为提升流域水污染管控水平,弥补现有入河污染物优化减排技术的不足,以受纳水体为研究对象,耦合双向长短记忆网络(Bi-LSTM)与贝叶斯优化(BO)算法,构建了基于数据驱动理论的河流水污染优化消减模型.将该模型应用于松花江流域摆渡镇/三道-宏克利段,结果表明:Bi-LSTM算法可有效训练水质参数空间网络拓扑结构,提升水质参数空间响应关系学习精度.当下游目标断面执行Ⅲ类水体标准时,摆渡镇和三道TN的削减率分别为[14.12%, 38.84%]和[15.01%, 38.98%],TP几乎不需要削减.当执行Ⅱ类水体标准时,摆渡镇TN、TP的削减率分别为[19.08%, 39.72%]和[0.00%, 41.93%],三道TN、TP的削减率分别为[18.43%, 40.09%]和[0.00%, 36.24%].该消减模型可以提出不同消减情景下的污染物最优削减策略,为河流水污染精准消减和智能管控提供决策支持.
Abstract
In order to improve the water pollution control level in the basin and solve the shortcomings of existing optimization emission reduction technologies, a data driven theory-based optimization emission reduction model for river pollution was developed with the combination of Bidirectional Long and Short Memory Network (Bi-LSTM) algorithm and Bayesian Optimization (BO) algorithm considering the receiving water body as the research object. The developed optimization emission reduction model was then applied to Songhua River basin from Baidu Town and Sandao to Hongkeli section. The results show that the Bi-LSTM algorithm can effectively train the spatial neural network architecture, and improve the study accuracy of water quality spatial relationship. The reduction rates of TN in Baidu Town and Sandao are [14.12%, 38.84%] and [15.01%, 38.98%], respectively, and TP hardly needs to be reduced under the water quality standard III. While the reduction rates of TN and TP in Baidu Town are [19.08%, 39.72%] and [0.00%, 41.93%], and the ones in Sandao are [18.43%, 40.09%] and [0.00%, 36.24%], respectively under the water quality standard II. The optimal reduction strategies can be obtained by the developed optimization emission reduction model under different reduction scenarios to provide decision support for accurate reduction and intelligent control of river pollution.
关键词
河流水污染 /
精准减排 /
智能管控 /
双向长短记忆网络 /
贝叶斯优化
Key words
river pollution /
precise emission reduction /
intelligent control /
bidirectional long and short memory network /
bayesian optimization
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 史斌,姜继平,王鹏.基于高频在线水质数据异常的突发污染预警[J]. 中国环境科学, 2017,37(11):4394-4400. Shi B, Jing J P, Wang P. Early warning of water pollution incidents based on abnormal change of water quality data from high frequency online monitoring[J]. China Environmental Science, 2017,37(11): 4394-4400.
[2] 陈颖珊.变化环境下基于农业面源污染与经济效益协同的水土资源优化配置[D]. 哈尔滨:东北农业大学, 2023. Chen Y S. Optimal allocation of soil and water resources based on synergistic agricultural surface pollution and economic benefits in a changing environment[D]. Harbin: Northeast Agricultural University, 2023.
[3] Xie Y L, Huang G H. A multi-period optimization model for energy planning with CO2 emission consideration[J]. Journal of Environmental Management, 2010,91(5):1063-1070.
[4] 胡雪红.耦合贝叶斯的两阶段随机优化方法在水质管理中的应用研究[D]. 北京:华北电力大学, 2016. Hu X H. A Bayesian-based two-stage inexact optimization method for water quality management[D]. Beijing: North China Electric Power University, 2016.
[5] Cao D, Chan M, Ng S. Modeling and forecasting of nano FeCu treated sewage quality using recurrent neural network (RNN)[J]. Computation, 2023,11(2):39.
[6] 徐桂烽.基于LSTM混合模型的河流中多环芳烃含量预测研究[D]. 重庆:重庆大学, 2022. Xu G F. Prediction research of polycyclic aromatic hydrocarbons in rivers based on LSTM hybrid model[D]. Chongqing: Chongqing University, 2022.
[7] Chen S Y, Huang J L, Wang P, et al. A coupled model to improve river water quality prediction towards addressing non-stationarity and data limitation[J]. Water Research, 2024,248:120895.
[8] Ruan J L, Cui Y, Song Y C, et al. A novel RF-CEEMD-LSTM model for predicting water pollution[J]. Scientific Reports, 2023,13:20901.
[9] Zheng L, Wang H P, Liu C, et al. Prediction of harmful algal blooms in large water bodies using the combined EFDC and LSTM models[J]. Journal of Environmental Management, 2021,295:113060.
[10] 汪昊,高凯璐,尹高飞,等.海河流域河流氮素变化趋势及关键因子识别[J]. 中国环境科学, 2025,45(7):3968-3977. Wang H, Gao K L, Yin G F, et al. Identification of river nitrogen trends and key factors in the Haihe River[J]. China Environmental Science, 2025,45(7):3968-3977.
[11] 商俊蓓.基于双向长短时记忆递归神经网络的联机手写数字公式字符识别[D]. 广州:华南理工大学, 2015. Shang J B. Online handwritten math expression label recognition based on long short term memory recurrent neural network[d]. Guangzhou: South China University of Technology, 2015.
[12] Chen D L, Luo Y Q, Yuan X G. Cascade refrigeration system synthesis based on hybrid simulated annealing and particle swarm optimization algorithm[J]. Chinese Journal of Chemical Engineering, 2023,58:244- 255.
[13] Cho J H, Sung K S, Ha S R. A river water quality management model for optimising regional wastewater treatment using a genetic algorithm[J]. Journal of Environmental Management, 2004,73(3):229-242.
[14] Zacharias C R, Lemes M R, Pino Jr D A. Combining genetic algorithm and simulated annealing: a molecular geometry optimization study[J]. Journal of Molecular Structure: THEOCHEM, 1998,430:29-39.
[15] 许琳娟,褚俊英,周祖昊,等.松花江流域水环境质量特征分析[J]. 水资源保护, 2012,28(6):55-58. Xu L J, Chu J Y, Zhou Z H, et al. Analysis of water environmental quality features in Songhua River Basin[J]. Water Resources Protection, 2012,28(6):55-58.
[16] 李玮,褚俊英,秦大庸,等.松花江流域水污染特征及其调控对策[J]. 中国水利水电科学研究院学报, 2010,8(3):229-232. Li W, Chu J Y, Qin D Y, et al. The water pollution features of Songhua River basin and the regulation strategy[J]. Journal of China Institute of Water Resources and Hydropower Research, 2010,8(3):229-232.
[17] Naheliya B, Redhu P, Kumar K. MFOA-Bi-LSTM: An optimized bidirectional long short-term memory model for short-term traffic flow prediction[J]. Physica A: Statistical Mechanics and its Applications, 2024,634:129448.
[18] Alfwzan W F, Selim M M, Almalki A S, et al. Water quality assessment using Bi-LSTM and computational fluid dynamics (CFD) techniques[J]. Alexandria Engineering Journal, 2024,97:346-359.
[19] Singla P, Duhan M, Saroha S. An ensemble method to forecast 24-h ahead solar irradiance using wavelet decomposition and BiLSTM deep learning network[J]. Earth Science Informatics, 2022,15:291-306.
[20] Wang Q L, Han J Q, Chen F, et al. Real-time risk prediction of chemical processes based on attention-based Bi-LSTM[J]. Chinese Journal of Chemical Engineering, 2024,75:131-141.
[21] Yin J, Deng Z, Ines A V M, et al. Forecast of short-term daily reference evapotranspiration under limited meteorological variables using a hybrid bi-directional long short-term memory model (Bi-LSTM)[J]. Agricultural Water Management, 2020,242:106386.
[22] Wang K F, Liu L, Ben X C, et al. Hybrid deep learning based prediction for water quality of plain watershed[J]. Environmental Research, 2024,262(2):119911.
[23] Khullar S, Singh N. Water quality assessment of a river using deep learning Bi-LSTM methodology: forecasting and validation[J]. Environmental Science and Pollution Research, 2022,29:12875- 12889.
[24] Sadoune H, Rihani R, Marra F S. DNN model development of biogas production from an anaerobic wastewater treatment plant using Bayesian hyperparameter optimization[J]. Chemical Engineering Journal, 2023,471:144671.
[25] 李涛宇,许秀春,杨轩,等.利用机器学习预测华北地区冬小麦农田氮淋失[J]. 中国环境科学, 2025,45(1):343-354. Li T Y, Xu X C, Yang X, et al. Prediction of nitrogen leaching from winter-wheat production in North China based on random forest and XGBoost[J]. China Environmental Science, 2025,45(1):343-354.
[26] Kitahara M, Dang C, Beer M. Bayesian updating with two-step parallel Bayesian optimization and quadrature[J]. Computer Methods in Applied Mechanics and Engineering, 2023,403:115735.
[27] Ma J F, Zhang J, Li R N, et al. Using Bayesian optimization to automate the calibration of complex hydrological models: Framework and application[J]. Environmental Modelling & Software, 2022,147: 105235.
[28] Morita Y, Rezaeiravesh S, Tabatabaei N, et al. Applying Bayesian optimization with Gaussian process regression to computational fluid dynamics problems[J]. Journal of Computational Physics, 2022,449: 110788.
[29] Uddin M G, Bamal A, Diganta M T M, et al. The role of optimizers in developing data-driven model for predicting lake water quality incorporating advanced water quality model[J]. Alexandria Engineering Journal, 2025,122:411-435.
[30] Shams R, Alimohammadi S, Yazdi J. Optimizing surfactant-enhanced aquifer remediation based on Gaussian process surrogate model in DNAPL-contaminated sites considering different wells patterns[J]. Groundwater for Sustainable Development, 2021,15:100675.
[31] 国务院.国务院关于松花江流域水污染防治规划(2006~2010年)的批复[EB/OL]. https://www.gov.cn/gongbao/content/2006/content_ 413980.htm, 2006-08-23. State Council. The State Council's approval of the water pollution prevention and control plan for Songhua River Basin (2006~2010)[EB/OL].
[32] GB3838-2002地表水环境标准[S]. GB3838-2002 Surface water environmental standards[S].
[33] Jiang D X, Zhu H, Liu J. Dynamic surface river pollution identification by a hybrid multivariate-based anomaly detection algorithm[J]. Journal of Cleaner Production, 2024,467:142923.
基金
黑龙江省自然科学基金联合引导项目(LH2023E030);中国博士后科学基金项目(2022T150105)
PDF(2286 KB)
京公网安备 11010802030352号 
