论文部分内容阅读
为控制锅炉燃烧向环境排放NOx造成的污染,提出了分级燃烧技术的综合优化方案。建立了基于人工神经网络及模拟进化算法的100MW火电机组锅炉分级燃烧优化模型,选取16个影响因子进行了分级燃烧的7个可调节参数优化,以达到机组的性能优化目标。锅炉负荷为100%、90%、80%及70%,相应神经网络训练次数分别为11523、14810、13410及19732时满足均方差要求。该神经网络模型优化时采用的种群数为80,交叉概率为0.8,变异概率为0.15。结果表明:锅炉效率和NOx排放量优化计算值同实测值相对误差低于1%;NOx平均排放量由原来的812mg/m3降为645mg/m3。
In order to control the pollution caused by NOx emission from boiler combustion to the environment, a comprehensive optimization scheme of staged combustion technology was proposed. Based on artificial neural network and simulated evolutionary algorithm, a staged combustion optimization model of a 100MW thermal power unit boiler was established. Sixteen influential factors were selected to optimize the seven adjustable parameters of the staged combustion so as to achieve the goal of performance optimization. The boiler load was 100%, 90%, 80% and 70%, respectively, and the corresponding neural network training times were 11523, 14810, 13410 and 19732, respectively. The neural network model is optimized with population of 80, crossover probability of 0.8 and mutation probability of 0.15. The results show that the relative error between the calculated value of boiler efficiency and NOx emission is less than 1% with the measured value, and the average NOx emission is reduced from 812mg / m3 to 645mg / m3.