摘要
以某300MW亚临界循环流化床锅炉为研究对象,对锅炉的NOx排放量进行预测。利用模拟退火混合鸡群算法(SACSO)和核极端学习机(KELM)对不同工况下NOx的排放量进行建模;对比了差分进化算法,粒子群算法和原始鸡群算法,证明了改进后算法的优越性;之后,又对传统BP算法,支持向量机,极端学习机和核极端学习机模型进行对比;最终确定的SACSO-KELM模型具有更高的预测精度和稳定性以及更好的泛化能力,可选择将此模型用于锅炉NOx排放的建模预测。
Taking a 300 MW subcritical circulating fluidized bed boiler as an object of study,the NOxemission of the boiler was predicted accurately. A model about NOxemission from different working conditions was established using hybrid chicken swarm optimization based on simulated annealing(SACSO) and kernel extreme learning machine(KELM). By comparing in the differential evolution algorithm(DE),the particle swarm optimization(PSO) and the original chicken swarm optimization(CSO),the superiority of the improved algorithm was proved. Then,several models were compared in traditional BP algorithm,support vector machine(SVM),extreme learning machine(ELM) and KELM. The Finally determinded SACOS-KELM model has higher prediction accuracy,stability and better generalization ability,so this model is a good choice for boiler NOxemission in modeling and prediction.
作者
牛培峰
丁翔
刘楠
常玲芳
张先臣
NIU Pei-feng;DING Xiang;LIU Nan;CHANG Ling-fang;ZHANG Xian-chen(College of Electrical Engineering,Yanshan University,Qinhuangdao,Hebei 066004,China)
出处
《计量学报》
CSCD
北大核心
2019年第5期929-936,共8页
Acta Metrologica Sinica
基金
国家自然科学基金(61573306,61403331)
关键词
计量学
氮氧化物排放
循环流化床锅炉
模拟退火算法
鸡群算法
支持向量机
核极端学习机
metrology
NOx emission
circulating fluidized bed boiler
simulated annealing algorithm
chicken swarm algorithm
SVM
extreme learning machine with kernels
