摘要
为了准确地预测循环流化床锅炉NO_x排放量，以某热电厂循环流化床锅炉燃烧数据为样本，提出了基于支持向量机（SVM）的循环流化床锅炉NO_x排放特性GSA—SVM模型。由于SVM精度及泛化能力依赖于参数选择，故将万有引力搜索算法（GSA）运用到模型参数寻优过程中，利用不同工况下的样本数据检验了模型的预测性能，并将该模型分别与BP神经网络、粒子群（PSO）和遗传算法（GA）优化的SVM模型进行比较，仿真实验证明GSA—SVM模型具有很好的辨识能力及良好的泛化能力。
In order to accurately predict the NOx emissions of circulating fluidized bed boiler, a GSA-SVM model based on support vector machine （ SVM ） of NOx emission was established on the base of the sample data of circulating fluidized bed boilers in a power plant. The gravitation search algorithm （GSA） was applied to the process of the model parameters optimization because the accuracy and generalization ability of SVM model depended on the parameters. The prediction performance of the model was tested by sample datas under different experimental conditions. The regression model was compared with BP and the SVM model whose parameters were optomizated by particle swarm optimization（PSO） and genetic algorithms （GA）. Finally the simulation result shows that the GSA-SVM model has a good recognition ability and generalization ability.
出处
《计量学报》
CSCD
北大核心
2013年第6期602-606,共5页
Acta Metrologica Sinica
基金
国家自然科学基金（60774028）
河北省自然科学基金（F2010001318）
关键词
计量学
氮氧化物排放特性
万有引力搜索算法
支持向量机
循环流化床锅炉
Metrology
NOx emission characteristic1
Gravitation search algorithm
Support vector machine
Circulating fluidized bed boilers
