【作者】 谢禁； 蒋朝辉； 陈致蓬； 桂卫华； 严文莉；

【Author】 Jin Xie;Zhaohui Jiang;Zhipeng Chen;Weihua Gui;Wenli Yan;School of Information Science and Engineering,Central South University;

【机构】 中南大学信息科学与工程学院；

【摘要】 转炉炼钢终点的准确控制对保证钢水质量、减少生产返工、降低冶炼成本非常重要,然而因为在冶炼过程温度极高,难以在线及时检测,导致终点碳含量和温度一直是转炉冶炼最难控制的变量。为此,本文采用一种动态自调整的RBF神经网络建立了转炉钢水终点碳含量预报模型、静态控制模型,其中静态控制模型采取联合训练的方式,提高了模型精度;并结合两个模型,对转炉冶炼过程进行双阶段控制。现场数据验证了所建模型的有效性,且双阶段控制具有良好的控制效果,对实际生产控制有一定意义,提高生产效益。更多还原

【Abstract】 Accurate control of the end point of BOF steelmaking is very important to guarantee the quality of molten steel, reduce rework and cut smelting cost, however, because the temperature is very high in the smelting process, it is difficult to detect online in time, causing that the end-point carbon and temperature is the most difiicult variable to control. In this paper, a dynamic self-tuning RBF neural network is used to establish a prediction model of end-point carbon and a static control model for BOF. The static control model adopts the method of joint training, which improves the precision of the model, and a two-stage control of BOF smelting process is carried out with two models. The data verifies the validity of the model, and the two-stage control has good control effect, which is necessary to the actual production control and improves the production efficiency.更多还原