热连轧精轧带钢厚度预报模型优化研究

Optimization of Prediction Model of Thickness in Hot Continuous Precise Rolling Strip Steel

为了提高某9机架热连轧机组厚度精度,依据轧制基本理论,优化修正了厚控模型。采用指数平滑自学习方法,通过自学习综合冷却系数优化了温度模型;采用变形抗力估计法,利用工艺参数建立了更为准确的变形抗力计算模型;通过对传统Hitchcock轧辊压扁半径与轧制力关系的研究,提出了轧辊压扁半径的显示算法,此方法计算过程不需迭代,计算精度高且响应速度快;在线控制生产结果表明,新模型能提高厚度预测精度3.3%。

In order to enhance thickness precision of some 9 hot continuous rolling strip mill,according to the basic theory of rolling, the model of thickness control was improved. Adopting the exponential smoothing estimation method, the temperature model was optimized by the self-learning of the comprehensive cooling coefficient; Adopting theory of the deformation resistance estimating method, the more accurate computational model of the deformation resistance was established by using the technological parameters; through study of the relationship between the flattened radius of roller for traditional Hitchcock and rolling force ,the explicit algorithm of the flattened radius of roller was put forward, which is not iteration in the calculation process. The computational accuracy was high and response speed was fast. Production results in online control show that the new model can improve prediction accuracy 3.3% for thickness estimation.