应用神经网络技术预报VD炉终点钢水温度

Prediction of End Temperature on VD by Neural Network Method

应用神经网络方法建立了VD炉钢水温度预报模型。用该模型在线连续预报的76炉次钢水中,预报温度与实际测量温度之差在±4℃和±5℃之内的炉次分别占67 1%和80 3%。分析了各工艺参数对终点温度的影响,据此对宝钢的生产实践提出了一些降低能耗的措施。

A prediction model for end temperature on VD furnace has been developed by neural network method. For the molten steel of 76 heats, the hittng ratio within ±4 ℃ and ±5 ℃ is 671 % and 803 % respectively. The influence of all factors on end temperature was analyzed, and some measures for energy saving aimed were put forward.