燃煤电站锅炉NO_X排放的人工神经网络建模及优化分析  

Artificial neural network models of NO_X emission for coal-fired boilers and optimization of staged-air combustion

作　　者：朱群志[1] 李芳芹[1] 任建兴[1] 潘卫国[1] 李琦芬[1] 

机构地区：[1]上海电力学院能源与环境工程学院,上海200090

出　　处：《华东电力》2008年第10期128-131,共4页East China Electric Power

基　　金：上海市重点学科建设项目(P1302);上海高校选拔培养优秀青年教师科研专项基金资助项目

摘　　要：影响燃煤锅炉氮氧化物生成的因素很多且规律复杂。利用人工神经网络技术,使用某一电厂低NOX排放燃烧优化试验的数据,建立了该锅炉氮氧化物的排放模型。该模型预测精度较高、结果可信。通过建立的神经网络模型分析了配风方式的影响。结果表明:缩腰型配风方式较佳,而倒宝塔型配风方式优于正宝塔型配风方式。建立的神经网络模型可以为燃煤锅炉通过优化燃烧降低NOX排放提供理论指导。Many factors can affect NOX generation in coal-fired boilers. To solve the problem, artificial neural network technique was applied to construct the NOX emission model based on field test data, and effects of vertical distributions of secondary air are analyzed based on neural network predictions. It is concluded that a concave-shape distribution or low percentages of secondary air at low stories of a combustor are beneficial to NOX reduction. The constructed neural network model can provide theoretical support for NOX reduction through optimization of staged-air combustion in coal boilers.

关 键 词：燃煤锅炉 人工神经网络 氮氧化物排放 空气分级燃烧 

分 类 号：TK223[动力工程及工程热物理—动力机械及工程]