欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
将支持向量机方法应用于电站锅炉SNCR系统的参数调节，建立了锅炉热态参数和SNCR系统参数与脱硝率之问的关系模型，并与神经网络方法预测结果进行了对比结果表明，平均相对误差和均方根误差都较后者降低了60％以上，同时线性相关系数r也提高了11％．通过该模型研究了典型工况下尿素用量和稀释水流量对脱硝率的影响，将其与机理性研究得到的结论进行比对，表明该模型所包含的信息很好地反映了样本数据中的规律．最后，研究了两个重要参数——核参数和边界参数对预测性能的影响，发现核参数取值应在[2，6]，此时，误差水平较低且对边界参数不敏感．
英文摘要：
SVM method was employed to adjust parameters in the SNCR system of utility boiler and to model the relationship between boiler thermal state parameter, SNCR system parameter and NOx removal efficiency. The prediction results was compared with those of neural network model, which shows that both mean relative error and root mean square error decrease by over 60% while linear correlation increases by 11%. Then, the effect of urea dosage and dilution water flow on NOx removal efficiency in some typical states was studied using this SVM model, which was compared with the conclusion drawn from the mechanism research. It is found that the information abtained from this model can perfectly reflect the sam- ple result. Finally, the effect of two important parameters, core parameter and boundary parameter in the model, on predic- tion function was examined. The conclusion indicates that core parameter should be fixed within [ 2, 6 ] interval, where error level is relatively low and core parameter is rather insensitive to the value of boundary parameter.
同期刊论文项目
同项目期刊论文
Physicochemical Characterizations and Desulfurization Properties in Coal Combustion of Three Calcium
Ceria as a catalyst for hydrogen iodide decomposition in sulfur-iodine cycle for hydrogen production
Experimental and modeling study on the influence of H2 addition on selective non-catalytic reduction
Catalytic mechanism of sodium compounds in black liquor during gasification of coal black liquor slu
期刊信息
位置：