones of finishing mill was predicted by mathematical model combined with BP neural network for optimization of the model of interstand cooling. The standard deviation between the predicted and measured finishing temperature was reduced by 14.08 %. The method can be used for on-line prediction.
出处
《钢铁研究学报》
CAS
CSCD
北大核心
2004年第3期75-78,共4页
Journal of Iron and Steel Research
基金
国家自然科学基金资助项目(59995440)
关键词
机架间冷却
BP神经网络
数学模型
热流密度
终轧温度
热轧
带钢
interstand cooling
BP neural network
mathematical model
heat-flow density
finishing temperature
hot rolling
strip
