作者:梁涛1, 靳云杰1, 姜文2, 刘子豪1
作者单位:1. 河北工业大学人工智能与数据科学学院,天津 300130;
2. 河北建投能源投资股份有限公司,河北 石家庄 050011
关键词:燃煤锅炉;NOx排放量预测;燃烧优化;多元宇宙优化算法;加权最小二乘支持向量机
摘要:
为降低电厂燃煤锅炉的NOx排放浓度,提出一种基于改进多元宇宙优化算法(improved multi-verse optimizer algorithm,IMVO)和加权最小二乘支持向量机(weighted least squares support vector machine,WLSSVM)的锅炉NOx排放优化方法。首先,针对多元宇宙优化算法TDR值下降速度较慢而导致旅行距离增加的问题,提出一种改进的多元宇宙算法;然后,采用IMVO算法对WLSSVM模型参数进行寻优,建立基于IMVO-WLSSVM的NOx排放量预测模型;最后,基于所建预测模型,采用IMVO算法对锅炉运行可调参数进行寻优来降低NOx排放浓度。采用某330 MW机组燃煤锅炉的运行数据对模型进行验证,结果表明:所建预测模型的平均绝对百分比误差为1.09%,相对于其他几种预测模型具有更高的预测精度,改进的多元宇宙优化算法可以使优化后的NOx排放浓度更低,具有更好的寻优效果。
Optimization of NOx emissions from coal-fired boilers based on improved MVO and WLSSVM
LIANG Tao1, JIN Yunjie1, JIANG Wen2, LIU Zihao1
1. College of Artificial Intelligence and Data Science, Hebei University of Technology, Tianjin 300130, China;
2. Hebei Jiantou Energy Investment Co., Ltd., Shijiazhuang 050011, China
Abstract: In order to reduce the NOx emission concentration of coal-fired boilers in power plants, a method for boiler NOx emission optimization based on improved multi-verse optimizer algorithm (IMVO) and weighted least squares support vector machine (WLSSVM) is proposed. First of all, to solve the problem that the TDR value of the multi-verse optimizer algorithm decreases slowly and the travel distance increases, an improved multi-verse optimizer algorithm is proposed. Then, the IMVO algorithm is used to optimize the WLSSVM model parameters, and the NOx emission prediction model based on IMVO-WLSSVM is established. Finally, based on the built prediction model, the IMVO algorithm is used to optimize the adjustable parameters of the boiler operation to reduce the NOx emission concentration. The model was verified with the operating data of a 330 MW unit coal-fired boiler, and the results showed that the average absolute percentage error of the built prediction model is 1.09%, which has a higher prediction accuracy than other prediction models. And the improved multi-verse optimization algorithm can make the optimized NOx emission concentration lower and have a better optimization effect.
Keywords: coal-fired boiler;NOx emission forecast;combustion optimization;multi-verse optimizer algorithm;weighted least squares support vector machine
2021, 47(10):148-154 收稿日期: 2020-12-07;收到修改稿日期: 2021-03-02
基金项目: 河北省科技支撑计划资助项目(19210108D,19214501D,20314501D)
作者简介: 梁涛(1975-),男,河北石家庄市人,教授,博士,研究方向为火力发电、数据挖掘、新能源、人工智能
参考文献
[1] WANG J, FAN W, LI Y, et al. The effect of air staged combustion on NO x emissions in dried lignite combustion[J]. Energy, 2012, 37(1): 725-736
[2] TANG Y, MA X, LAI Z, et al. NO x and SO2 emissions from municipal solid waste (MSW) combustion in CO2/O2 atmosphere[J]. Energy, 2012, 40(1): 300-306
[3] 余廷芳, 刘冉. 基于RBF神经网络和BP神经网络的燃煤锅炉NOx排放预测[J]. 热力发电, 2016, 45(8): 94-98
[4] VAPNIK V. Statistical learning theory[M]. New York: Wiley-Interscience, 1998.
[5] 但长林, 李三雁, 张彬. 基于样本熵和SVM的滚动轴承故障诊断方法研究[J]. 中国测试, 2020, 46(11): 37-42
[6] 李爱莲, 郭志斌, 解韶峰, 等. 蚁群和粒子群混合优化SVM的钢板表面缺陷分类研究[J]. 中国测试, 2020, 46(1): 110-116
[7] SUYKENS J, VANDEWALLE J. Least squares support vector machine classifiers[J]. Neural processing letters, 1999, 9(3): 293-300
[8] TAN P, XIA J, ZHANG C, et al. Modeling and reduction of NO x emissions for a 700 MW coal-fired boiler with the advanced machine learning method[J]. Energy, 2016, 94: 672-679
[9] TANG Z, ZHANG H. Modeling NOx emission of coal-fired boiler with differential evolution optimized least square support vector machine[C]//2018 Chinese Control And Decision Conference (CCDC). IEEE, 2018: 3364-3367.
[10] 顾燕萍, 赵文杰, 吴占松. 基于最小二乘支持向量机的电站锅炉燃烧优化[J]. 中国电机工程学报, 2010, 30(17): 91-97
[11] SUYKENS J, DE BRABANTER J, LUKAS L, et al. Weighted least squares support vector machines: robustness and sparse approximation[J]. Neurocomputing, 2002, 48(1-4): 85-105
[12] MIRJALILI S, MIRJALILI S M, HATAMLOU A. Multi-verse optimizer: a nature-inspired algorithm for global optimization[J]. Neural Computing and Applications, 2016, 27(2): 495-513
[13] BENMESSAHEL I, XIE K, CHELLAL M. A new evolutionary neural networks based on intrusion detection systems using multiverse optimization[J]. Applied Intelligence, 2018, 48(8): 2315-2327
[14] 聂颖, 任楚苏, 赵杨峰. 多元宇宙优化算法改进SVM参数[J]. 辽宁工程技术大学学报(自然科学版), 2016, 35(12): 1507-1511