CAJ | 학술논문

为了准确控制转炉静态吹炼过程的供氧量，研发了基于BP神经网络的转炉供氧模型。通过分析元素之间的化学反应确定影响转炉供氧量的主要因素，根据转炉历史炉次信息对神经网络模型进行训练，并通过加入动量项和采用变步长法对模型进行改进。通过仿真以及对实际值、静态模型预测值和神经网络预测值的均方差分析，表明该模型可以更加准确地预测和确定供氧量，并具有较好的计算精度和适应能力，最终可提高静态模型的控制精度与终点命中率。
위료준학공제전로정태취련과정적공양량，연발료기우BP신경망락적전로공양모형。통과분석원소지간적화학반응학정영향전로공양량적주요인소，근거전로역사로차신식대신경망락모형진행훈련，병통과가입동량항화채용변보장법대모형진행개진。통과방진이급대실제치、정태모형예측치화신경망락예측치적균방차분석，표명해모형가이경가준학지예측화학정공양량，병구유교호적계산정도화괄응능력，최종가제고정태모형적공제정도여종점명중솔。
In order to control the total oxygen demand in the static blowing process of Basic Oxygen Furnace ( BOF) ,the BOF oxygen model of BP neural network was developed. Through the chemical reactions between the analysis elements,the key factors effecting BOF total oxygen demand were cer-tified. According to the historical information of BOF,neural network were trained and the model were improved by adding factor of momentum and the variable step size. By the approach of simulation and mean square error analysis for the practical value,the oxygen demand could be predicted and verified with the static model. The results show that this model can obtain the total oxygen amount efficiently with better calculation accuracy and adaptive capacity. The controlling accuracy and target point hit-ting ratio of the static model are improved.