摘要
对自某热电厂350MW燃煤机组所采集的数据进行分析。先将现场采集的数据进行预处理,并运用最小二乘法建立了系统的预测模型。该模型只需通过输入数据,便能够预测出燃烧产生的氮氧化物的浓度。通过仿真和实测数据的对比可知预测结果较好。此外,通过偏最小二乘法建模来说明输入变量之间不存在冗余,即本文中采用最小二乘法不存在数据饱和问题。
The data collected from a 350MW coal-fired unit in a thermal power plant are analyzed. Firstly, the data of the field data are preprocessed, and the prediction model of the system is established by using the least square. The model can predict the concentration of nitrogen oxides produced by burning just through the input data. The comparison between simulation and measured data shows that the prediction results are good. In addition, the model established through partial least squares to illustrate the absence of redundancy between input variables, that is, this article uses the least squares method does not exist data saturation problem.
出处
《自动化技术与应用》
2017年第12期5-9,共5页
Techniques of Automation and Applications
基金
国网吉林省电力有限公司电力科学研究院科技项目(编号KY-GS-15-02-05)
关键词
选择性催化还原
最小二乘
预测控制
模型辨识
selective catalytic reduction （SCR）
least squares method
predictive control
model identification
