Abstract: Method,effect and function relation of static model and dynamic model for controlling end-point of converter steelmaking have been analyzed. According to the handicapped present situation of sensor for medium and small size converters,the comparison between optimal static model based on established GA-BP hybrid algorithm for predicting end-point of converter steelmaking and quasi-dynamic control model for predicting end-point of converter steelmaking has been compared from modeling principle,methodology and application effect. It is believed that prediction effect of a static model may be improved with GA-BP hybrid algorithm and the quasi-dynamic control model can reflect dynamic operation and the influence of operation in melting and charging process on first turndown time and end-point of converter steelmaking.

Key words: Converter Steelmaking,End-point Controlling,BP Neural Network,Cenetic Algorithm

摘要： 分析了转炉炼钢静态控制与动态控制模型的方法、作用及功能关系。针对中小型炼钢转炉检测设备受限制的现状,从建模原理、方法特点和应用效果方面对建立的基于遗传算法-神经网络混合算法(CA-BP)的转炉炼钢终点优化控制静态模型和转炉炼钢终点预测的准动态控制模型进行了比较,认为混合算法可以提高静态模型的预测效果,而准动态模型可以反映转炉冶炼的动态操作和加料过程对一次倒炉时间和终点的影响。

关键词: 转炉炼钢, 终点控制, BP人工神经网络, 遗传算法