基于L-M算法BP神经网络的转炉炼钢终点磷含量预报

来源期刊：钢铁2011年第4期

论文作者：李长荣 赵浩文 谢祥 尹青

文章页码：23 - 25

关键词：BP神经网络； 终点磷含量； Levenberg-Marquardt算法； 预报模型

Key words：BP neural network； phosphorus content of end-point； Levenberg-Marquardt(LM) algorithm； predictive model

摘    要：转炉炼钢过程是一个非常复杂的物理化学变化过程,人工控制很难一次达到终点目标值,通常需要经过多次补吹才能出钢。通过研究影响转炉冶炼终点磷含量的主要因素,确定了影响转炉终点磷含量的参数,建立了基于Levenberg-Marquardt(LM)算法BP神经网络转炉终点磷含量的预报模型。结果表明:在预报误差目标精度为±0.002%内,命中率达到了90%。

Abstract: BOF steelmaking is a very complex physical chemistry process;it is hard to achieve the target value of end-point by manual control.Multiple reblowing operations were usually necessary to taping off.Based on analyzing the influence major factors of phosphorus end-point in converter,the dominative factors of prediction model of end-point for Conrerter smelting were fixed.A prediction model of end-point phosphorus content for BOF process is established based on Levenberg-Marquardt(LM) algorithm of BP neural network.The results show that the phosphorus content of end-point hitting rates could be reached 90% if the accuracy of target error were ±0.002%.