摘要
AOD炉冶炼终点成分和温度是冶炼的控制目标,与多个变量存在非线性关系且无法在线连续测量。利用RBF神经网络建立终点温度和碳含量的预报模型,以冷却剂加入量最少和终点温度与碳含量同时到达为目标,克服传统终点控制借助经验来判断并不准确的缺点,提高终点的命中率,缩短冶炼时间。
The overall control objects of the AOD ferroalloy refining are end-point component and temperature which have non-linear relationship with many parameters and are difficult measured on-line.Here we establish a model to predict the end temperature and carbon content based on the RBF neural network.With the minimum contend of coolant we can make the end point temperature and carbon content get the fixed value at same time to overcome the shortcomings of the traditional end-point control.Our method can improve the efficiency and reduce the refining time.
出处
《长春工业大学学报》
CAS
2011年第4期370-374,共5页
Journal of Changchun University of Technology
基金
吉林省科技发展计划基金资助项目(20090402)