CAWOA-ELM混合模型的锅炉NO_x排放量预测  

Boiler NO_x Emission Prediction Based on a Hybrid Model of CAWOA and ELM

作　　者：赖敏 陈国彬[2] 刘超[3] 牛培峰[3] LAIMin;CHENGuobin;LIUChao;NIUPeifeng(School of Software Engineering,Chongqing Institute of Engineering,Chongqing 400056,China;Rongzhi College,Chongqing Technology and Business University,Chongqing 400033,China;School of Electrical Engineering,Yanshan University,Qinhuangdao 066004,Hebei Province,China)

机构地区：[1]重庆工程学院软件学院,重庆400056 [2]重庆工商大学融智学院,重庆400033 [3]燕山大学电气工程学院,河北秦皇岛066004

出　　处：《动力工程学报》2018年第11期874-879,共6页Journal of Chinese Society of Power Engineering

基　　金：国家自然科学基金资助项目(61403331;61573306)

摘　　要：针对燃烧过程中变量之间的强非线性和耦合性,利用极限学习机(ELM)和改进的鲸鱼优化算法(WOA)进行混合建模。该方法利用Sin混沌自适应鲸鱼优化算法(CAWOA)对极限学习机的模型参数进行搜索和优化,以提高极限学习机的泛化性能。在CAWOA算法中,通过引入Sin混沌搜索策略和自适应惯性权值来改善WOA算法的全局优化性能。在此基础上,利用优化后的极限学习机对330MW煤粉锅炉的NO_x排放质量浓度进行预测,建立了CAWOA-ELM的NO_x排放量预测模型,并与同类算法模型进行对比研究。结果表明:该方法具有更好的泛化能力,能更加精确地预测NO_x排放量。Considering the strong nonlinearity and coupling characteristics of input variables in combustion process of thermal power plants,an integrated model was established to predict the NOx emission based on extreme learning machine（ELM）and modified whale optimization algorithm（WOA）,by using Sin chaos adaptive whale optimization algorithm（CAWOA）to search and optimize the model parameters of ELM to improve its generalization performance.In CAWOA algorithm,the Sin chaotic search strategy and adaptive inertia weight were introduced to improve the global optimization performance of WOA algorithm.The optimized ELM was finally used to predict the NOx emission of a 330 MW coal-fired boiler,and the CAWOA-ELM scheme was compared with congeneric methods.Results show that the method proposed has better regression precision and generalization capability,which therefore can accurately predict the NOx emission of power plant boilers.

关 键 词：极限学习机 鲸鱼优化算法 混沌搜索 自适应惯性权值 NOX排放量 

分 类 号：TM621.2[电气工程—电力系统及自动化]