摘要
NO_(X)对环境和人体健康危害极大,环保法规对电厂NO_(X)排放有严格的要求,降低NO_(X)的前提条件是对其精准的预测。以宁夏京能某电厂660 MW锅炉为研究对象,从DCS中提取到一段时间的数据,首先对数据进行预处理操作,通过归一化、特征提取等方法得到满意的数据,后续使用WOA-SVR算法建立了锅炉SCR入口处NO_(X)浓度预测模型。研究结果表明,通过WOA-SVR建立的预测模型预测精度高,拟合能力和泛化能力优秀,其中决定系数R2的大小为0.939,接近1;均方根误差RMSE为5.80,误差小。该模型为本机组低NO_(X)排放提供了有力的指导。
NO_(X)poses great harm to the environment and human health,and environmental regulations have strict requirements for NO_(X)emissions from power plants.The prerequisite for reducing NO_(X)is accurate prediction.Taking a 660 MW boiler in a power plant in Jingneng,Ningxia as the research object,a period of data was extracted from DCS.Firstly,the data was preprocessed and satisfactory data was obtained through normalization,feature extraction,and other methods.Subsequently,a WOA-SVR algorithm was used to establish a NO_(X)concentration prediction model at the SCR inlet of the boiler.The research results indicate that the prediction model established through WOA-SVR has high prediction accuracy,excellent fitting and generalization abilities,with a determination coefficient R2 of 0.939,close to 1;The root mean square error(RMSE)is 5.80,indicating a small error.This model provides strong guidance for the low NO_(X)emissions of our unit.
作者
赵启智
白鹤
汤吉昀
Zhao Qizhi;Bai He;Tang Jiyun(College of Energy and Control Engineering,Changji University,Changji,China)
出处
《科学技术创新》
2025年第18期88-91,共4页
Scientific and Technological Innovation
基金
新疆维吾尔自治区高等学校科学研究计划资助(XJEDU2024P077)
庭州英才(2024QN003)。
