材料期刊网

在中厚板生产过程中，轧后冷却的温度控制是决定产品组织性能的关键工艺技术。换热系数是温度控制模型的核心参数，由于其影响因素多且复杂，故很难有一个固定的模型来计算。提出一种简单有效的换热系数的自学习模型：先通过参数节点化和插值法得出全范围内的各换热系数影响因素的特征值；再基于k-NN原理，寻找待冷目标钢板与各已冷样本钢板之间的相似度；最后，通过IDW加权平均算法，预估出目标冷却钢板所需的换热系数。现场实际的应用情况表明，使用该自学习模型相比以往可提高控制精度约5%。

Temperature control of cooling after rolling is a key technology for determining products' properties during hot plate rolling. Furthermore,the coefficient of heat transfer is a core parameter in temperature control process. It can-not be determined by a fixed mode because there are many complex influence factors. A simple and valid self-learning model is proposed to determine the coefficient of heat transfer. Firstly,characteristic value of influence factors in each range of the coefficient of heat transfer is obtained by using node interpolation. Secondly,the correlationship between target and each cooled plate is found by k-NN method. Finally,the coefficient of heat transfer of target plate is estimat-ed through IDW weight average algorithm. The results of applications with real industrial data show the proposed model can increase temperature control precision by about 5%compared with the past.