综合神经网络在热连轧机组轧制压力预报中的应用

• 作 者: 王秀梅 王国栋 刘相
• 机 构: 华东北大学
• 出 处: 钢铁研究学报
• 发表日期: 1998年04期
• 关键词: 轧制压力 神经网络 热连轧机组 精轧机组 计算模型 数学模型 预报精度 误差反向传播 自适应技术 平坦区

由于传统的轧制压力计算模型结构简单，即使采用自适应技术，也难以适应不断提高的尺寸精度的要求。为了提高精轧机组轧制压力预设定精度，采用人工神经网络和传统数学模型相结合的综合神经网络方法。经离线仿真证明，该方法的预报精度优于传统方法，预报结果的相对误差基本限制在±５％以内。

Due to the simple structure of the traditional rolling pressure calculation model, even with adaptive technology, it is difficult to adapt to the ever-increasing requirements of dimensional accuracy. In order to improve the preset accuracy of the rolling pressure of the finishing mill, an artificial neural network (ANN) combined with the traditional mathematical model is applied to the neural network. The off-line simulation shows that the accuracy of this method is better than the traditional method, and the relative error of the prediction results is limited within ± 5%.