基于神经网络的锅炉对流受热面灰污监测研究  

Monitoring Ash Fouling on the Boiler Convective Surfaces Based on the BP Neural Network

作　　者：吴观辉[1] 向文国[1] 

机构地区：[1]东南大学洁净煤发电及燃烧技术教育部重点实验室,江苏南京210096

出　　处：《锅炉技术》2005年第2期18-21,32,共5页Boiler Technology

摘　　要：采用多层前向型神经网络,对电站锅炉对流受热面的实时污染状况建立了监测模型。模型选取合适的参数组成输入向量,利用电站数据采集系统下载的实时机组数据,经规格化处理后对神经网络进行训练。结果表明,训练后的神经网络可以较准确地实现锅炉对流受热面的积灰状态的在线监测,为吹灰方案的最优化打下了良好的基础。A Fouling Monitoring Modeling System based on the BP neural network for boiler convection surfaces is presented in this paper. The real-time parameters of the boiler are selected as the inputs of the neural network instead of the simulated parameters of power boiler. The neural network is trained through the online data from DCS after normalization. Using the data from DCS, the neural network is simulated. The neural model can successfully monitor the soot-blowing process under different loads. From the simulation results , it is concluded that the trained neural model can be used to monitor the fouling state of boiler convection surfaces accurately and it can be used to optimize the soot-blowing process.

关 键 词：神经网络 对流受热面 积灰 监测 

分 类 号：O157.5[理学—数学]