摘要
近几年为控制发电成本和污染物排放,发电企业选择混煤掺烧和低氮燃烧器改造的方案,燃烧不稳定局部热传导变差等问题引起的受热面汽温偏差现象频发。分析了汽温偏差的理论影响因素,并基于某600 MW机组汽温偏差的实际问题,理论选择输入变量共56个,建立了PLS-GA-SVR的汽温偏差回归预测模型,分析了汽温偏差对主要输入参数的敏感度,筛选了影响汽温偏差的关键参数。计算结果表明:经过变量筛选可以在不影响回归效果的前提下大幅缩减模型的输入变量个数,10个变量输入的GA-SVR模型预测结果的均方根误差和可决系数分别为5.152和0.879,氧量和CD磨煤机的给煤量是影响汽温偏差最关键的参数。现场试验结果表明:磨煤机的煤粉分配不均是导致汽温偏差大的主要原因。
In recent years, in order to control the cost and pollutant emissions, the scheme of coal mixing and low nitrogen burner modification is applied. Steam temperature deviation on heating surface caused by combustion instability and poor heat conduction variation is frequent . In this paper, the theoretical influencing factors of steam temperature deviation were analyzed, 56 input variables of the prediction model were selected. The model based on PLS-GA-SVR algorithm is established. The sensitivity of the steam temperature deviation to the main input parameters is analyzed, and the key parameters are screened. The calculation results show that the number of input variables can be greatly reduced without affecting the regression effect. The root mean square error and R-squared of the results predicted by GA-SVR model with 10 variables were 5.152 and 0.879. The oxygen and the coal feed of mill C and D are the most critical parameters which affects the steam temperature deviation. The field measurement results show that the uneven distribution of pulverized coal is the main reason in this question.
作者
丁皓轩
李松山
唐文
严文龙
郝志兵
马永昱
DING Haoxuan;LI Songshan;TANG Wen;YAN Wenlong;HAO Zhibing;MA Yongyu(China Power Huachuang Electricity Technology Research Company Ltd.,Suzhou 215000,China;Huanggang Dabieshan Power Generation Co.,Ltd.,Huanggang 438000,China;Huainan Pingwei Power Generation Co.,Ltd.,Huainan 232000,China)
出处
《锅炉技术》
北大核心
2024年第2期32-38,共7页
Boiler Technology
关键词
电站锅炉
主蒸汽温度偏差
预测模型
变量筛选
遗传算法
power plant boiler
main steam temperature deviation
prediction model
variable screening
genetic algorithm(GA)
