查询结果： 张子阳，孙彦广．基于灰色Elman神经网络转炉吹氧量的预测[J]．计算机应用与软件，2018，35(11)：103 - 107．
PREDICTION OF OXYGEN AMOUNT IN CONVERTER BASED ON GREY ELMAN NEURAL NETWORK
中国钢研科技集团有限公司冶金自动化研究设计院 北京 100071
Automation Research and Design Institute of Metallurgical Industry, China Iron & Steel Research Institute Group, Beijing 100071, China
Oxygen amount in the converter
Grey Elman neural network
钢铁企业转炉动态吹炼的过程复杂，其冶炼过程中存在的非线性问题导致消耗的氧气量难以准确估量造成能源浪费。传统方法不能准确预测，一些简单的神经网络预测的精度较低。为了提高预测效果，提出采用灰色Elman神经网络来建立转炉吹氧量预测模型。通过优化权值和阈值，解决普通的神经网络局部最优和过拟合等问题。结合转炉炼钢用氧的特点对实际生产数据进行仿真计算，预测用氧量的平均误差为334 m3，远小于BP神经网络预测用氧量的平均误差976 m3。结果显示所建立的预测模型可有效快捷地确定转炉吹氧量，验证方法是有效的，具有更高的预测精度。
The process of dynamic blowing of converters is complex in iron and steel companies. The non-linear problems in the smelting process make it difficult to accurately estimate the amount of oxygen consumed, which causes energy waste. Traditional methods cannot accurately predict, and some simple neural network predictions have lower accuracy. In order to improve the prediction accuracy, we proposed a prediction model of oxygen amount in converter based on gray Elman neural network. The problems such as local optimality and overfitting of ordinary neural networks were solved by optimizing the weights and enthalpy values. Combined with the characteristics of oxygen in converter steelmaking, the simulation was carried out with actual production data. The average error of predicted oxygen consumption was 334 m3, which was much less than that of oxygen consumption predicted by BP neural network which was 976 m3. The results show that the established prediction model can efficiently and quickly determine the oxygen amount in converter. The method is verified to be effective and has higher prediction accuracy.