摘要
针对如何提高氨法烟气脱硫装置脱硫效率的预测精度,建立了以烟气量、喷淋浆液密度、液气比、喷淋塔pH值、预洗涤塔pH值和浓缩槽pH值为输入变量,以脱硫效率为输出量的最小二乘支持向量机(LS-SVM)脱硫效率的预测模型,并运用这一模型对某135 MW机组氨法烟气脱硫装置的脱硫效率进行了预测评判.结果表明:与传统的支持向量机模型进行比较,LS-SVM模型预测结果的最大误差只有传统模型的14%左右;LS-SVM模型所得预测值与实际值的误差小于5%,完全在工程所允许的误差范围之内,模型是可行的.
In order to improve the prediction precision desulfurization efficiency of an ammonia flue gas desulphurization system,a least squares-support vector machine（LS-SVM） model has been established for efficiency prediction purposes,which takes following parameters as the input variables,such as the flue gas flow rate,the slurry density,the liquid-gas ratio,the pH values respectively in spraying tower,pre-washing tower and in concentration basin,and the desulfurization efficiency as the output variable.The newly built model has been applied for efficiency prediction of the flue gas desulfurization system in a 135 MW unit.Results show that the new model is proved to be applicable.Compared with traditional SVM model,the maximum predicted error of new model is only 14% of the traditional one,and the predicted error against actual measurements is less than 5%,which is completely within the permitted range of engineering application.
出处
《动力工程学报》
CAS
CSCD
北大核心
2011年第11期846-850,共5页
Journal of Chinese Society of Power Engineering
