摘要
针对目前锅炉飞灰含碳量测量方法存在时间滞后和精度不高等问题,在分析锅炉飞灰含碳量影响因素和做锅炉燃烧特性实验的基础上,利用最小二乘支持向量机这种新的机器学习工具,建立了飞灰含碳量的软测量模型。应用该模型对燃煤电站锅炉的飞灰含碳量进行研究,理论分析和仿真计算表明,该方法学习速度快、泛化能力强、对样本的依赖程度低,比BP神经网络的软测量建模更具有推广力。
Unburned carbon content in the fly ash is a main factor that has great impacts on the boiler efficiency. The soil- sensing model for the unburned carbon in fly ash was built by using the least square support vector machines （LS-SVM）. The simulation was conducted on a 600 MW power plant boiler. The procedure of simulation and theoretical analysis indicate that the proposed method was effective verified. Good predicting performance was achieved with the proper learning parameters. The modeling results showed that support vector machine was a good toll for building combustion models and had better generalization ability and higher calculation speed comparing with other modeling approaches.
出处
《电力科学与工程》
2010年第1期39-43,59,共6页
Electric Power Science and Engineering
基金
中国国家高技术研究发展计划863计划(2007AA041106)
关键词
飞灰含碳量
支持向量机
软测量
单纯形算法
unburned carbon
support vector machine
soft-sensing
simplex algorithm
