摘要
由于锅炉设备庞大、操作参数复杂和煤质多变,通过普通方法建立NOx排放特性的数学模型非常困难。借助正交实验法获取了某200 MW燃煤锅炉的多工况运行数据,应用神经网络的非线性特性建立了该锅炉NOx排放的神经网络模型。通过引入成本系数,将NOx排放特性和反映锅炉效率的特征量统一起来,应用基于实数编码的遗传算法对锅炉高效低NOx运行特性进行建模并寻优。理论分析和实验结果表明:该方法可获得锅炉的最佳运行参数,为锅炉的经济与环保运行提供可靠的决策依据。
Based on the orthogonal experimental data ofa 200MW utility boiler, the artificial neural network was used to describe NOx emission property. In combination with a cost coefficient, a neural network based model is set up. A decimal genetic algorithm was used to search the optimal solution of the neural network model. The calculated optimum results were tested and verified on the utility boiler. This method can provide boiler with advisable operating parameters.
出处
《电站系统工程》
北大核心
2006年第5期5-7,11,共4页
Power System Engineering
关键词
锅炉
燃烧优化
神经网络
遗传算法
氮氧化物
boiler
combustion optimization
neural network
genetic algorithms
NOx
