一种基于灰色预测与神经网络组合模型转 炉氧耗量预测方法，属于钢铁企业转炉氧气预测 技术领域。通过对影响转炉氧气消耗量的多个因 素分析后进行预测。利用Elman反馈神经网络非 线性计算特点，能够较好拟合预测复杂情况下的 非线性预测问题，发挥其反映系统动态特性的能 力，利用灰色系统理论不仅能够加快人工神经网 络预测模型的收敛速度，而且更有效展现转炉氧 气消耗量的变化规律，提高了预测准确性。结合 灰色预测与Elman神经网络建立的组合模型，该 组合模型全局搜索能力较强，同时对网络进行优 化，并且加入了适当的算子，增强了搜索局部最 优解和全局最优解的能力，能有效地提高预测精 度，指导炼钢生产合理用氧、提高生产效率。 

The invention relates to a prediction method of converter oxygen consumption based on the combined model of grey prediction and neural network, which belongs to the technical field of converter oxygen prediction in iron and steel enterprises. Through the analysis of many factors affecting the oxygen consumption of converter, it is predicted. Using the nonlinear calculation characteristics of Elman feedback neural network can better fit the nonlinear prediction problems in complex situations and give play to its ability to reflect the dynamic characteristics of the system. Using the grey system theory can not only accelerate the convergence speed of the artificial neural network prediction model, but also more effectively show the change law of converter oxygen consumption and improve the prediction accuracy. The combined model established by combining grey prediction and Elman neural network has strong global search ability, optimizes the network at the same time, and adds appropriate operators to enhance the ability to search local and global optimal solutions, which can effectively improve the prediction accuracy, guide the rational use of oxygen in steelmaking production and improve production efficiency.