基于高斯回归的百万二次再热机组NO_(x)排放预测分析及应用  

Prediction analysis and application of NO_(x) emissionof 1000 MW unit with double-reheat cycle based on Gaussian regression

作　　者：徐民 XU Min

机构地区：[1]中国大唐集团科学技术研究总院有限公司华东电力试验研究院,安徽合肥230088

出　　处：《节能》2022年第10期57-59,共3页Energy Conservation

摘　　要：研究百万二次再热机组的NO_(x)排放量预测模型。采用高斯回归对锅炉燃烧过程进行建模,采用Spearman相关性系数对建模历史数据进行相关性分析,从49项测点数据中选取相关性较高的28项测点数据作为自变量,NO_(x)生成量为标签变量。结果显示:测试数据显示搭建回归模型的R~2为0.913 271,相对误差在10%范围内;现场部署该模型并实时计算,结果的准确性较高,为运行调整提供参考意见。The predictive models on NO_(x) emission of 1 000 MW ultra-supercritical unit with double-reheat cycle were studied. The boiler combustion process was modeled by Gaussian regression, and the Spearman correlation coefficient was used for correlation analysis of the modeled historical data. 28 measurement points with high correlation, were selected from 49 measurement points as independent variables, and NO_(x) emission as label variables. The results show that, the built regression model R~2 is 0.913 271, and the relative error is within the range of 10%. The model is deployed on site and calculated in real time, and the calculation results have high accuracy, which can provide reference for operation adjustment.

关 键 词：NO_(x) 高斯回归 相关性系数 

分 类 号：X773[环境科学与工程—环境工程]