Optimizing water reuse in integrated refining and petrochemical enterprises: high-precision prediction of water quality enabling a novel proactive warning index

Abstract

To address the challenges of low prediction accuracy and limited generalization capability in forecasting complex water quality at refineries, this study proposes a novel hybrid neural network model (CBG). This model integrates convolutional neural networks, bidirectional long short-term memory networks, and grey wolf optimization algorithms. The CBG model demonstrates excellent accuracy in predicting key pollutants such as chemical oxygen demand (COD), oil, and ammonia nitrogen (NH₃-N). Its correlation coefficients reach 0.95, 0.89, and 0.91 respectively, and the nash sutcliffe efficiency coefficients stand at 0.91, 0.79, and 0.83 respectively, which are significantly superior to those of other benchmark models. Additionally, the study innovatively developed a comprehensive warning water quality index (WWQI). This index, together with the CBG model, forms an integrated prediction and warning framework that triggers alerts when water quality indices exceed pre-set thresholds. This framework provides a valuable tool for the early detection and proactive intervention of risks within the water systems of integrated refining and petrochemical enterprises. This study holds significant practical implications for enhancing water resource utilization and maintaining the stability of production operations. By providing intelligent early warning and proactive risk management tools, this research contributes to improving the operational safety and resource efficiency of industrial water circulation systems. This comprehensive approach provides a clear, quantifiable method for forward-thinking, science-based decision-making in water systems management for integrated refining and petrochemical enterprises, ultimately helping to drive more sustainable industrial practices.

Graphical abstract

Keywords

Deep learning / Convolutional neural network / Bidirectional long short-term memory network / Grey wolf optimizer / Warning water quality index

Highlight

	● Hybrid deep learning for spatio-temporal analysis of complex wastewater.
	● A weighted water quality index, WWQI, enables proactive, integrated risk warning.
	● Model interpretation identifies key drivers to construct a weighted warning index.
	● Robust operational thresholds are validated by both historical and simulated data.

References

[1] Al-Dahidi S , Alrbai M , Al-Ghussain L , Alahmer A , Hayajneh H S . (2024). Data-driven analysis and prediction of wastewater treatment plant performance: insights and forecasting for sustainable operations. Bioresource Technology, 391: 129937

[2] Boobier S , Hose D R J , Blacker A J , Nguyen B N . (2020). Machine learning with physicochemical relationships: solubility prediction in organic solvents and water. Nature Communications, 11(1): 5753

[3] Cabrera S M , Winnubst L , Richter H , Voigt I , McCutcheon J , Nijmeijer A . (2022). Performance evaluation of an industrial ceramic nanofiltration unit for wastewater treatment in oil production. Water Research, 220: 118593

[4] Chen J , Wan J Q , Ye G , Wang Y . (2024). Prediction and optimization of wastewater treatment process effluent chemical oxygen demand and energy consumption based on typical ensemble learning models. Bioresource Technology, 411: 131362

[5] Chen K Y , Chen H X , Zhou C L , Huang Y C , Qi X Y , Shen R Q , Liu F R , Zuo M , Zou X Y , Wang J F . et al. (2020). Comparative analysis of surface water quality prediction performance and identification of key water parameters using different machine learning models based on big data. Water Research, 171: 115454

[6] Chen Y , Qiu R , Wang J Q , Chen P , Zheng M , Guo H G . (2025). Synergistic efficiency in greenhouse gas emission reduction and water pollution control: evaluating policy impacts in China. Frontiers of Environmental Science & Engineering, 19(10): 132

[7] Dicataldo G , Desmond P , Al-Maas M , Adham S . (2025). Feasibility and application of membrane aerated biofilm reactors for industrial wastewater treatment. Water Research, 280: 123523

[8] Emary E , Zawbaa H M , Grosan C . (2018). Experienced gray wolf optimization through reinforcement learning and neural networks. IEEE Transactions on Neural Networks and Learning Systems, 29(3): 681–694

[9] Fatima S A , Ramli N , Taqvi S A A , Zabiri H . (2021). Prediction of industrial debutanizer column compositions using data-driven ANFIS- and ANN-based approaches. Neural Computing and Applications, 33(14): 8375–8387

[10] Fu W L , Feng M H , Guo C B , Zhou J E , Zhang X Y , Lv S Y , Huo Y Q , Wang F . (2024). Machine learning-driven prediction of phosphorus removal performance of metal-modified biochar and optimization of preparation processes considering water quality management objectives. Bioresource Technology, 403: 130861

[11] Gao K X , Yang Y , Li A , Pu J , Takizawa S , Graham N J D , Hou L A . (2024). Fouling behavior of BTEX in petrochemical wastewater treated by nanofiltration (NF). Journal of Hazardous Materials, 476: 135185

[12] Hashemi F , Hashemi H , Abbasi A , Schreiber M E . (2022). Life cycle and economic assessments of petroleum refineries wastewater recycling using membrane, resin and on site disinfection (UF-IXMB-MOX) processes. Process Safety and Environmental Protection, 162: 419–425

[13] Hashemi F , Hashemi H , Shahbazi M , Dehghani M , Hoseini M , Shafeie A . (2020). Reclamation of real oil refinery effluent as makeup water in cooling towers using ultrafiltration, ion exchange and multioxidant disinfectant. Water Resources and Industry, 23: 100123

[14] Helm W , Zhong S F , Reid E , Igou T , Chen Y S . (2024). Development of gradient boosting-assisted machine learning data-driven model for free chlorine residual prediction. Frontiers of Environmental Science & Engineering, 18(2): 17

[15] Hochreiter S , Schmidhuber J . (1997). Long short-term memory. Neural Computation, 9(8): 1735–1780

[16] Hou B J , Zhou Z H . (2020). Learning with interpretable structure from gated RNN. IEEE Transactions on Neural Networks and Learning Systems, 31(7): 2267–2279

[17] Hu J T , Fu W Y , Ni F , Zhang X H , Yang C P , Sang J Q . (2020). An integrated process for the advanced treatment of hypersaline petrochemical wastewater: a pilot study. Water Research, 182: 116019

[18] Huang Y T , Yuan B X , Wang X Q , Dai Y S , Wang D M , Gong Z J , Chen J M , Shen L , Fan M K , Li Z L . (2023). Industrial wastewater source tracing: the initiative of SERS spectral signature aided by a one-dimensional convolutional neural network. Water Research, 232: 119662

[19] Huang Y X , Bu L J , Zhu S M , Zhou S Q . (2024). Integration of nontarget analysis with machine learning modeling for prioritization of odorous volatile organic compounds in surface water. Journal of Hazardous Materials, 471: 134367

[20] Khan U A , Löffler P , Spilsbury F , Wiberg K , Stålsby Lundborg C , Lai F Y . (2024). Towards sustainable water reuse: a critical review and meta-analysis of emerging chemical contaminants with risk-based evaluation, health hazard prediction and prioritization for assessment of effluent water quality. Journal of Hazardous Materials, 480: 136175

[21] Koksal E S , Aydin E . (2025). A hybrid approach of transfer learning and physics-informed modelling: improving dissolved oxygen concentration prediction in an industrial wastewater treatment plant. Chemical Engineering Science, 304: 121088

[22] Li Z L , Liu H X , Zhang C , Fu G T . (2024). Real-time water quality prediction in water distribution networks using graph neural networks with sparse monitoring data. Water Research, 250: 121018

[23] Lin S B , Kim J , Hua C B , Park M H , Kang S . (2023). Coagulant dosage determination using deep learning-based graph attention multivariate time series forecasting model. Water Research, 232: 119665

[24] Liu X , Yue F J , Guo T L , Li S L . (2024). High-frequency data significantly enhances the prediction ability of point and interval estimation. Science of the Total Environment, 912: 169289

[25] Ma W C , Zhang S B , Deng L M , Zhong D , Li K F , Liu X T , Li J X , Zhang J N , Ma J . (2023). Cu-based perovskite as a novel CWPO catalyst for petroleum refining wastewater treatment: performance, toxicity and mechanism. Journal of Hazardous Materials, 448: 130824

[26] Mao J E , Chen H Y , Xu X Y , Zhu L . (2024). Assessing greenhouse gas emissions from the printing and dyeing wastewater treatment and reuse system: potential pathways towards carbon neutrality. Science of the Total Environment, 927: 172301

[27] Monje V , Owsianiak M , Junicke H , Kjellberg K , Gernaey K V , Flores-Alsina X . (2022). Economic, technical, and environmental evaluation of retrofitting scenarios in a full-scale industrial wastewater treatment system. Water Research, 223: 118997

[28] Nong X Z , Lai C , Chen L H , Wei J H . (2024). A novel coupling interpretable machine learning framework for water quality prediction and environmental effect understanding in different flow discharge regulations of hydro-projects. Science of the Total Environment, 950: 175281

[29] Pang H J , Ben Y , Cao Y , Qu S , Hu C Z . (2025). Time series-based machine learning for forecasting multivariate water quality in full-scale drinking water treatment with various reagent dosages. Water Research, 268: 122777

[30] Peng C , Wu Z M , Zhang S D , Lin B R , Nie L , Tian W L , Zang H C . (2025). Online monitoring of water quality in industrial wastewater treatment process based on near-infrared spectro-scopy. Water Research, 275: 123165

[31] Peng L , Wu H , Gao M , Yi H L , Xiong Q Y , Yang L D , Cheng S P . (2022). TLT: recurrent fine-tuning transfer learning for water quality long-term prediction. Water Research, 225: 119171

[32] Pyo J , Park L J , Pachepsky Y , Baek S S , Kim K , Cho K H . (2020). Using convolutional neural network for predicting cyanobacteria concentrations in river water. Water Research, 186: 116349

[33] Sang S , Li L . (2024). A novel variant of LSTM stock prediction method incorporating attention mechanism. Mathematics, 12(7): 945

[34] Santos A V , Amaral M C S , Oliveira S M A C . (2024). Artificial Neural Networks and Multivariate Statistical Process Control to improve ammonia removal on membrane bioreactors treating refinery wastewater. Journal of Water Process Engineering, 67: 106126

[35] Tarpani R R Z , Azapagic A . (2023). Life cycle sustainability assessment of advanced treatment techniques for urban waste-water reuse and sewage sludge resource recovery. Science of the Total Environment, 869: 161771

[36] Tian Y Q , Wen Z G , Zhao Y H . (2025). Novel knowledge for identifying point pollution sources in watershed environmental management. Water Research, 275: 123168

[37] Visser H , Evers N , Bontsema A , Rost J , de Niet A , Vethman P , Mylius S , van der Linden A , van den Roovaart J , van Gaalen F . et al. (2022). What drives the ecological quality of surface waters? A review of 11 predictive modeling tools. Water Research, 208: 117851

[38] Wang J H , Zhang Y T , Li C L , Duan H P , Wang W H . (2025). Predicting dissolved oxygen in water areas using transfer learning and visual information from real-time surveillance videos. Journal of Cleaner Production, 507: 145547

[39] Wang X Y , Tang X Y , Zhu M , Liu Z N , Wang G Q . (2024). Predicting abrupt depletion of dissolved oxygen in Chaohu lake using CNN-BiLSTM with improved attention mechanism. Water Research, 261: 122027

[40] Wu C Y , Zhou Y X , Sun Q L , Fu L Y , Xi H B , Yu Y , Yu R Z . (2016). Appling hydrolysis acidification-anoxic–oxic process in the treatment of petrochemical wastewater: from bench scale reactor to full scale wastewater treatment plant. Journal of Hazardous Materials, 309: 185–191

[41] Xie Y F , Chen Y Q , Wei Q , Yin H L . (2024). A hybrid deep learning approach to improve real-time effluent quality prediction in wastewater treatment plant. Water Research, 250: 121092

[42] Xu J , Lai Z L , Zhang W , Liu T C , Xiao S Z , Yang L B , Yu Z J , Zhou X F . (2025). Moving bed biofilm reactor for blackwater treatment: insights into pollutant removal, microbial communities, and water quality prediction through machine learning. Frontiers of Environmental Science & Engineering, 19(8): 102

[43] Zandavi S M , Chung V Y Y , Anaissi A . (2021). Stochastic dual simplex algorithm: a novel heuristic optimization algorithm. IEEE Transactions on Cybernetics, 51(5): 2725–2734

[44] Zanoni M G , Majone B , Bellin A . (2022). A catchment-scale model of river water quality by Machine Learning. Science of the Total Environment, 838: 156377

[45] Zhang Y T , Wang J H , Li C L , Duan H P , Wang W H . (2025). Attention-based deep learning models for predicting anomalous shock of wastewater treatment plants. Water Research, 275: 123192

RIGHTS & PERMISSIONS

Higher Education Press 2026