基于信息融合算法的LF炉钢水温度预测

Temperature Prediction of Molten Steel for LF Based on Information Fusion Technique

LF炉钢水温度的控制对钢的质量和连铸操作的顺行都很重要,而LF炉钢水温度的预报是LF炉钢水温度控制的前提。针对LF炉冶炼过程中物理化学反应过程及传热过程的复杂性,以宝山钢铁股份有限公司300 tLF炉为研究对象,在分析了影响LF炉钢水温度的主要因素的基础上,应用基于BP神经网络的信息融合算法,开发了用C语言编写的预测程序,预测了LF炉的钢水温度。实验表明,此算法可以提高预测的速度和精度,预测结果为误差不大于±5℃的炉次大于90%。

It is very important to control the temperature of molten steel during ladle furnace （LF） refining. The prediction of temperature on LF is precondition for temperature control. Due to the complexity of physico-chemical reactions and heat transfer in LF refining, the main influencing factors of temperature were analyzed, based on the data Baoshan Iron and Steel collected on 300 t LF at the temperature was predicted by new algorithms——information fusion based on improved back propagation network （BP）. C program for predicting the temperature was dcveloped. Experiment results indicated that the new algorithms can improve the precision of temperature predic tion, and the hitting rate can be more than 90% at an error less than ±5 ℃.