Environmental Science and Pollution Research

, Volume 25, Issue 10, pp 9360–9370 | Cite as

Application of experimental design for the optimization of artificial neural network-based water quality model: a case study of dissolved oxygen prediction

  • Aleksandra Šiljić Tomić
  • Davor Antanasijević
  • Mirjana Ristić
  • Aleksandra Perić-Grujić
  • Viktor Pocajt
Research Article
  • 80 Downloads

Abstract

This paper presents an application of experimental design for the optimization of artificial neural network (ANN) for the prediction of dissolved oxygen (DO) content in the Danube River. The aim of this research was to obtain a more reliable ANN model that uses fewer monitoring records, by simultaneous optimization of the following model parameters: number of monitoring sites, number of historical monitoring data (expressed in years), and number of input water quality parameters used. Box–Behnken three-factor at three levels experimental design was applied for simultaneous spatial, temporal, and input variables optimization of the ANN model. The prediction of DO was performed using a feed-forward back-propagation neural network (BPNN), while the selection of most important inputs was done off-model using multi-filter approach that combines a chi-square ranking in the first step with a correlation-based elimination in the second step. The contour plots of absolute and relative error response surfaces were utilized to determine the optimal values of design factors. From the contour plots, two BPNN models that cover entire Danube flow through Serbia are proposed: an upstream model (BPNN-UP) that covers 8 monitoring sites prior to Belgrade and uses 12 inputs measured in the 7-year period and a downstream model (BPNN-DOWN) which covers 9 monitoring sites and uses 11 input parameters measured in the 6-year period. The main difference between the two models is that BPNN-UP utilizes inputs such as BOD, P, and PO43−, which is in accordance with the fact that this model covers northern part of Serbia (Vojvodina Autonomous Province) which is well-known for agricultural production and extensive use of fertilizers. Both models have shown very good agreement between measured and predicted DO (with R2 ≥ 0.86) and demonstrated that they can effectively forecast DO content in the Danube River.

Keywords

Dissolved oxygen Modeling ANN Design of experiment Parameter selection 

Notes

Acknowledgments

This work was supported by the Ministry of Education, Science and Technological Development of the Republic of Serbia [Project No. 172007].

Supplementary material

11356_2018_1246_MOESM1_ESM.pdf (4.6 mb)
ESM 1 (PDF 4733 kb)

References

  1. Alagumurthi N, Palaniradja K, Soundararajan V (2006) Optimization of grinding process through design of experiment (DOE)—a comparative study. Mater Manuf Process 21:19–21.  https://doi.org/10.1081/AMP-200060605 CrossRefGoogle Scholar
  2. Annadurai G, Sheeja RY (1998) Use of Box-Behnken design of experiments for the adsorption of vetofix red using biopolymer. Bioprocess Eng 18(6):463–466.  https://doi.org/10.1007/s004490050472 CrossRefGoogle Scholar
  3. Antanasijević D, Pocajt V, Povrenović D, Perić-Grujić A, Ristić M (2013) Modelling of dissolved oxygen content using artificial neural networks: Danube River, North Serbia, case study. Environ Sci Pollut Res Int 20(12):9006–9013.  https://doi.org/10.1007/s11356-013-1876-6 CrossRefGoogle Scholar
  4. Antanasijević D, Pocajt V, Perić-Grujić A, Ristić M (2014) Modelling of dissolved oxygen in the Danube River using artificial neural networks and Monte Carlo simulation uncertainty analysis. J Hydrol 519:1895–1907.  https://doi.org/10.1016/j.jhydrol.2014.10.009 CrossRefGoogle Scholar
  5. Aslan N, Cebeci Y (2007) Application of Box–Behnken design and response surface methodology for modeling of some Turkish coals. Fuel 86(1-2):90–97.  https://doi.org/10.1016/j.fuel.2006.06.010 CrossRefGoogle Scholar
  6. Ay M, Kişi O (2017) Estimation of dissolved oxygen by using neural networks and neuro fuzzy computing techniques. KSCE J Civ Eng 21(5):1631–1639.  https://doi.org/10.1007/s12205-016-0728-6 CrossRefGoogle Scholar
  7. Balestrassi PP, Popova E, Paiva AP, Marangon Lima JW (2009) Design of experiments on neural network’s training for nonlinear time series forecasting. Neurocomputing 72(4-6):1160–1178.  https://doi.org/10.1016/j.neucom.2008.02.002 CrossRefGoogle Scholar
  8. Basant N, Gupta S, Malik A, Singh KP (2010) Linear and nonlinear modeling for simultaneous prediction of dissolved oxygen and biochemical oxygen demand of the surface water—a case study. Chemom Intell Lab Syst 104(2):172–180.  https://doi.org/10.1016/j.chemolab.2010.08.005 CrossRefGoogle Scholar
  9. Batalon JT, Madamba PS (2001) Optimization of coir dust compaction using the response surface methodology approach. J Agric Eng Res 78(2):167–175.  https://doi.org/10.1006/bioe.2002.0066 CrossRefGoogle Scholar
  10. Bezerra MA, Santelli RE, Oliveira EP, Villar LS, Escaleira LA (2008) Response surface methodology (RSM) as a tool for optimization in analytical chemistry. Talanta 76(5):965–977.  https://doi.org/10.1016/j.talanta.2008.05.019 CrossRefGoogle Scholar
  11. Bolanča T, Cerjan-Stefanović Š, Ukić Š, Rogošić M, Luša M (2008) Application of different training methodologies for the development of a back propagation artificial neural network retention model in ion chromatography. J Chemom 22(2):106–113.  https://doi.org/10.1002/cem.1096 CrossRefGoogle Scholar
  12. Box GEP, Behnken DW (1960) Some new three level designs for the study of quantitative variables. Technometrics 2(4):455–475.  https://doi.org/10.1080/00401706.1960.10489912 CrossRefGoogle Scholar
  13. Box GEP, Draper NR (1987) Empirical model—building and response surfaces. John Wiley & Sons, Wiley, New YorkGoogle Scholar
  14. Can MY, Kaya Y, Algur OF (2006) Response surface optimization of the removal of nickel from aqueous solution by cone biomass of Pinus sylvestris. Bioresour Technol 97(14):1761–1765.  https://doi.org/10.1016/j.biortech.2005.07.017 CrossRefGoogle Scholar
  15. Carton JG, Olabi AG (2010) Design of experiment study of the parameters that affect performance of three flow plate configurations of a proton exchange membrane fuel cell. Energy 35(7):2796–2806.  https://doi.org/10.1016/j.energy.2010.02.044 CrossRefGoogle Scholar
  16. Costa M, Gonçalves AM (2011) Clustering and forecasting of dissolved oxygen concentration on a river basin. Stoch Environ Res Risk Assess 25(2):151–163.  https://doi.org/10.1007/s00477-010-0429-5 CrossRefGoogle Scholar
  17. Csábrági A, Molnár S, Tanos P, Kovács J (2017) Application of artificial neural networks to the forecasting of dissolved oxygen content in the Hungarian section of the river Danube. Ecol Eng 100:63–72.  https://doi.org/10.1016/j.ecoleng.2016.12.027 CrossRefGoogle Scholar
  18. Ehret A, Hochstuhl D, Gianola D, Thaller G (2015) Application of neural networks with back-propagation to genome-enabled prediction of complex traits in Holstein-Friesian and German Fleckvieh cattle. Genet Sel Evol 47(1):22.  https://doi.org/10.1186/s12711-015-0097-5 CrossRefGoogle Scholar
  19. Ferreira SLC, Bruns RE, Ferreira HS, Matos GD, David JM, Brandão GC, da Silva EGP, Portugal LA, dos Reis PS, Souza AS, dos Santos WNL (2007) Box-Behnken design: an alternative for the optimization of analytical methods. Anal Chim Acta 597(2):179–186.  https://doi.org/10.1016/j.aca.2007.07.011 CrossRefGoogle Scholar
  20. Ghadge SV, Raheman H (2006) Process optimization for biodiesel production from mahua (Madhuca indica) oil using response surface methodology. Bioresour Technol 97(3):379–384.  https://doi.org/10.1016/j.biortech.2005.03.014 CrossRefGoogle Scholar
  21. Heddam S, Kisi O (2017) Extreme learning machines: a new approach for modeling dissolved oxygen (DO) concentration with and without water quality variables as predictors. Environ Sci Pollut Res 24(20):16702–16724.  https://doi.org/10.1007/s11356-017-9283-z CrossRefGoogle Scholar
  22. Hornik K, Stinchcombe M, White H (1989) Multilayer feedforward networks are universal approximators. Neural Netw 2(5):359–366.  https://doi.org/10.1016/0893-6080(89)90020-8 CrossRefGoogle Scholar
  23. ICPDR (2012) Strategy on adaptation to climate change. https://www.icpdr.org/flowpaper/viewer/default/files/nodes/documents/icpdr_climate-adaptation-strategy.pdf. Accessed 14 May 2017
  24. ICPDR (2015) The Danube River Basin District Management Plan. https://www.icpdr.org/flowpaper/viewer/default/files/nodes/documents/drbmp-update2015.pdf. Accessed 14 May 2017
  25. Ji X, Shang X, Dahlgren RA, Zhang M (2017) Prediction of dissolved oxygen concentration in hypoxic river systems using support vector machine: a case study of Wen-Rui Tang River, China. Environ Sci Pollut Res 24(19):16062–16076.  https://doi.org/10.1007/s11356-017-9243-7 CrossRefGoogle Scholar
  26. Keshtegar B, Heddam S (2017) Modeling daily dissolved oxygen concentration using modified response surface method and artificial neural network: a comparative study. Neural Comput Appl.  https://doi.org/10.1007/s00521-017-2917-8
  27. Khaw JFC, Lim BS, Lim LEN (1995) Optimal design of neural networks using the Taguchi method. Neurocomputing 7(3):225–245.  https://doi.org/10.1016/0925-2312(94)00013-I CrossRefGoogle Scholar
  28. Kia MB, Pirasteh S, Pradhan B, Mahmud AR, Sulaiman WNA, Moradi A (2012) An artificial neural network model for flood simulation using GIS: Johor River Basin, Malaysia. Environ Earth Sci 67(1):251–264.  https://doi.org/10.1007/s12665-011-1504-z CrossRefGoogle Scholar
  29. Kim Y-S, Yum B-J (2004) Robust design of multilayer feedforward neural networks: an experimental approach. Eng Appl Artif Intell 17(3):249–263.  https://doi.org/10.1016/j.engappai.2003.12.005 CrossRefGoogle Scholar
  30. Lek S, Guegan JF (1999) Artificial neural network as a tool in ecological modelling, an introduction. Ecol Model 120(2-3):65–73.  https://doi.org/10.1016/S0304-3800(99)00092-7 CrossRefGoogle Scholar
  31. Millie DF, Weckman GR, Young WA, Ivey JE, Carrick HJ, Fahnenstiel GL (2012) Modeling microalgal abundance with artificial neural networks: demonstration of a heuristic “Grey-Box” to deconvolve and quantify environmental influences. Environ Model Softw 38:27–39.  https://doi.org/10.1016/j.envsoft.2012.04.009 CrossRefGoogle Scholar
  32. Natchkov I (1997) A guide to the use of water quality management principles, case study IX—the Danube Basin. Water Pollut Control, WHO/UNEP. http://www.who.int/water_sanitation_health/resourcesquality/wpccasestudy9.pdf. Accessed 16 Apr 2017
  33. NIST/SEMATECH (2012) e-handbook of statistical methods. http://www.itl.nist.gov/div898/handbook/. Accessed 16 April 2017
  34. Olden J, Jackson D (2002) Illuminating the “black box”: a ramdomization approach for understanding variable contributions in artificial neuronal networks. Ecol Model 154(1-2):135–150.  https://doi.org/10.1016/S0304-3800(02)00064-9 CrossRefGoogle Scholar
  35. Olyaie E, Zare Abyaneh H, Danandeh Mehr A (2017) A comparative analysis among computational intelligence techniques for dissolved oxygen prediction in Delaware River. Geosci Front 8:517–527CrossRefGoogle Scholar
  36. Onkal-Engin G, Demir I, Engin SN (2005) Determination of the relationship between sewage odour and BOD by neural networks. Environ Model Softw 20(7):843–850.  https://doi.org/10.1016/j.envsoft.2004.04.012 CrossRefGoogle Scholar
  37. Packianather MS, Drake PR, Rowlands H (2000) Optimizing the parameters of multilayered feedforward neural networks through Taguchi design of experiments. Qual Reliab Eng Int 16(6):461–473.  https://doi.org/10.1002/1099-1638(200011/12)16:6<461::AID-QRE341>3.0.CO;2-G CrossRefGoogle Scholar
  38. Paschalidou AK, Karakitsios S, Kleanthous S, Kassomenos PA (2011) Forecasting hourly PM(10) concentration in Cyprus through artificial neural networks and multiple regression models: implications to local environmental management. Environ Sci Pollut Res Int 18(2):316–327.  https://doi.org/10.1007/s11356-010-0375-2 CrossRefGoogle Scholar
  39. Phansalkar VV, Sastry PS (1994) Analysis of the back-propagation algorithm with momentum. IEEE Trans Neural Netw 5(3):505–506.  https://doi.org/10.1109/72.286925 CrossRefGoogle Scholar
  40. Raheli B, Aalami MT, El-Shafie A, Ghorbani MA, Deo RC (2017) Uncertainty assessment of the multilayer perceptron (MLP) neural network model with implementation of the novel hybrid MLP-FFA method for prediction of biochemical oxygen demand and dissolved oxygen: a case study of Langat River. Environ Earth Sci 76(503).  https://doi.org/10.1007/s12665-017-6842-z
  41. Rene ER, Saidutta MB (2008) Prediction of water quality indices by regression analysis and artificial neural networks. Int J Environ Res 2:183–188Google Scholar
  42. Rumelhart DE, Hinton GE, Williams RJ (1986) Learning representations by back-propagating errors. Nature 323(6088):533–536.  https://doi.org/10.1038/323533a0 CrossRefGoogle Scholar
  43. SEPA (2015) State of surface waters of Serbia—analysis and elements for monitoring design. http://www.sepa.gov.rs/download/VodeSrbije/StatusPovrsinskihVodaSrbije.pdf. Accessed on 19 March 2017
  44. Shahin MA, Jaksa MB, Maier HR (2008) State of the art of artificial neural networks in geotechnical engineering. Electron J Geotech Eng 8:1–26Google Scholar
  45. Šiljić Tomić AN, Antanasijević DZ, Ristić M, Perić-Grujić AA, Pocajt VV (2016) Modeling the BOD of Danube River in Serbia using spatial, temporal, and input variables optimized artificial neural network models. Environ Monit Assess 188(5):300.  https://doi.org/10.1007/s10661-016-5308-1 CrossRefGoogle Scholar
  46. Šiljić Tomić A, Antanasijević D, Ristić M, Perić-Grujić A, Pocajt V (2018) A linear and non-linear polynomial neural network modeling of dissolved oxygen content in surface water: inter- and extrapolation performance with inputs’ significance analysis. Sci Total Environ 610–611:1038–1046Google Scholar
  47. Šiljić A, Antanasijević D, Perić-Grujić A, Ristić M, Pocajt V (2014) Artificial neural network modelling of biological oxygen demand in rivers at the national level with input selection based on Monte Carlo simulations. Environ Sci Pollut Res 22(6):4230–4241.  https://doi.org/10.1007/s11356-014-3669-y Google Scholar
  48. Statsoft (2010) Underst SStatistica 10. http://statistica.io/
  49. Sukthomya W, Tannock J (2005) The optimisation of neural network parameters using Taguchi’s design of experiments approach: an application in manufacturing process modelling. Neural Comput Applic 14(4):337–344.  https://doi.org/10.1007/s00521-005-0470-3 CrossRefGoogle Scholar
  50. Walczak S, Cerpa N (1999) Heuristic principles for the design of artificial neural networks. Inf Softw Technol 41(2):107–117.  https://doi.org/10.1016/S0950-5849(98)00116-5 CrossRefGoogle Scholar
  51. Wen X, Fang J, Diao M, Zhang C (2013) Artificial neural network modeling of dissolved oxygen in the Heihe River, Northwestern China. Environ Monit Assess 185(5):4361–4371.  https://doi.org/10.1007/s10661-012-2874-8 CrossRefGoogle Scholar
  52. Yang SM, Lee GS (1999) Neural network design by using Taguchi method. J Dyn Syst Meas Control 121:560–563.  https://doi.org/10.1115/1.2802515 CrossRefGoogle Scholar
  53. Yetilmezsoy K, Demirel S, Vanderbei RJ (2009) Response surface modeling of Pb(II) removal from aqueous solution by Pistacia vera L.: Box-Behnken experimental design. J Hazard Mater 171(1-3):551–562.  https://doi.org/10.1016/j.jhazmat.2009.06.035 CrossRefGoogle Scholar
  54. Zinatizadeh AAL, Mohamed AR, Abdullah AZ, Mashitah MD, Hasnain Isa M, Najafpour GD (2006) Process modeling and analysis of palm oil mill effluent treatment in an up-flow anaerobic sludge fixed film bioreactor using response surface methodology (RSM). Water Res 40(17):3193–3208.  https://doi.org/10.1016/j.watres.2006.07.005 CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Faculty of Technology and MetallurgyUniversity of BelgradeBelgradeSerbia
  2. 2.Innovation Center of the Faculty of Technology and MetallurgyUniversity of BelgradeBelgradeSerbia

Personalised recommendations