基于分层多项式模型的飞灰含碳量预测
张先炼，肖禹，潘波，周雨生
（湖北华电西塞山发电有限公司，湖北 黄石 435000）
摘要： 火力电厂中飞灰含碳量的软测量是有难度且意义重大的一个课题。提出基于主要飞灰影响因素的分层多项式
回归模型作为飞灰含碳量的预测模型。该模型通过分析含氧量对飞灰含碳量的影响，在此基础上添加其他工况参数的影
响，建立了适用于多层燃烧器多台磨煤机的分层多项式回归模型。以首阳山680 MW 电厂1 220 条实际工况数据作为原
始数据，训练并验证了该模型的预测效果，飞灰含碳量的平均绝对误差达到了0.45%，能满足飞灰含碳量软测量的需求。
关键词： 火力电厂；飞灰含碳量；含氧量；软测量；多项式回归
中图分类号： TK222     文献标识码： A     文章编号： 2095-0802-(2021)12-0166-04
Prediction of Carbon Content in Fly Ash Based on Hierarchical and Polynomial Model
ZHANG Xianlian, XIAO Yu, PAN Bo, ZHOU Yusheng
(Hubei Huadian Xisaishan Power Generation Co., Ltd., Huangshi 435000, Hubei, China)
Abstract: The soft measurement of carbon content in fly ash in thermal power plants is a difficult and significant subject. This
paper proposed a hierarchical and polynomial regression model based on the main influencing factors of fly ash as the prediction
model of carbon content in fly ash. By analyzing the influence of oxygen content on carbon content of fly ash, and adding the
influence of other working conditions, a hierarchical and polynomial regression model suitable for multi-layer burners and multiple
pulverizers was established. Taking 1 220 actual working condition data of Shouyangshan 680 MW Power Plant as the original
data, the prediction effect of the model was trained and verified. The average absolute error of carbon content in fly ash reaches
0.45%, which can meet the needs of soft measurement of carbon content in fly ash.

Key words: thermal power plant; carbon content in fly ash; oxygen content; soft measurement; polynomial regression