摘要
建立以终点碳和温度为间距的氧气用量非等间距灰色数列预测模型,并利用广义回归神经网络对灰模预测结果进行非线性组合优化,得到氧气质量的综合预测值。通过采集到的某钢厂实际生产数据,建立氧气质量的组合预测模型,得到平均相对精度达到97.39%的一步预测值。验证结果表明该组合模型是准确而有效的。
The non-equidistant grey sequence prediction model for the oxygen dosage was established,which takes terminal carbon and temperature as spacing and employs a generalized regression neural network to nonlinearly optimize grey prediction results so that acomprehensive prediction value of the oxygen dosage can be reached.Basing on the sampled data from a steel works,a combined prediction model for the oxygen weight was established to get a predicted value which boasting of 97.39% in average and relative accuracy.The verification results show that the combined model is effective and accurate.
出处
《化工自动化及仪表》
CAS
2013年第4期505-507,共3页
Control and Instruments in Chemical Industry
基金
天津市科技支撑计划重点项目(10ZCKFGX03400)
关键词
氧气质量预测
非等间距灰色模型
广义回归神经网络
组合模型
oxygen weight prediction
non-equidistant grey model
generalized regression neural network
combined model
