摘要
针对中小型转炉不宜增设副枪、难以对钢水成分和温度进行连续检测、难以建立动态模型的实际情况,本文将传统增量模型和神经网络模型有机结合,提出了一种基于增量式神经网络的转炉静态控制模型,对钢水终点进行控制。在该模型引入了RBF神经网络对钢水终点温度和碳含量进行实时预报,使得对增量式神经网络控制模型的训练以预报模型的输出值与所要求的钢水终点温度和碳含量之差为最小,克服了常规静态控制模型存在的不足,改善了控制效果,提高了炼钢一倒命中率。
A BOF static model based on the incremental neural network is proposed to control the steel endpoint. The RBF neural network is introduced into the static model to predict the steel endpoint temperature and carbon content timely, so that the incremental neural network control model can be trained based on the minimum difference of the temperature and carbon content between the output of the predicted model and the required steel endpoint.
出处
《自动化技术与应用》
2005年第5期17-19,共3页
Techniques of Automation and Applications
关键词
转炉
增量模型
神经网络
预报模型
Converter
Incremental model
Neural network
Prediction model
