基于机器学习的短程硝化/短程反硝化-厌氧氨氧化工艺脱氮性能预测与关键参数识别

吴宇伦; 李泽敏; 成晓倩; 邱光磊; 韦朝海

doi:10.13205/j.hjgc.202409017

基于机器学习的短程硝化/短程反硝化-厌氧氨氧化工艺脱氮性能预测与关键参数识别

doi: 10.13205/j.hjgc.202409017

1. 华南理工大学环境与能源学院, 广州 510006;
2. 华南师范大学环境学院, 广州 510006

基金项目:

国家自然科学基金项目“工业废水的形成、资源化利用与污染控制——以钢铁、冶金、矿山废水为例”(U1901218)

详细信息

作者简介:
吴宇伦(1999-),男,硕士研究生,主要研究方向为废水生物处理。mr_wuyulun@sina.com

通讯作者:
韦朝海(1962-),男,教授,主要研究方向为水污染控制理论及技术。cechwei@scut.edu.cn

计量
- 文章访问数: 45
- HTML全文浏览量: 11
- PDF下载量: 1
- 被引次数: 0
出版历程
- 收稿日期: 2024-07-23
- 网络出版日期: 2024-12-02

PREDICTION OF NITROGEN REMOVAL PERFORMANCE AND IDENTIFICATION OF KEY PARAMETERS OF PARTIAL NITRIFICATION/PARTIAL DENITRIFICATION-ANAMMOX PROCESS BASED ON MACHINE LEARNING

1. School of Environment and Energy, South China University of Technology, Guangzhou 510006, China;
2. School of Environment, South China Normal University, Guangzhou 510006, China

摘要

摘要: 短程硝化-厌氧氨氧化(PNA)与短程反硝化-厌氧氨氧化(PDA)工艺的脱氮性能会受到许多参数的影响。在综合考虑各种参数的基础上,对2种工艺的脱氮性能进行预测,并识别关键参数,能够为其实际工程应用提供优化目标。解决上述问题时,实验方法耗时耗力,而传统数学模型难以处理非线性关系。因此采用机器学习技术,构建的随机森林(RF)机器学习模型对2个工艺的出水总氮(TN)浓度进行了高精度预测,对PNA和PDA工艺出水TN浓度预测结果的决定系数(R²)分别为0.728、0.812。SHAP方法能够较好地解释模型的预测过程,并对各参数进行了重要性排序。在PNA工艺中,出水TN浓度主要受到进水TN浓度及COD浓度的影响。在PDA工艺中,出水TN浓度首先受进水TN浓度及氮负荷的约束。在此基础上,进水COD浓度作为另一重要因素影响着工艺的出水TN浓度。进水COD浓度在2个工艺中的共同重要性表明,2种工艺在实际应用时需要预先做好污废水中碳源的管理与分配,预分离与应用策略非常重要。该研究采用机器学习模型为PNA与PDA工艺脱氮性能的预测提供了方法指导,并基于SHAP的模型解释为2种工艺在实际应用时的关键参数识别与优化提供了选择依据。
- 短程硝化-厌氧氨氧化 /
- 短程反硝化-厌氧氨氧化 /
- 机器学习 /
- SHAP分析 /
- 工艺参数
Abstract: The nitrogen removal performance of the partial nitrification-Anammox (PNA) and partial denitrification-Anammox (PDA) processes are affected by many parameters. Predicting the performance of the two processes and identifying the key parameters based on a comprehensive consideration of various parameters can provide an optimization target for their practical engineering applications. When solving the above problems, experimental methods are usually time-consuming and labor-intensive, while traditional mathematical models are difficult to deal with non-linear relationships. Therefore, in this study, machine learning techniques were used. The constructed Random Forest (RF) machine learning model predicted the effluent nitrogen (TN) concentration of the two processes with high accuracy, and the coefficient of determination (R²) of the PNA and the PDA processes were 0.728 and 0.812, respectively. The SHAP method explained the prediction process of the model well and ranked the importance of each parameter. In the PNA process, the effluent TN concentration was mainly influenced by the influent TN concentration and COD concentration; in the PDA process, the effluent TN concentration was firstly constrained by the influent TN concentration and nitrogen load. On this basis, influent COD concentration is another important factor that affects the effluent TN concentration of the PDA process. The common importance of the influent COD concentration in both processes indicated that both processes should be managed and allocated to the carbon source in the wastewater well in advance of practical application. It is of significant importance to consider the pre-separation and application strategies. The machine learning model used in this study can provide methodological guidance for the prediction of the nitrogen removal performance of the PNA and PDA process. The SHAP-based model interpretation can provide a foundation for the identification and optimization of key parameters for the two processes in practical application.
- partial nitrification-Anammox /
- partial denitrification Anammox /
- machine learning /
- SHAP analysis /
- process parameters

HTML全文

参考文献(45)

[1]	AHMAD H A, AHMAD S, GAO L, et al. Energy-efficient and carbon neutral anammox-based nitrogen removal by coupling with nitrate reduction pathways: a review[J]. Science of The Total Environment, 2023,889: 164213.
[2]	IZADI P, SINHA P, ANDALIB M, et al. Coupling fundamental mechanisms and operational controls in mainstream partial denitrification for partial denitrification anammox applications: a review[J]. Journal of Cleaner Production, 2023,400: 136741.
[3]	WU P, CHEN J, GARLAPATI V K, et al. Novel insights into Anammox-based processes: a critical review[J]. Chemical Engineering Journal, 2022,444: 136534.
[4]	ZHANG X, ZHANG X, CHEN J, et al. A critical review of improving mainstream anammox systems: based on macroscopic process regulation and microscopic enhancement mechanisms[J]. Environmental Research, 2023,236: 116770.
[5]	GHOLAMI-SHIRI J, AZARI M, DEHGHANI S, et al. A technical review on the adaptability of mainstream partial nitrification and anammox: substrate management and aeration control in cold weather[J]. Journal of Environmental Chemical Engineering, 2021,9(6): 106468.
[6]	LAW Y, MATYSIK A, CHEN X, et al. High dissolved oxygen selection against Nitrospira Sublineage Ⅰ in full-scale activated sludge[J]. Environmental Science & Technology, 2019,53(14): 8157-8166.
[7]	郝晓地, 杨万邦, 李季, 等. 厌氧氨氧化技术研究与应用反差现象归因[J]. 环境科学学报, 2023,43(9): 1-13.
[8]	HUAGUANG L, WENYI D, ZILONG Z, et al. Anammox-based technologies for municipal sewage nitrogen removal: advances in implementation strategies and existing obstacles[J]. Journal of Water Process Engineering, 2023,55: 104090.
[9]	YOU Q, WANG J, QI G, et al. Anammox and partial denitrification coupling: a review[J]. RSC Advances, 2020,10(21): 12554-12572.
[10]	GANI K M, AWOLUSI O O, KHAN A A, et al. Potential strategies for the mainstream application of anammox in treatment of anaerobic effluents: a review[J]. Critical Reviews in Environmental Science and Technology, 2021,51(21): 2567-2594.
[11]	朱腾义, 张玉, 程浩淼, 等. 基于可解释性机器学习的纳滤膜去除有机微污染物研究[J]. 环境科学学报, 2023,43(7): 194-203.
[12]	YANG J, CHEN Z, WANG X, et al. Elucidating nitrogen removal performance and response mechanisms of anammox under heavy metal stress using big data analysis and machine learning[J]. Bioresource Technology, 2023,382: 129143.
[13]	LIU S, WU F, GUO M, et al. A comprehensive literature mining and analysis of nitrous oxide emissions from different innovative mainstream anammox-based biological nitrogen removal processes[J]. Science of The Total Environment, 2023,904: 166295.
[14]	LV J, DU L, LIN H, et al. Enhancing effluent quality prediction in wastewater treatment plants through the integration of factor analysis and machine learning[J]. Bioresource Technology, 2024,393: 130008.
[15]	RUDIN C. Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead[J]. Nature Machine Intelligence, 2018,1: 206-215.
[16]	KAO C, LI J, GAO R, et al. Advanced nitrogen removal from real municipal wastewater by multiple coupling nitritation, denitritation and endogenous denitritation with anammox in a single suspended sludge bioreactor[J]. Water Research, 2022,221: 118749.
[17]	SU B, LIU Q, LIANG H, et al. Simultaneous partial nitrification, Anammox, and denitrification in an upflow microaerobic membrane bioreactor treating middle concentration of ammonia nitrogen wastewater with low COD/TN ratio[J]. Chemosphere, 2022,295: 133832.
[18]	THONGSAI A, KRISHNAN S, NOOPHAN P L, et al. Performance and microbial analysis of a fluidized bed membrane bioreactor operated in the partial nitrification and anammox (PN/A) mode for polishing anaerobically treated effluent to industrial discharge standard[J]. Journal of Environmental Chemical Engineering, 2023,11(3): 109808.
[19]	WANG Y, LI B, XUE F, et al. Partial nitrification coupled with denitrification and anammox to treat landfill leachate in a tower biofilter reactor (TBFR)[J]. Journal of Water Process Engineering, 2021,42: 102155.
[20]	XU P, MENG J, LI X, et al. Insights into complete nitrate removal in one-stage nitritation-Anammox by coupling heterotrophic denitrification[J]. Journal of Environmental Management, 2021,298: 113431.
[21]	XU X, QIU L, WANG C, et al. Achieving mainstream nitrogen and phosphorus removal through Simultaneous partial Nitrification, Anammox, Denitrification, and Denitrifying Phosphorus Removal (SNADPR) process in a single-tank integrative reactor[J]. Bioresource Technology, 2019,284: 80-89.
[22]	JI J, PENG Y, WANG B, et al. Synergistic Partial-Denitrification, Anammox, and in-situ Fermentation (SPDAF) Process for Advanced Nitrogen Removal from Domestic and Nitrate-Containing Wastewater[J]. Environmental Science & Technology, 2020,54(6): 3702-3713.
[23]	LI X, DU R, ZHANG J, et al. Verifying the applicability of PD/A unit for ultimate sidestream and mainstream polishing: operating performance, granular characteristics and active microbes[J]. Chemical Engineering Journal, 2023,466: 143300.
[24]	LIN L, LUO Z, ZHANG Y, et al. Partial denitrification-anammox granular sludge domesticated from high-strength anammox granules and the high-efficiency performance in treating low-nitrogen wastewater[J]. Chemical Engineering Journal, 2023,477: 147281.
[25]	LIU Q, PENG Y, ZHAO Y, et al. Excellent anammox performance driven by stable partial denitrification when encountering seasonal decreasing temperature[J]. Bioresource Technology, 2022,364: 128041.
[26]	TAO Y, SHI R, LI L, et al. Performance optimization and nitrogen removal mechanism of up-flow partial denitrification/anammox process[J]. Journal of Environmental Management, 2023,348: 119191.
[27]	WANG D, HE Y, ZHANG X. A comprehensive insight into the functional bacteria and genes and their roles in simultaneous denitrification and anammox system at varying substrate loadings[J]. Applied Microbiology and Biotechnology, 2019,103(3): 1523-1533.
[28]	ZHANG J, PENG Y, LI X, et al. Feasibility of partial-denitrification/anammox for pharmaceutical wastewater treatment in a hybrid biofilm reactor[J]. Water Research, 2022,208: 117856.
[29]	ZHANG L, ZHANG Q, LI X, et al. Enhanced nitrogen removal from municipal wastewater via a novel combined process driven by partial nitrification/anammox (PN/A) and partial denitrification/anammox (PD/A) with an ultra-low hydraulic retention time (HRT)[J]. Bioresource Technology, 2022,363: 127950.
[30]	YE G, WAN J, DENG Z, et al. Prediction of effluent total nitrogen and energy consumption in wastewater treatment plants: bayesian optimization machine learning methods[J]. Bioresource Technology, 2024,395: 130361.
[31]	PARK J, LEE W H, KIM K T, et al. Interpretation of ensemble learning to predict water quality using explainable artificial intelligence[J]. Science of the Total Environment, 2022,832: 155070.
[32]	ZHANG X, LIU Y, LI Z R, et al. Impact of COD/N on anammox granular sludge with different biological carriers[J]. Science of The Total Environment, 2020,728: 138557.
[33]	JIN R C, YANG G F, YU J J, et al. The inhibition of the Anammox process: a review[J]. Chemical Engineering Journal, 2012,197(none): 67-79.
[34]	ZHANG Y, DENG J, XIAO X, et al. Insights on pretreatment technologies for partial nitrification/anammox processes: a critical review and future perspectives[J]. Bioresource Technology, 2023,384: 129351.
[35]	薛同站, 全志道, 李卫华, 等. 短程反硝化强化脱氮的影响因素及其耦合工艺应用进展[J]. 环境工程技术学报, 2024,14(2): 663-671.
[36]	杨京月, 郑照明, 李军, 等. 厌氧氨氧化耦合反硝化底物竞争抑制特性[J]. 中国环境科学, 2018,38(8): 2947-2953.
[37]	LIANG Y, LI Z, ZHANG B, et al. Decryption for nitrogen removal in Anammox-based coupled systems: nitrite-induced mechanisms[J]. Bioresource Technology, 2023,384: 129274.
[38]	TRINH H P, LEE S, JEONG G, et al. Recent developments of the mainstream anammox processes: challenges and opportunities[J]. Journal of Environmental Chemical Engineering, 2021,9(4): 105583.
[39]	WEI C, LI Z, PAN J, et al. An oxic-hydrolytic-oxic process at the nexus of sludge spatial segmentation, microbial functionality, and pollutants removal in the treatment of coking wastewater[J]. ACS ES&T Water, 2021,1(5): 1252-1262.
[40]	TABOADA-SANTOS A, RIVADULLA E, PAREDES L, et al. Comprehensive comparison of chemically enhanced primary treatment and high-rate activated sludge in novel wastewater treatment plant configurations[J]. Water Research, 2020,169: 115258.
[41]	HE X, KE X, WEI T, et al. Process energy and material consumption determined by reaction sequence: from AAO to OHO[J]. Water, 2024,16: 1796.
[42]	CHEN Y, CHEN A, LI Z, et al. O/H/H/O process for total nitrogen removal: an upgrade of the A/A/O process for coking wastewater treatment[J]. ACS ES&T Engineering, 2023,3(9): 1236-1247.
[43]	WEI T, PAN J, KE X, et al. Evaluation of carbon and nitrogen removal performance of the oxic-hydrolytic and denitrification-oxic process in coking wastewater treatment[J]. ACS ES&T Water, 2023,3(1): 236-245.
[44]	LI Z, WEI T, PAN J, et al. Physicochemical pre- and post-treatment of coking wastewater combined for energy recovery and reduced environmental risk[J]. Journal of Hazardous Materials, 2023,447: 130802.
[45]	ZHU S, DENG J, JIN X, et al. Diverse and distinct bacterial community involved in a full-scale A/O1/H/O2 combination of bioreactors with simultaneous decarbonation and denitrogenation of coking wastewater[J]. Environmental Science and Pollution Research, 2022,30.