【摘要】 为了提高卷取温度的精度,采用BP神经网络方法并结合大量的现场数据,对热轧带钢层流冷却水冷数学模型中的综合换热系数因子进行预报,将预报结果应用于计算卷取温度的数学模型中,可将卷取温度的计算值控制在目标值的±15℃之间,大大提高了卷取温度的精度,具有在线应用的前景。更多还原

【Abstract】 By using BP neural network,the integrative exchange heat coefficient of mathematical model of hot rolled strip was predicted,and the result was applied to calculation of mathematical model of coiling temperature.The results indicated that the difference between the calculated temperature and target coiling temperature was controlled in the range between-15 ℃ and +15 ℃ after integrative exchange heat coefficient was predicted by BP neural network.The method obviously improves the accuracy of coiling temperature,so the on-line application of this method has a gseat future.更多还原