真空精炼过程的钢水温度预测模型分析

Prediction and Analysis of Molten Steel Temperature for Refining Process

通过分析RH-MFB精炼过程中钢水温度和钢水传热规律,基于人工神经网络、模糊逻辑建模方法,建立了针对RH-MFB精炼复杂过程的温度预测模型。通过对比15炉300t钢水温度应用模型计算值与实际测量值,结果发现最大温差为7.2℃,最小温差为0.3℃,平均温差为4.52℃,两者误差在±5℃内的数据占到82.7%。另外根据实际要求,在RH-MFB精炼过程中通过添加铝、冷却材料实现对钢水温度的补偿。

Heat transfer and temperature variation rules of molten steel during RH-MFB refining process were developed.And a model for forecasting temperature of molten steel was established to simulate the actual process by using BP neural network and fuzzy algorithm.The forecasting results of temperature of 15 heats molten steel in 300t ladle RH-MFB refining show that temperature maximum error of molten steel at decarburization end-point of RH process is 7.2℃,a minimum 0.3℃,average error of 4.52℃ and within ±5℃ by 82.7%.Also,according to the actual situation,temperature compensation was achieved by adding aluminum or cooling materials.