摘要
为实现游离氧化钙(f-CaO)含量的持续稳定预测,采用集成学习算法研究软测量实现方法。针对烧成系统中的复杂工况,首先展开工况分类,然后对每一类工况构建集成学习预测模型,同时引入在线建模的方式提高模型的泛化能力和时间有效性,解决了模型短期有效和重复性建模的问题。集成学习算法基于bagging的思想进行对多种弱学习器构建模型,通过检验发现模型效果显著优于单一模型效果。该算法融合了工艺生产特点和多种回归算法,具有较好的稳定性和提前性,实现了水泥质量的实时控制,助力水泥厂高质量稳定生产。
In order to achieve continuous and stable prediction of free calcium oxide(f-CaO)content,this paper uses ensemble learn⁃ing algorithm to study the soft measurement methods.For the complex working conditions in the firing system,the classification of work⁃ing conditions is firstly carried out,and then an integrated learning prediction model is constructed for each type of working condition,at the same time,the introduction of online modeling improves the generalization ability and time effectiveness of the model,solving the problems of short-term and repetitive modeling.The integrated learning algorithm is based on the idea of bagging to construct models for various weak learners,and it is found that the effect of the model is significantly better than that of a single model through inspec⁃tion.The algorithm combines the characteristics of process production and multiple regression algorithms,has good stability and ad⁃vanceness,realizes real-time control of cement quality,and helps cement plants to produce high-quality and stable products.
作者
崔保华
张成伟
李慧霞
陈克政
郭文洁
孙战军
Cui Baohua;Zhang Chengwei;Li Huixia;Chen Kezheng;Guo Wenjie;Sun Zhanjun(Sinoma International intelligent Technology Co.,Ltd.,NanJing,210036,China;不详)
出处
《水泥工程》
CAS
2024年第1期1-5,15,共6页
Cement Engineering
基金
中国建材集团攻关专项。
关键词
F-CAO
集成学习
软测量
工况分类
在线建模
f-CaO
ensemble learning
soft measurement
classification of working conditions
online modeling
