The accurate prediction of effluent quality, which is of great significance for sewage treatment plants, helps optimize the treatment process and supports process control to achieve compliance. In this paper, the artificial gorilla army optimization algorithm (GTO) is used to optimize the hybrid deep learning model BiLSTM-GTO, which is a bidirectional long short-term memory network (BiLSTM) with optimized hyperparameters, so as to realize the prediction of effluent quality of sewage treatment plants. The hybrid model is based on BiLSTM and implements two-way modeling of time series so that it can utilize the context information before and after. In order to improve the parameter optimization ability, the GTO global search mechanism is introduced to optimize the key hyperparameters of BiLSTM so that the prediction performance can be improved. So that the accuracy and robustness of the BiLSTM-GTO model can be verified, the actual operation data of a sewage treatment plant in Shaanxi Province and the simulation data of the benchmark simulation model (BSM2) were used to verify the proposed method. The experimental results show that the proposed method exhibited performance advantages in predicting both chemical oxygen demand (COD) and total nitrogen (TN) compared with the bench mark models. In real data, the determination coefficients (R2) of COD and TN reached 0.932 5 and 0.912 4 respectively; In simulated data, the improvement under heavy rain conditions was more significant than that under other weather conditions, with R2 values of COD and TN reaching 0.927 1 and 0.915 9 respectively. Thus, the feasibility and effectiveness of the BiLSTM-GTO model have been verified.
HANH, LIUY, QIAOJ. Mechanism-data-driven multiobjective optimization for wastewater treatment process[J]. IEEE Transactions on Industrial Informatics,2024,20(5):7810-7819.
[2]
LIR, FENGK, ANT,et al. Enhanced insights into effluent prediction in wastewater treatment plants:Comprehensive deep learning model explanation based on shap[J]. ACS ES&T Water,2024,4(4):1904-1915.
[3]
JIANGZ, XIAZ, LIUS,et al. The effect of fine grits and fine debris concentrations on the MLVSS/MLSS ratio of an activated sludge system[J]. Journal of Environmental Sciences,2025,147:607-616.
[4]
LIUD, SUNK. Random forest solar power forecast based on classification optimization[J]. Energy,2019,187:115940.
[5]
JIANGY, ZHENGL, DINGX. Ultra-short-term prediction of photovoltaic output based on an LSTM-ARMA combined model driven by EEMD[J]. Journal of Renewable and Sustainable Energy,2021,13(4):046103.
[6]
ZHAON, LIC, JIAH,et al. Acoustic emission-based flow noise detection and mechanism analysis for gas-liquid two-phase flow[J]. Measurement,2021,179:109480.
[7]
SONIS, KHURSHIDA, MINASEA M,et al. A TinyML Approach for Quantification of BOD and COD in Water[C]//2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems,Machine Learning and Signal Processing (PCEMS). Piscataway:IEEE,2023:1-6.
[8]
MEKAOUSSIH, HEDDAMS, BOUSLIMANNIN,et al. Predicting biochemical oxygen demand in wastewater treatment plant using advance extreme learning machine optimized by Bat algorithm[J]. Heliyon,2023,9(11):e21351.
[9]
LIM, CHENGH, WANGX,et al. Research on soft sensor modeling of support vector machine for wastewater treatment[C]//2023 8th International Conference on Automation,Control and Robotics Engineering (CACRE). Piscataway:IEEE,2023:162-166.
[10]
ZHAOY, LIH, ZHOUH,et al. A review of graph neural network applications in mechanics-related domains[J]. Artificial Intelligence Review,2024,57(11):315.
[11]
WAQASM, HUMPHRIESU W. A critical review of RNN and LSTM variants in hydrological time series predictions[J]. MethodsX,2024,13:102946.
[12]
ZHANGY, LIC, DUANH,et al. Deep learning based data-driven model for detecting time-delay water quality indicators of wastewater treatment plant influent[J]. Chemical Engineering Journal,2023,467:143483.
[13]
TAWHIDM A, SAVSANIV. Multi-objective sine-cosine algorithm (MO-SCA) for multi-objective engineering design problems[J]. Neural Computing and Applications,2019,31(2):915-929.
[14]
SALAMATTALABM M, ZONOOZIM H, MOLAVI-ARABSHAHIM. Innovative approach for predicting biogas production from large-scale anaerobic digester using long-short term memory (LSTM) coupled with genetic algorithm (GA)[J]. Waste Management,2024,175:30-41.
[15]
DEBK, AGRAWALS. A niched-penalty approach for constraint handling in genetic algorithms[C]//Artificial Neural Nets and Genetic Algorithms:Proceedings of the International Conference in Portorož. Vienna:Springer-Verlag,1999:235-243.
[16]
LIUX, SHIQ, LIUZ,et al. Using LSTM neural network based on improved PSO and attention mechanism for predicting the effluent COD in a wastewater treatment plant[J]. IEEE Access,2021,9:146082-146096.
[17]
KENNEDYJ, EBERHARTR C. A discrete binary version of the particle swarm algorithm[C]//1997 IEEE International conference on systems,man,and cybernetics. Computational cybernetics and simulation. Piscataway:IEEE,1997:4104-4108.
[18]
VRECKOD, GERNAEYK V, ROSENC,et al. Benchmark simulation model No 2 in Matlab-Simulink:towards plant-wide WWTP control strategy evaluation[J]. Water Science and Technology,2006,54(8):65-72.
[19]
JIANGJ, ZHANGX, YUANZ. Feature selection for classification with Spearman’s rank correlation coefficient-based self-information in divergence-based fuzzy rough sets[J]. Expert Systems with Applications,2024,249:123633.