论文部分内容阅读
针对铁水硅含量预测高炉炉温的不科学性,本文以高炉铁水温度为研究对象,建立基于小波神经网络的预测高炉炉温模型。首先利用数理统计的方法对海量的样本数据进行预处理和特征提取,然后利用相关分析的方法归纳出影响炉温的关键变量,进而建立以出铁批次为时间序列的高炉铁水温度小波神经网络预测模型。最后,应用某钢厂高炉数据做模型试验,相对传统的BP神经网络模型,小波神经网络模型有较好的命中率和预测精度。
In order to predict the blast furnace temperature unscientific for the content of silicon in hot metal, this paper takes blast furnace hot metal temperature as the research object, and establishes the forecast blast furnace temperature model based on wavelet neural network. First of all, the method of mathematical statistics is used to pretreat and extract the large amount of sample data. Then, the correlation analysis is used to summarize the key variables that affect the furnace temperature, and then establish a wavelet neural network Predictive model. Finally, using a blast furnace data from a steel mill as a model test, the wavelet neural network model has better hit rate and prediction accuracy than the traditional BP neural network model.