本文 通 过 对己有模型和锅炉运行现状的分析，尝试用BP神经网络的方法分析讨论锅炉效率在线计算和运行优化等问题。研究工作主要包括:基于BP神经网络进行煤的工业分析结果和元素分析结果之间的转换 考虑到煤在锅炉中燃烧时有固体未燃碳存在，给出了煤组成成分的实际结果的概念 详细讨论了过量空气系数的两种定义及特点，说明运行中用烟气分析结果确定过量空气系数时，对测得的湿烟气含氧量进行修正的必要性和方法:以反平衡法为基础给出了改进的锅炉效率在线计算模型 以效率在线计算为基础，尝试运用神经网络方法确定锅炉运行中的最佳过量空气系数。-Ac co rdi ngt ot hea nalysiso fa vailablem odelsa ndb oilerru nning

conditions, the problems in relation to the on-line calculation for

boiler efficiency and optimization operation can be discussed with

BPn euraln etworki nt histh esis.T her esearchw orkm ainlyc onclude:

The transformation between the industrial analysis results and the

element analysis results of coals based on neural network is discussed.

The concept of the actual results of coal ingredients is put forward

because there are some solid carbon not burned when coals burn in

boiler furnace. The two definitions of the excess air coefficient is

discusses in details. The necessity and methods of modifying the

oxygen content in damp gas are explained when the excess air

coefficient are determined with the results of gas analysis. The

modified on-line calculation model for the boiler efficiency on the

basis of the negative equation way is developed. Neural network

methods to determine the optimum excess air coef