高炉炼铁含硅量的动态预测

Dynamic Prediction of Silicon Content in Blast Furnace Ironmaking

针对高炉炼铁过程中含硅量预测滞后性和复杂性等问题,建立以改进型Elman神经网络为基础的铁水含硅量动态预测数学模型,利用MATLAB工具箱对动态预测模型进行仿真和预测。首先利用蒙特卡洛法随机选取一组数据作为仿真数据样本,对改进后的Elman神经网络进行离线训练。然后用训练后的Elman神经网络进行铁水含硅量预测,再用残差分析法分析预测值与数据样本之间的显著性差异,用统计检验法讨论模型的可行性。结果表明,铁水含硅量数值预测误差较小,铁水含硅量的实际值和预测值无显著性差异,含硅量的测结果是可信的。

Aiming at the problems of hysteresis and complexity of silicon content prediction during ironmaking in blast furnace,a mathematical model for dynamic prediction of silicon content in molten iron based on Elman neural network was established,and the model was simulated and predicted using MATLAB toolbox.Firstly,uses Monte Carlo method to randomly select a set of data as simulation data samples,offline training of the improved Elman neural network.Then,the trained Elman neural network is used to predict the silicon content of molten iron.The residual difference analysis method was used to analyze the significant difference between the predicted value of silicon content in molten iron and the data samples.The feasibility of dynamic predictive control of silicon content in molten iron was discussed by statistical test method.The results show that the numerical prediction error of the silicon content in molten iron is small.There is no significant difference between the actual value and the predicted value of silicon content in molten iron.The prediction result is credible.