基于神经网络的锅炉燃烧过程多参数软测量  

Multi-Parameter Soft Measurement of Boiler Combustion Process Based on Neural Network

作　　者：田沛[1] 游欣佩[1] 

机构地区：[1]华北电力大学河北省发电过程仿真与优化控制重点实验室,河北保定071003

出　　处：《计算机仿真》2013年第10期162-166,共5页Computer Simulation

摘　　要：在分析锅炉燃烧经济性影响因素的基础上,应用主元分析方法精简建模变量数量,然后采用神经网络和现场数据建立锅炉相关过程模型,模型输出为表征锅炉燃烧工况的烟气含氧量等参数。仿真结果表明该模型具有较高的精度和一定的泛化能力。Based on the analysis of the boiler combustion efficiency factors, apply the principal component analy- sis method to streamline the number of modeling variables, then use neural networks and field data to establish boiler - related process model, and the model outputs are flue gas oxygen content and other parameters which characterize the working condition of boiler combustion. Simulation results show that this model has high precision and certain generalization.

关 键 词：燃烧经济性 主成分分析 神经网络 

分 类 号：TM621.7[电气工程—电力系统及自动化]