SCR烟气脱硝系统动态建模方法比较
发布时间：2019-04-04 20:33
【摘要】：SCR系统脱硝过程普遍应用的建模方法分为机理建模和数据建模两种,但这两种方法之间的对比研究相对较少。确定两种方法各自的适用范围,从而选取适宜的建模方法是确保模型准确性的前提。文中根据E-R反应机理以及支持向量机(support vector machine,SVM)、BP神经网络(BP neural network,BPNN)、核偏最小二乘(kernel partial least squares,KPLS)等方法分别建立了SCR系统的机理和数据动态模型。采用现场实际运行数据对模型进行验证,并根据模型精度评价指标以及建模耗时对各种建模方法进行对比。结果表明,在局部工况样本条件下,机理模型计算精度更高,计算量较小。在全局工况样本条件下,数据模型的拟合和泛化能力更强,但是数据模型的计算量更大,建模耗时更长。
[Abstract]:The modeling methods widely used in the denitrification process of SCR system can be divided into two types: mechanism modeling and data modeling, but the comparative study between the two methods is relatively few. In order to ensure the accuracy of the model, it is necessary to determine the applicable scope of the two methods, and then select the appropriate modeling method to ensure the accuracy of the model. In this paper, the mechanism and data dynamic models of SCR system are established according to Eur reaction mechanism, support vector machine (support vector machine,SVM), BP) neural network (BP neural network,BPNN), kernel partial least squares (kernel partial least squares,KPLS) and so on. The model is verified by field running data, and various modeling methods are compared according to the evaluation index of model precision and the time consuming of modeling. The results show that the calculation accuracy of the mechanism model is higher and the calculation amount is less under the local working conditions. The fitting and generalization ability of the data model is stronger under the condition of the global working condition sample, but the calculation of the data model is larger and the modeling time is longer.
【作者单位】： 新能源电力系统国家重点实验室(华北电力大学);海南电力技术研究院;
【基金】：中央高校基本科研业务费专项资金资助(2016MS47,2015XS69) 中国南方电网有限责任公司科技项目(073000KK51140001)~~
【分类号】：X773
[Abstract]:The modeling methods widely used in the denitrification process of SCR system can be divided into two types: mechanism modeling and data modeling, but the comparative study between the two methods is relatively few. In order to ensure the accuracy of the model, it is necessary to determine the applicable scope of the two methods, and then select the appropriate modeling method to ensure the accuracy of the model. In this paper, the mechanism and data dynamic models of SCR system are established according to Eur reaction mechanism, support vector machine (support vector machine,SVM), BP) neural network (BP neural network,BPNN), kernel partial least squares (kernel partial least squares,KPLS) and so on. The model is verified by field running data, and various modeling methods are compared according to the evaluation index of model precision and the time consuming of modeling. The results show that the calculation accuracy of the mechanism model is higher and the calculation amount is less under the local working conditions. The fitting and generalization ability of the data model is stronger under the condition of the global working condition sample, but the calculation of the data model is larger and the modeling time is longer.
【作者单位】： 新能源电力系统国家重点实验室(华北电力大学);海南电力技术研究院;
【基金】：中央高校基本科研业务费专项资金资助(2016MS47,2015XS69) 中国南方电网有限责任公司科技项目(073000KK51140001)~~
【分类号】：X773
【相似文献】
相关期刊论文 前10条
1 吴碧君;刘晓勤;王述刚;万志星;盛永校;;烟气脱硝工艺及其化学反应机理[J];电力环境保护;2006年02期
2 杨旭中;;关于烟气脱硝的几点认识[J];电力勘测设计;2008年01期
3 周W，
本文编号：2454147
本文链接：https://www.wllwen.com/kejilunwen/dianlidianqilunwen/2454147.html
