系统仿真学报 ›› 2020, Vol. 32 ›› Issue (1): 35-43.doi: 10.16182/j.issn1004731x.joss.19-0235

• 仿真建模理论与方法 • 上一篇    下一篇

水泥窑协同处置生活垃圾的燃烧特性分析优化

吴敬兵, 唐汉卿, 胥军   

  1. 武汉理工大学 机电工程学院,湖北 武汉 430070
  • 收稿日期:2019-06-03 修回日期:2019-07-07 发布日期:2020-01-17
  • 作者简介:吴敬兵(1971-),男,湖北黄冈,博士,副教授,研究方向为CAD/CAM、计算机仿真、建材工业;唐汉卿(1994-),男,江苏徐州,硕士生,研究方向为水泥厂大数据应用;胥军(1977-),男,湖北襄阳,博士后,副教授,研究方向为建材企业信息化。

Analysis and Optimization of Combustion Characteristics of Cement Kiln Cooperatively Disposing Domestic Refuse

Wu Jingbing, Tang Hanqing, Xu Jun   

  1. School of Mechanical and Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China
  • Received:2019-06-03 Revised:2019-07-07 Published:2020-01-17

摘要: 针对传统方法难以分析掺烧生活垃圾后的水泥窑复杂燃烧特性的问题,引入数据挖掘技术,以国内某水泥厂为对象,采集相关参数数据,使用稳定性选择算法分析各参数对煤耗与NOx排量的影响系数,通过随机森林算法建立煤耗与NOx排量的数学模型,结合K-means聚类算法得出关键优化参数及其最优值。结果表明,该方法能够建立精确的煤耗与NOx排量模型,挖掘出节能减排的关键优化参数及其最优目标值。通过改善关键优化参数至最优值,能够大大降低煤耗与NOx排量,可指导水泥厂优化窑内燃烧特性。

关键词: 水泥窑, 生活垃圾, 燃烧性能, 数据挖掘, 稳定性选择, 随机森林, k-means

Abstract: Because the traditional methods can hardly analyze the complex combustion characteristics of cement kiln mixed with domestic refuse, a data mining technology is introduced. A domestic cement plant is selected as the object, and its operating data and relevant parameters are collected. The influence coefficient of each parameter on coal consumption and NOx emission is analyzed by using Stability Selection algorithm. The mathematical model of coal consumption and NOx emission is established with Random Forest algorithm, and the key optimization parameters and their optimal values are obtained by K-means clustering algorithm. The result shows that this method can establish accurate models of coal consumption and NOx emission, and can find out the key optimization parameters and their optimal values for energy saving and emission reduction. By adjusting the key optimization parameters, coal consumption and NOx emission can be greatly reduced. This method can guide cement plant to optimize kiln combustion performance.

Key words: cement kiln, domestic refuse, combustion performance, data mining, stability selection, random forest, k-means

中图分类号: