-
Abstract: Grain-size distribution data, as a substitute for measuring hydraulic conductivity (K), has often been used to get K value indirectly. With grain-size distribution data of 150 sets of samples being input data, this study combined the Artificial Neural Network technology (ANN) and Markov Chain Monte Carlo method (MCMC), which replaced the Monte Carlo method (MC) of Generalized Likelihood Uncertainty Estimation (GLUE), to establish the GLUE-ANN model for hydraulic conductivity prediction and uncertainty analysis. By means of applying the GLUE-ANN model to a typical piedmont region and central region of North China Plain, and being compared with actually measured values of hydraulic conductivity, the relative error ranges are between 1.55% and 23.53% and between 14.08% and 27.22% respectively, the accuracy of which can meet the requirements of groundwater resources assessment. The global best parameter gained through posterior distribution test indicates that the GLUE-ANN model, which has satisfying sampling efficiency and optimization capability, is able to reasonably reflect the uncertainty of hydrogeological parameters. Furthermore, the influence of stochastic observation error (SOE) in grain-size analysis upon prediction of hydraulic conductivity was discussed, and it is believed that the influence can not be neglected.
-
Key words:
- Grain-size distribution /
- Hydraulic conductivity /
- ANN /
- GLUE /
- MCMC /
- Stochastic Observation Error (SOE)
-
Table 1. Parameters in the ANN model
Parameter Physical meaning Value range D2000+ Gravel content, diameter more than 2 mm 0~0.02 D500 Coarse sand content, diameter between 0.5~2 mm 0.01~0.46 D250 Medium sand content, diameter between 0.25~0.5 mm 0.01~0.82 D75 Fine sand content, diameter between 0.075~0.25 mm 0.02~0.51 D5 Silt content, diameter between 0.005~0.075 mm 0.01~0.68 D5- Clay content, diameter less than 0.005 mm 0.01~0.65 N2 The number of hidden layer nodes 7~21 Wpi, j Network weight initial value -1~1 bpi Bias -1~1 Table 2. Comparison between the ANN model's output and measured values
Sample number Gravel Coarse sand Medium sand Fine sand Silt Clay Predicted K 10-6 (cm/s) Measured K 10-6 (cm/s) Relative error % > 2 2~0.5 0.5~0.25 0.25~0.075 0.075~0.005 < 0.005 content % zk11-01 1.15 4.05 12.33 5.10 64.06 13.31 18.01 18.33 1.75 zk11-02 0.00 4.10 14.59 8.18 57.12 16.01 13.34 13.59 1.84 zk11-03 0.00 3.35 8.09 10.17 33.12 45.27 2.42 2.98 18.79 zk11-04 0.53 7.26 20.20 4.11 49.07 18.83 9.29 9.10 2.08 zk11-05 0.00 10.32 13.52 24.97 42.15 9.04 28.91 28.47 1.55 zk11-06 0.00 15.16 46.94 28.05 8.14 1.71 602.56 665.30 9.43 zk11-07 0.00 8.20 1.08 3.34 62.30 25.08 2.65 2.51 5.58 zk11-08 1.00 1.13 1.19 5.17 65.40 26.11 3.80 3.92 3.06 zk11-09 0.00 31.22 39.14 1.35 26.54 1.75 536.11 589.78 9.10 zk11-10 0.00 1.26 5.79 1.62 54.45 36.88 1.87 1.97 5.08 zk11-11 0.00 32.49 27.55 6.88 30.87 2.21 410.41 456.98 10.19 zk11-12 0.00 28.01 38.40 10.02 21.91 1.66 445.09 415.70 7.07 zk11-13 0.00 20.34 42.15 4.37 31.51 1.63 404.02 448.31 9.88 zk11-14 0.00 2.30 2.27 13.25 35.90 46.28 2.21 2.89 23.53 zk11-15 0.00 3.36 1.33 1.03 54.73 39.55 1.23 1.36 9.56 zk11-16 0.00 39.22 20.68 7.21 30.70 2.19 394.80 439.18 10.11 zk11-17 0.00 24.98 55.17 11.15 6.97 1.73 589.85 644.71 8.51 zk11-18 0.00 2.24 1.12 2.33 56.66 37.65 1.18 1.05 12.38 zk11-19 0.00 1.27 2.97 2.10 60.15 33.51 0.88 0.92 4.35 zk11-20 2.06 5.10 4.55 8.51 53.28 26.50 5.03 5.24 4.01 Table 3. The comparison between the ANN model's output and measured values of samples from middle region of NCP
Sample number Gravel Coarse sand Medium sand Fine sand Silt Clay Predicted K 10-6 (cm/s) Measured K 10-6 (cm/s) Relative error % > 2 2~0.5 0.5~0.25 0.25~0.075 0.075~0.005 < 0.005 content % S18-01 0.00 23.52 20.58 33.54 12.25 10.11 17.33 20.17 14.08 S18-02 0.00 25.08 17.41 45.88 9.38 2.25 398.75 475.32 16.11 S18-03 0.00 25.49 22.72 28.10 10.89 12.80 42.36 36.74 15.30 S18-04 0.00 29.37 25.26 21.65 14.17 9.55 31.91 40.47 21.15 S18-05 0.00 28.90 19.21 41.20 8.92 1.77 332.97 412.55 19.29 S18-06 0.00 28.83 19.88 37.21 11.12 2.96 274.10 376.61 27.22 S18-07 0.00 41.28 22.43 17.57 4.22 14.50 13.73 18.39 25.34 -
Alfaro Soto MA, Chang HK, Th M, et al. 2017. Fractal-based models for the unsaturated soil hydraulic functions. Geoderma, 306: 144-151. doi: 10.1016/j.geoderma.2017.07.019 Awad HS, Bassam AM. 2001. A computer program to calculate hydraulic conductivity from grain size data in Saudi Arabia. International Journal of Water Resources Development, 17(2): 237-246. doi: 10.1080/07900620120031298 Das SK, Samui P, Sabat AK. 2012. Prediction of field hydraulic conductivity of clay liners using an artificial neural network and support vector machine. International Journal of Geomechanics, 12(5): 606-611. doi: 10.1061/(ASCE)GM.1943-5622.0000129 David CW, Asce F. 2003. Goodbye, Hazen; Hello, Kozeny-Carman. Journal of Geotechnical and Geoenvironmental Engineering, 129(11): 1054-1056. doi: 10.1061/(ASCE)1090-0241(2003)129:11(1054) DONG Pei. 2010. Laboratory studies of sand column on the dynamic evaporation on a shallow water table. M.S. thesis, Beijing: China University of Geosciences: 10-22. (in Chinese) Erzin Y, Gumaste SD, Gupta AK, et al. 2009. Artificial neural network (ANN) models for determining hydraulic conductivity of compacted fine-grained soil. Canadian Geotechnical Journal, 46: 955-968 doi: 10.1139/T09-035 FAN Gui-sheng, XING Ri-xian, ZHANG Ming-bin. 2012. Experimental study on permeability of the sandy gravel media with different gradation. Journal of Taiyuan University of Technology, 43(3): 373-378. (in Chinese) http://www.cqvip.com/QK/90007A/20123/41765758.html GONG Guang-lu, QIAN Min-ping. 2003. Applied stochastic coursebook and its application to algorithm and intelligent computing. Beijing: Tsinghua University Press: 35-58. (in Chinese) Haario H, Saksman E, Tamminen J. 2005. Compon-entwise adaptation for high dimensional MCMC. Computational Statistics, 20(2): 265-273. doi: 10.1007/BF02789703 Hasan M, Ozer C, Ramazan M, et al. 2006. Comparison of artificial neural and regression pedotransfer functions for prediction of soil water retention and saturated hydraulic conductivity. Soil and Tillage Research, 90(2): 108-116. http://www.sciencedirect.com/science/article/pii/S0167198705002436 Haykin S. 2004. Neural Networks: A compre-hensive foundation, second edition. Beijing: China Machine Press: 46-55. (in Chinese) Isik Y, Marian M, Martin B. 2012. Neural com-puting models for prediction of permeability coefficient of coarse-grained soils. Neural Computing and Applications, 21(5): 957-968. doi: 10.1007/s00521-011-0535-4 JI Rui-li, ZHANG Ming, SU Rui, 2016. Research of in-situ hydraulic test method by using double packer equipment. Journal of Ground-water Science and Engineering, 4(1): 41-51. Justine O. 2007. Evaluation of empirical formulae for determination of hydraulic conductivity based on Grain-size analysis. Journal of American Science, 3(3): 54-60. http://www.researchgate.net/publication/254793785_Evaluation_of_Empirical_Formulae_for_Determination_of_Hydraulic_Conductivity_based_on_Grain-Size_Analysis Keith B. 2006. A manifesto for the equifinality thesis. Journal of Hydrology, 320(1-2): 18-36. doi: 10.1016/j.jhydrol.2005.07.007 Koekkoek JW, Booltink H. 1999. Neural network models to predict soil water retention. European Journal of Soil Science, 50(3): 489-495. doi: 10.1046/j.1365-2389.1999.00247.x LI Shou-ju, LIU Ying-xi, WANG Deng-gang, et al. 2002. Inversion algorithm of permeability coefficients of rockmass and its application based on artificial neural network. Chinese Journal of Rock Mechanics and Engineering, 21(4): 479-483. (in Chinese) http://www.researchgate.net/publication/297311638_Inversion_algorithm_of_permeability_coefficients_of_rockmass_and_its_application_based_on_artificial_neural_network LU Le, WU Ji-chun. 2010. Bayesian analysis of uncertainties in groundwater numerical simulation. Journal of Hydraulic Engineering, 41(3): 264-271. http://en.cnki.com.cn/Article_en/CJFDTOTAL-SLXB201003003.htm LU Le, WU Ji-chun, CHEN Jing-Ya. 2008. Identification of hydrogeological parameters based on the bayesian method. Hydrogeology and Engineering Geology, 58(5): 58-63. (in Chinese) http://en.cnki.com.cn/Article_en/CJFDTOTAL-SWDG200805017.htm Mahmoud S, Alyaman I, Zekai S. 1993. Deter-mination of hydraulic conductivity from complete Grain-Size Distribution Curves. Ground Water, 31(4): 551-555. doi: 10.1111/j.1745-6584.1993.tb00587.x MAO Shi-song. 1999. Bayesian statistics. Beijing: China Statistics Press. (in Chinese) Nakhaei M. 2005. Estimating the saturated hy-draulic conductivity of granular material using artificial neural network based on grain size distribution curve. Islam Repub Iran, 16(1): 55-62. http://www.researchgate.net/publication/228371456_Estimating_the_Saturated_Hydraulic_Conductivity_of_Granular_Material_Using_Artificial_Neural_Network_Based_on_Grain_Size_Distribution_Curve Namunu JM, Ian PK, Arulanandan K. 1989. An expression for the permeability of anisotropic granular media. International Journal for Numerical and Analytical Methods in Geo-mechanics, 13(6): 575-598. doi: 10.1002/nag.1610130602 Park HI. 2011. Development of neural network model to estimate the permeability coe-fficient of soils. Marine Georesources & Geotechnology, 29(4): 267-278. Russel GS. 1989. Correlations of permeability and grain size. Ground Water, 27(5): 633-636. doi: 10.1111/j.1745-6584.1989.tb00476.x Salarashayeri AF, Siosemarde M. 2012. Prediction of soil hydraulic conductivity from particle-size distribution. World Academy of Science, Engineering and Technology, 61: 454-457. Smiles DE, Youngs EG. 1963. A Multiple-well method for determining the hydraulic con-ductivity of a saturated soil in Situ. Journal of Hydrology, 1(4): 279-287. doi: 10.1016/0022-1694(63)90019-2 Smith AFM, Roberts GO. 1993. Bayesian com-putation via the Gibbs sampler and related Markov chain Monte Carlo Methods. Journal of Royal Statistical Society, Series B, 55: 3-24. doi: 10.1111/j.2517-6161.1993.tb01466.x TANG Xiao-song, ZHENG Ying-ren, DONG Cheng. 2007. The prediction of seepage co-efficient of coarse-grained soil by neurotic network. Rock and Soil Mechanics, 28: 133-136, 143. (in Chinese) http://en.cnki.com.cn/Article_en/CJFDTOTAL-YTLX2007S1028.htm WANG Dong, Vijay PS, ZHU Yuan-shen, et al. 2009. Stochastic observation error and uncertainty in water quality evaluation. Ad-vances in Water Resources, 32: 1526-1534. doi: 10.1016/j.advwatres.2009.07.004 XU Peng, QIU Shu-xia, JIANG Zhou-ting, et al. 2011. Fractal analysis of Kozeny-Carman constant in the homogenous porous media. Journal of Chongqing University, 34(4): 78-82. (in Chinese)