摘要
在分析焦炉火道温度特性的基础上,提出了一种基于线性回归和神经网络模型的火道温度软测量集成模型;分析生产工艺得到典型蓄热室的选取原则,从典型蓄热室获得蓄顶温度,建立一元和二元线性回归模型反映蓄顶温度和火道温度的线性关系;建立神经网络模型拟和蓄顶温度和火道温度的非线性关系;最后利用误差最小法将线性回归模型和神经网络模型集成,提高软测量精度;模型实际运行效果验证了所建模型的有效性。
An integrated model combining linear regress and neural network based on the features of coke oven flue temperature are proposed. Rules of selecting typical regenerators are put forward by analyzing features of process. Linear regression models are built to map the linear relationship between flue temperature and top of regenerator temperature; neural network models are built to map the nonlinear relationship. At last, least error method is employed to integrate the output of linear regression and neural network models. The run results of the models validate the method.
出处
《计算机测量与控制》
CSCD
2006年第2期149-151,共3页
Computer Measurement &Control
基金
国家杰出青年科学基金项目(60425310)
教育部青年教师奖项目(教人[2002]5号)。
关键词
焦炉
软测量
神经网络
模型集成
coke oven
soft-sensing
neural network
integration of models
