-
燃煤发电脱硫系统的污染排放治理与系统能效的协同耦合机理对实现节能优先的目标具有决定性作用,也是集中治理燃煤污染、解决环境问题的关键。石灰石-石膏湿法烟气脱硫技术是国内外大型燃煤电厂应用范围最广的脱硫技术[1-4],而脱硫过程是涉及三传一反等多因素相互关联的复杂工艺流程,因此,湿法脱硫系统的污染物治理与能耗协同评价仍是一项技术挑战。目前,国内外脱硫系统的能效评价指标体系主要以统计方法构建数学模型,单一地从脱硫系统能耗或者脱硫效率的角度进行分析和评价。在对脱硫效率的评价及分析中,GUO等[5]将数学模型与人工神经网络(ANN)[6-7]联合,建立了一种预测二氧化硫排放的新型WFGD模型;王俊[8]通过SPSS利用多元统计学的方法[9]对某机组脱硫系统的运行数据进行了多元回归建模,确定了影响脱硫效率的关键因子。在对脱硫系统能耗评价方面,杨勇平等[10]以我国典型火电机组湿法脱硫系统实际运行数据[11]为基础,对主要参数进行多元回归分析,建立了脱硫能耗的数学模型;王岳宸[12]设计了基于电耗的能耗计算方法,将石灰石消耗和工艺水耗折算为电耗[13],构建了基于单位SO2脱除量的系统综合能效指标。尽管张健华[14]在建立火电能效综合评价指标体系的基础上,借助灰色关联度分析[15-17]、模糊综合分析[18]、矢量投影理论分析对电厂的能效利用水平进行全面综合的评价分析[19],但只是单纯围绕能效评价而未涉及污染物治理层次。可以看出,现有的评价指标主要存在的不足是不能综合反映脱硫系统污染物治理和系统能耗的协同影响问题。
本研究以某盐化公司220 t·h−1锅炉脱硫塔为例,利用脱硫控制系统在线运行数据,基于支持向量机的深度自学习方法[20],构建了SO2污染排放与系统能效的智能自适应预测模型,系统研究了脱硫系统的污染排放治理与系统能效的协同耦合作用机理,提出了综合考虑污染与能耗协同的评价方法,为实际工业锅炉的能效评价提供参考。
基于支持向量机的湿法脱硫塔污染治理优化与能效评价
Synergetic mechanism and evaluation of pollution and energy consumption of wet desulfurization tower based on support vector machine regression
-
摘要: 环境污染排放治理与系统综合能耗之间的协同优化与评价是解决我国环境与能源不堪重负的理论前提。针对燃煤电厂脱硫系统协同优化与评价的关键问题,基于湿法脱硫控制系统在线运行数据的支持向量机(SVM)智能深度自学习,构建了脱硫系统实时在线污染排放指标和系统综合能耗指标的协同预测模型,明晰了关键调控参数-污染排放治理-系统综合能耗之间的协同耦合规律,据此提出了工业级锅炉脱硫塔的污染排放治理与系统综合能耗的创新协同评估方法。结果表明:出口二氧化硫排放浓度和系统综合能耗指标与氧含量呈负关联协同耦合关系,与浆液密度呈先减后增的协同耦合关系,而污染排放治理与系统综合能耗指标呈负关联协同耦合关系;提高氧含量并将浆液密度控制在约1 250 kg·m−3,可使系统污染排放治理与综合能耗指标均处于最优状态,协同优化可使系统综合能耗指标最大降幅达8.3%。示范工程验证表明:脱硫系统污染排放与综合能耗指标的协同预测模型精准可靠,其最大误差小于10%。综合上述结果,基于支持向量机的回归方法可应用于工业级湿法脱硫锅炉的污染排放治理与能效评价,对实际脱硫工程的优化运行具有指导意义。Abstract: The synergistic optimization and evaluation between environmental pollution discharge control and comprehensive energy consumption of the system is the theoretical premise to solve the overburdened environment and energy in China. To solve the key scientific problem for the synergistic optimization and evaluation of desulfurization system of coal-fired power plants, based on online operation data of wet desulfurization control system, the support vector machine (SVM) intelligent deep self-learning method was used to build a synergistic prediction model for real-time online pollution emission control indices and comprehensive energy consumption indices of desulfurization systems, and clarify the synergistic coupling rules among the key control parameters-polluting emission control-the comprehensive energy consumption of the system. Then an innovative synergistic assessment method for pollution emission control and comprehensive energy consumption of industrial-grade boiler desulfurization towers was proposed. The results showed that the sulfur dioxide emission concentration in the outlet and the comprehensive energy consumption index of the system were negatively correlated with the oxygen content, and they firstly decreased then increased with the density of the slurry, while the pollution emission control was negatively correlated with the comprehensive energy consumption index of the system. Both the system pollution discharge control and comprehensive energy consumption indices were in the optimal state when increasing the oxygen content and controlling the slurry density at about 1 250 kg·m−3. The synergistic optimization could result in the maximum decline of 8.3% in the comprehensive energy consumption index of the system. The demonstration project verification showed that the synergistic prediction model of the pollution discharge and comprehensive energy consumption index of the desulfurization system was accurate and reliable, and its maximum error was less than 10%. Based on the above results, the support vector machine regression method can be applied to the pollution control and energy efficiency evaluation of industrial-grade wet desulfurization boilers, and has guiding significance for the optimal operation of actual desulfurization projects.
-
表 1 实际运行工况下参数比较
Table 1. Comparison of parameters under actual operating conditions
序号 烟气温度/
℃入口烟气流量/
(km3·h−1)入口SO2浓度/
(mg·m−3)烟气
含氧量/%浆液
pH浆液密度/
(kg·m−3)浆液循环泵
数量/台出口SO2浓度/
(mg·m−3)总能耗/
(kWh)1 101.61 199.93 1 443 6.90 3.5 1 281 2 91.86 418.36 2 103.66 221.21 1 401 6.94 4.3 1 074 2 32.13 400.93 3 104.54 206.71 1 417 6.09 4.4 1 077 2 42.81 424.29 4 104.61 211.93 1 574 7.94 4.9 1 077 2 22.75 373.15 5 109.45 192.67 1 477 6.96 4.8 1 340 2 17.19 400.75 6 100.29 194.99 1 288 7.14 4.7 1 075 2 29.39 385.06 7 111.57 208.92 1 654 6.64 4.2 1 284 2 37.17 405.39 8 109.74 188.10 1 432 7.17 5.2 1 347 2 10.90 399.37 9 110.03 177.48 1 503 6.94 5.1 1 286 2 12.63 392.56 10 106.30 190.06 1 373 7.39 4.0 1 080 2 41.03 361.02 -
[1] 中国环境保护产业协会电除尘委员会. 燃煤电厂烟气超低排放技术[M]. 北京: 中国电力出版社, 2015. [2] LIU X W. Progress of desulfurization and denitration technology of flue gas in china[J]. Earth and Environmental Science, 2019, 242(4): 2447-2450. [3] 金平, 王昊辰, 李磊, 等. 烟气脱硫技术现状及展望[J]. 当代化工, 2019, 48(1): 119-121. doi: 10.3969/j.issn.1672-8114.2019.01.075 [4] SHI W X, LIN C, CHEN W, et al. Environmental effect of current desulfurization technology on fly dust emission in China[J]. Renewable and Sustainable Energy Reviews, 2017, 72(5): 1-9. doi: 10.1016/j.rser.2017.01.033 [5] GUO X D, ZHENG L, SHU Y X, et al. Modeling and optimization of wet flue gas desulfurization system based on a hybrid modeling method[J]. Journal of the Air & Waste Management Association, 2019, 69(5): 565-575. [6] JANG H, SHU L, SO S. Analysis the compressive strength of flue gas desulfurization gypsum using artificial neural network[J]. Journal of Nanoscience and Nanotechnology, 2020, 20(1): 485-490. doi: 10.1166/jnn.2020.17235 [7] FICHOU D, MORLOCK G E. Powerful artificial neural network for planar chromatographic image evaluation, shown for denoising and feature extraction[J]. Analytical Chemistry, 2018, 90(11): 6984-6991. doi: 10.1021/acs.analchem.8b01298 [8] 王俊. 火力发电厂石灰石-石膏湿法脱硫系统优化运行研究[D]. 北京: 北京交通大学, 2010. [9] SAMPADA D, VIJAI S, VIJAY P. Stepwise cox regression analysis in SPSS[J]. Cancer Research, Statistics, and Treatment, 2018, 1(2): 167-170. [10] 杨勇平, 袁星, 黄圣伟, 等. 火电机组湿法脱硫系统能耗的回归分析[J]. 工程热物理学报, 2012, 33(11): 1854-1859. [11] 胡满银, 刘松涛, 刘炳伟, 等. 湿式脱硫装置脱硫效率的回归分析[J]. 中国电力, 2004(7): 75-77. [12] 王岳宸. 超低排放条件下湿法脱硫塔能效分析研究[D]. 济南: 山东大学, 2018. [13] 刘京. 火电厂锅炉及辅助系统能效评价方法的研究[D]. 北京: 华北电力大学, 2014. [14] 张健华. 火电厂能效综合评价[D]. 北京: 华北电力大学, 2015. [15] 杨少龙, 韩志涛, 潘新祥, 等. 基于灰色层次分析法的船舶废气SOx减排技术评价[J]. 环境工程学报, 2015, 9(10): 4928-4934. doi: 10.12030/j.cjee.20151049 [16] 刘剑. 燃煤机组湿法脱硫系统能效评估方法研究[J]. 电气应用, 2013, 32(S2): 91-94. [17] 孙芳芳. 浅议灰色关联度分析方法及其应用[J]. 科技信息, 2010(17): 880-882. doi: 10.3969/j.issn.1001-9960.2010.17.0721 [18] 史梦洁. 石灰石-石膏湿法脱硫系统综合能效评估方法研究[D]. 北京: 华北电力大学, 2014. [19] 崔亚兵. 燃煤电厂烟气脱硫技术和经济性分析及其模糊综合评价研究[D]. 南京: 东南大学, 2005. [20] KRISHNAMOORTHY S, RUEDA L, SAAD S, et al. Identification of user behavioral biometrics for authentication using keystroke dynamics and machine learning[C]//International Conference on Biometric Engineering and Applications. Amsterdam, 2018: 50-57. [21] 马旭霞. 支持向量机理论及应用[J]. 科学技术创新, 2019(2): 13-14. [22] 李永娜. 基于支持向量机的回归预测综述[J]. 信息通信, 2014(11): 32-33. doi: 10.3969/j.issn.1673-1131.2014.11.017 [23] 陈金凤. 支持向量机回归算法的研究与应用[D]. 无锡: 江南大学, 2009. [24] DONG Y, CHENG W, LI S. A new regression method based on SVM classification[C]//IEEE. Eighth International Conference on Fuzzy Systems & Knowledge Discovery, IEEE. Shanghai, 2011: 978-1011. [25] ZENG X, CHEN X. SMO-based pruning methods for sparse least squares support vector machines[J]. IEEE transactions on Neural Networks, 2005, 16(6): 1541-1546. doi: 10.1109/TNN.2005.852239 [26] 陈柳, 李岁月. 对FGD系统的研究及脱硫效率影响因素分析[J]. 现代盐化工, 2018, 45(2): 41-42. doi: 10.3969/j.issn.1005-880X.2018.02.019 [27] 吕新锋. 石灰石-石膏湿法烟气脱硫设施常见故障及影响脱硫效率因素分析[J]. 电力科技与环保, 2018, 34(2): 27-29. doi: 10.3969/j.issn.1674-8069.2018.02.008 [28] HRASTEL I, GERBEC M, STERGARŠEK A. Technology optimization of wet flue gas desulfurization process[J]. Chemical Engineering & Technology, 2010, 30(2): 220-233. [29] 薛龙, 胡颖丽. 脱硫浆液品质对脱硫效果的影响因素研究[J]. 山西冶金, 2019, 42(2): 24-26.