Please wait a minute...
Shopping Cart Email List Chinese
Advanced Search Fig/Tab
Frontiers of Environmental Science & Engineering

ISSN 2095-2201

ISSN 2095-221X(Online)

CN 10-1013/X

Postal Subscription Code 80-973

2018 Impact Factor: 3.883

Featured Articles  |  Most Downloaded  |  Most Reading
Online Submission Subscribe to Our RSS Content Feed
Front. Environ. Sci. Eng.    2023, Vol. 17 Issue (6) : 67    https://doi.org/10.1007/s11783-023-1667-3
RESEARCH ARTICLE
Water quality soft-sensor prediction in anaerobic process using deep neural network optimized by Tree-structured Parzen Estimator
Junlang Li1, Zhenguo Chen1, Xiaoyong Li1, Xiaohui Yi1, Yingzhong Zhao2, Xinzhong He2, Zehua Huang2, Mohamed A. Hassaan3, Ahmed El Nemr3, Mingzhi Huang1,4()
1. SCNU Environmental Research Institute, Guangdong Provincial Key Laboratory of Chemical Pollution and Environmental Safety & MOE Key Laboratory of Theoretical Chemistry of Environment, School of Environment, South China Normal University, Guangzhou 510006, China
2. Fujian Environmental Protection Design Institute Co. Ltd, Fuzhou 350000, China
3. National Institute of Oceanography and Fisheries, NIOF, Alexandria 21556, Egypt
4. SCNU Qingyuan Institute of Science and Technology Innovation Co., Ltd., Qingyuan 511517, China
 Download: PDF(8318 KB)   HTML
 Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks
Abstract

● Hybrid deep-learning model is proposed for water quality prediction.

● Tree-structured Parzen Estimator is employed to optimize the neural network.

● Developed model performs well in accuracy and uncertainty.

● Usage of the proposed model can reduce carbon emission and energy consumption.

Anaerobic process is regarded as a green and sustainable process due to low carbon emission and minimal energy consumption in wastewater treatment plants (WWTPs). However, some water quality metrics are not measurable in real time, thus influencing the judgment of the operators and may increase energy consumption and carbon emission. One of the solutions is using a soft-sensor prediction technique. This article introduces a water quality soft-sensor prediction method based on Bidirectional Gated Recurrent Unit (BiGRU) combined with Gaussian Progress Regression (GPR) optimized by Tree-structured Parzen Estimator (TPE). TPE automatically optimizes the hyperparameters of BiGRU, and BiGRU is trained to obtain the point prediction with GPR for the interval prediction. Then, a case study applying this prediction method for an actual anaerobic process (2500 m3/d) is carried out. Results show that TPE effectively optimizes the hyperparameters of BiGRU. For point prediction of CODeff and biogas yield, R2 values of BiGRU, which are 0.973 and 0.939, respectively, are increased by 1.03%–7.61% and 1.28%–10.33%, compared with those of other models, and the valid prediction interval can be obtained. Besides, the proposed model is assessed as a reliable model for anaerobic process through the probability prediction and reliable evaluation. It is expected to provide high accuracy and reliable water quality prediction to offer basis for operators in WWTPs to control the reactor and minimize carbon emission and energy consumption.
Keywords Water quality prediction      Soft-sensor      Anaerobic process      Tree-structured Parzen Estimator     
Corresponding Author(s): Zhenguo Chen,Mingzhi Huang   
Issue Date: 22 December 2022
 Cite this article:   
Junlang Li,Zhenguo Chen,Xiaoyong Li, et al. Water quality soft-sensor prediction in anaerobic process using deep neural network optimized by Tree-structured Parzen Estimator[J]. Front. Environ. Sci. Eng., 2023, 17(6): 67.
 URL:  
https://academic.hep.com.cn/fese/EN/10.1007/s11783-023-1667-3
https://academic.hep.com.cn/fese/EN/Y2023/V17/I6/67
Fig.1  Structure of proposed network and algorithm: (a) Structure of GRU; (b) A full connected network after using dropout algorithm; (c) The whole structure of BiGRU.
Fig.2  Flowchart of construction process of proposed model. The blue arrows represent the construction process, and the red arrows represent dataset combination. X denotes the features, Y denotes the labels. The superscript “tr” and “te” respectively denote training set and test set. The subscript “1” and “2” respectively denote the first and the second.
Fig.3  A heatmap with correlation analyze.
Optimization methods Max R2 Average R2 Variance of R2 p value of the t-test
TPE 0.975 0.954 0.000297 0.0127
RS 0.970 0.935 0.000732
Tab.1  Comparison of R2 over the last 30 trials between TPE and RS
Fig.4  Comparison over 50 trials by TPE and RS: (a)–(e) Hyperparameters selection; (f) Training accuracy. Grey area is the initial iterations. Red denotes TPE and blue denotes RS.
Feature Model Point prediction Interval prediction Probability prediction Test time (s)
R2 RMSE* MAPE (%) CP MWP MC CRPS
CODeff BiGRU-GPR 0.973 24.94 2.38 0.93 0.12 0.13 0.0329 2093
BiLSTM-GPR 0.963 30.31 2.84 0.97 0.15 0.15 0.0412 2356
BiRNN-GPR 0.899 48.54 5.22 1 0.24 0.24 0.0504 1769
Biogas BiGRU-GPR 0.939 169.32 4.66 0.9 0.24 0.27 0.0988 1953
BiLSTM-GPR 0.927 184.89 5.26 0.97 0.31 0.31 0.110 2238
BiRNN-GPR 0.842 275.37 8.06 0.93 0.39 0.42 0.186 1641
Tab.2  Metrics of three models.
Fig.5  Point prediction results of three models. (a) CODeff; (b) Biogas.
Fig.6  Interval prediction results of three models. (a) CODeff; (b) Biogas.
Fig.7  Reliability evaluation of BiGRU-GPR. The red area denotes Kolmogorov 5% significance band, red diagonal line is the theoretical uniform distribution and the points are PIT values. (a) CODeff; (b) Biogas.
WWTPs Wastewater Treatment Plants
SVM Support Vector Machine
RF Random Forest
GPR Gaussian Process Regression
ANN Artificial Neural Network
RNN Recurrent Neural Network
LSTM Long Short-Time Memory
GRU Gated Recurrent Unit
BiGRU Bidirectional Gated Recurrent Unit
TPE Tree-structured Parzen Estimator
RS Random Search
GS Grid Search
RBF Radial Basis Function
SE Square Exponential Covariance function
RQ Rational Quadratic
ALK Alkalinity
OLR Organic Loading Rate
HRT Hydraulic Retention Time
COD Chemical Oxygen Demand
MSE Mean Square Error
RMSE Root Mean Square Error
MAPE Mean Absolute Percentage Error
CP Coverage Percentage
MWP Mean Width Percentage
CRPS Continuous Ranked Probability Score
CDF Cumulative Distribution Function
PDF Probability Density Function
PIT Probability Integral Transform
  
1 Q L Chen, W Chai, J F Qiao. (2010). Modeling of Wastewater Treatment Process Using Recurrent Neural Network. Jinan: IEEE, 5872–5876
2 P M L Ching, R H Y So, T Morck. (2021). Advances in soft sensors for wastewater treatment plants: a systematic review. Journal of Water Process Engineering, 44: 102367
https://doi.org/10.1016/j.jwpe.2021.102367
3 H Darvishi, D Ciuonzo, E R Eide, P S Rossi. (2021). Sensor-fault detection, isolation and accommodation for digital twins via modular data-driven architecture. IEEE Sensors Journal, 21(4): 4827–4838
https://doi.org/10.1109/JSEN.2020.3029459
4 F Di Maria, C Micale. (2015). The contribution to energy production of the aerobic bioconversion of organic waste by an organic Rankine cycle in an integrated anaerobic-aerobic facility. Renewable Energy, 81: 770–778
https://doi.org/10.1016/j.renene.2015.03.087
5 C A T Ferro. (2014). Fair scores for ensemble forecasts. Quarterly Journal of the Royal Meteorological Society, 140(683): 1917–1923
https://doi.org/10.1002/qj.2270
6 H G Han, J C Zhang, S L Du, H Y Sun, J F Qiao. (2021). Robust optimal control for anaerobic-anoxic-oxic reactors. Science China. Technological Sciences, 64(7): 1485–1499
https://doi.org/10.1007/s11431-020-1821-2
7 M Hauck, F A Maalcke-Luesken, M S M Jetten, M A J Huijbregts (2016). Removing nitrogen from wastewater with side stream anammox: What are the trade-offs between environmental impacts? Resources, Conservation and Recycling, 107: 212–219
https://doi.org/10.1016/j.resconrec.2015.11.019
8 B Heydari, E A Sharghi, S Rafiee, S S Mohtasebi. (2021). Use of artificial neural network and adaptive neuro-fuzzy inference system for prediction of biogas production from spearmint essential oil wastewater treatment in up-flow anaerobic sludge blanket reactor. Fuel, 306: 121734
https://doi.org/10.1016/j.fuel.2021.121734
9 S Hochreiter, J Schmidhuber. (1997). Long short-term memory. Neural Computation, 9(8): 1735–1780
https://doi.org/10.1162/neco.1997.9.8.1735 pmid: 9377276
10 Y Jiang, S Yin, J Dong, O Kaynak. (2021). A review on soft sensors for monitoring, control, and optimization of industrial processes, control, and optimization of industrial processes. IEEE Sensors Journal, 21(11): 12868–12881
https://doi.org/10.1109/JSEN.2020.3033153
11 P E Jupp, A Kume. (2020). Measures of goodness of fit obtained by almost-canonical transformations on Riemannian manifolds. Journal of Multivariate Analysis, 176: 104579
https://doi.org/10.1016/j.jmva.2019.104579
12 P Kadlec, B Gabrys, S Strandt. (2009). Data-driven Soft Sensors in the process industry. Computers & Chemical Engineering, 33(4): 795–814
https://doi.org/10.1016/j.compchemeng.2008.12.012
13 L Kang, R S Chen, N Xiong, Y C Chen, Y X Hu, C M Chen. (2019). Selecting hyper-parameters of gaussian process regression based on non-inertial particle swarm optimization in Internet of things. IEEE Access: Practical Innovations, Open Solutions, 7: 59504–59513
https://doi.org/10.1109/ACCESS.2019.2913757
14 M Kim, Y N Yang, M S Morikawa-Sakura, Q H Wang, M V Lee, D Y Lee, C P Feng, Y L Zhou, Z Y Zhang. (2012). Hydrogen production by anaerobic co-digestion of rice straw and sewage sludge. International Journal of Hydrogen Energy, 37(4): 3142–3149
https://doi.org/10.1016/j.ijhydene.2011.10.116
15 F Laio, S Tamea. (2007). Verification tools for probabilistic forecasts of continuous hydrological variables. Hydrology and Earth System Sciences, 11(4): 1267–1277
https://doi.org/10.5194/hess-11-1267-2007
16 X Y Li, X H Yi, Z H Liu, H B Liu, T Chen, G Q Niu, B Yan, C Chen, M Z Huang, G G Ying. (2021). Application of novel hybrid deep leaning model for cleaner production in a paper industrial wastewater treatment system. Journal of Cleaner Production, 294: 126343
https://doi.org/10.1016/j.jclepro.2021.126343
17 K B Newhart, R W Holloway, A S Hering, T Y Cath. (2019). Data-driven performance analyses of wastewater treatment plants: a review. Water Research, 157: 498–513
https://doi.org/10.1016/j.watres.2019.03.030 pmid: 30981980
18 H P Nguyen, J Liu, E Zio. (2020). A long-term prediction approach based on long short-term memory neural networks with automatic parameter optimization by Tree-structured Parzen Estimator and applied to time-series data of NPP steam generators. Applied Soft Computing, 89: 106116
https://doi.org/10.1016/j.asoc.2020.106116
19 G Ozcan, M Pajovic, Z Sahinoglu, Y B Wang, P V Orlik, T Wada. (2016). Online State of Charge Estimation for Lithium-Ion Batteries Using Gaussian Process Regression. Florence: IEEE, 998–1003
20 V Pham, T Bluche, C Kermorvant, J Louradour. (2014). Dropout Improves Recurrent Neural Networks for Handwriting Recognition. Hersonissos, Greece: IEEE, 285–290
21 S Putatunda, K Rama, Acm (2018). A Comparative Analysis of Hyperopt as Against Other Approaches for Hyper-Parameter Optimization of XGBoost. Shanghai: ACM
22 S Qiao, Q Wang, J Zhang, Z Pei. (2020). Detection and classification of early decay on blueberry based on improved deep residual 3D convolutional neural network in hyperspectral images. Scientific Programming, 2020: 1–12
https://doi.org/10.1155/2020/8895875
23 M Safari M a, N Masseran, M H A Majid. (2020). Robust reliability estimation for lindley distribution: a probability integral transform statistical approach. Mathematics, 8(9): 1634
https://doi.org/10.3390/math8091634
24 O Samuelsson, A Björk, J Zambrano, B Carlsson. (2017). Gaussian process regression for monitoring and fault detection of wastewater treatment processes. Water Science and Technology, 75(12): 2952–2963
https://doi.org/10.2166/wst.2017.162 pmid: 28659535
25 H Şenol. (2021). Methane yield prediction of ultrasonic pretreated sewage sludge by means of an artificial neural network. Energy, 215: 119173
https://doi.org/10.1016/j.energy.2020.119173
26 N Srivastava, G Hinton, A Krizhevsky, I Sutskever, R Salakhutdinov. (2014). Dropout: A simple way to prevent neural networks from overfitting. Journal of Machine Learning Research, 15: 1929–1958
27 B Szelag, A Gawdzik, A Gawdzik. (2017). Application of selected methods of black box for modelling the settleability process in wastewater treatment plant. Ecological Chemistry and Engineering S-Chemia I Inzynieria Ekologiczna S, 24(1): 119–127
28 H T Wang, Y Yang, A A Keller, X Li, S J Feng, Y N Dong, F T Li. (2016). Comparative analysis of energy intensity and carbon emissions in wastewater treatment in USA, Germany, China and South Africa. Applied Energy, 184: 873–881
https://doi.org/10.1016/j.apenergy.2016.07.061
29 J Wang, Q Cui, X Sun. (2021). A novel framework for carbon price prediction using comprehensive feature screening, bidirectional gate recurrent unit and Gaussian process regression. Journal of Cleaner Production, 314: 128024
https://doi.org/10.1016/j.jclepro.2021.128024
30 J P Wei, G F Liang, J Alex, T C Zhang, C B Ma. (2020). Research progress of energy utilization of agricultural waste in China: Bibliometric analysis by citespace. Sustainability (Basel), 12(3): 812
https://doi.org/10.3390/su12030812
31 X Wu, Y Wang, C Wang, W Wang, F Dong. (2021). Moving average convergence and divergence indexes based online intelligent expert diagnosis system for anaerobic wastewater treatment process. Bioresource Technology, 324: 124662
https://doi.org/10.1016/j.biortech.2020.124662 pmid: 33434874
32 Y Xu, W Gao, F Qian, Y Li. (2021). Potential analysis of the attention-based LSTM model in ultra-short-term forecasting of building HVAC energy consumption. Frontiers in Energy Research, 9: 730640
https://doi.org/10.3389/fenrg.2021.730640
33 K Yaginuma, S Tanabe, M Kano. (2022). Gray-box soft sensor for water content monitoring in fluidized bed granulation. Chemical & Pharmaceutical Bulletin, 70(1): 74–81
https://doi.org/10.1248/cpb.c21-00777 pmid: 34980737
34 G M Zeng, X D Li, R Jiang, J B Li, G H Huang. (2006). Fault diagnosis of WWTP based on improved support vector machine. Environmental Engineering Science, 23(6): 1044–1054
https://doi.org/10.1089/ees.2006.23.1044
35 C Zhang, H Wei, X Zhao, T Liu, K Zhang. (2016). A Gaussian process regression based hybrid approach for short-term wind speed prediction. Energy Conversion and Management, 126: 1084–1092
https://doi.org/10.1016/j.enconman.2016.08.086
36 Z Zhang, L Ye, H Qin, Y Liu, C Wang, X Yu, X Yin, J Li. (2019). Wind speed prediction method using shared weight long short-term memory network and gaussian process regression. Applied Energy, 247: 270–284
https://doi.org/10.1016/j.apenergy.2019.04.047
[1] FSE-22097-OF-LJL_suppl_1 Download
[1] Zhaocai Wang, Qingyu Wang, Tunhua Wu. A novel hybrid model for water quality prediction based on VMD and IGOA optimized for LSTM[J]. Front. Environ. Sci. Eng., 2023, 17(7): 88-.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed