عنوان مقاله [English]
The use of clay blanket in reservoirs is one of the main methods of seepage reducing. In this study, with clay blanket modeling in a proposed reservoir by finite element method, 350 dataset was obtained using SEEP/W. Validation of SEEP/W was carried out by comparing seepage results obtained from a laboratory tests. For evaluation of suitable model for predicting seepage values (results of modeling), used from five artificial intelligence techniques comprising: multilayer perceptron neural network (MLP), radial base function (RBF), gene expression programming (GEP), support vector regression (SVR) and a novel hybrid model of the firefly algorithm (FFA) with the multilayer perceptron (MLP-FFA). All the techniques were trained with 70% of available dataset and tested using the remaining 30% dataset. Different combinations of input data that include the ratio of the permeability coefficient of foundation to the permeability coefficient of clay blanket (K_f/K_b ), the ratio of the length of blanket to upstream head (L_1/H), the ratio of thickness of foundation to thickness of blanket (h_f/t), the ratio of length of blanket to thickness of core (L_1/L_2 ) and the ratio of horizontal to vertical permeability coefficient of foundation (K_(f_x )/K_(f_y ) ) were used for evaluation of mentioned methods. The results were evaluated using four performance criteria metrics: root mean square error (RMSE), mean absolute error (MAE), Nash-Sutcliffe efficiency (NS), Willmott’s Index of agreement (WI) and Taylor diagram. The results of study showed that the MLP-FFA method provides better estimation results than the other models and therefore, could be applied an optimized for predictive model of earth fill dam seepage.
Preventing water penetration and seepage control is of prime importance in hydraulic structures projects. Recent studies show that 30% of dam failures are due to the seepage from dam’s body or foundation. Seepage control inherently is controlling potential energy of water molecules causing seepage and related losses. Constructing a core with low rate of permeability can considerably control seepage from dam body. So foundation seepages are significantly more than body seepages. Foundation seepage control is done to prevent uplift and piping, two phenomena which led to dam failure. One of the methods for controlling seepage from bottom of earth dams which are mounted on alluvial foundation with high rate of permeability, is utilizing a covering layer with low permeability on bed of river, bottom of reservoir (in upstream) and connecting it to central core of dam. In fact, the role of such methods and mentioned covering layer is lengthening flow path for increasing potential losses and decreasing water energy which is terminated to decrease penetrated water and related losses. This covering layer is called clay blanket.
One of the longest upstream impermeable blankets is executed in Tarbela dam in Pakistan with 140 m height. This blanket has 1400 m length and its thickness is 1.52 m at the dam (WCD, 2000). Khalili and Amiri (2015) investigated cutoff effect in reducing leakage, exit gradient and uplift, both experimentally and numerically analyze by software GEOSTUDIO and referring that the results of the software are in acceptable agreement with the experimental results. Tayfu et al. (2005) used Finite Element Method (FEM) and Artificial Neural Network (ANN) models for flow through Jeziorsko Earth fill Dam in Poland. This case study offers insight into the adequacy of ANN as well as its competitiveness against FEM for seepage prediction through an earth fill dam body. Ahmed and Sattar (2014) used Gene expression models (GEP) for prediction of dam failure and results showed the superiority of the developed GEP models over existing regression-based models.
The goal in the proposed study is to introduce the best statistical model to predict the leakage from dams. For this purpose, all important and effective parameters for clay blanket including; permeability coefficient, blanket length and thickness, alluvial foundation thickness and its permeability coefficient, and the ratio of horizontal to vertical alluvial foundation permeability coefficient (which is very effective in seepage from foundation) were modeled with SEEP/W, and seepage values were obtained. Then for choosing the best statistical model, some of the most commonly neural network models comprising FFA, RBF, MLP, GEP and SVR were used. Based on the seepage values, the above-mentioned models were compared.
1-Ahmed, M. and Sattar, A., 2014. Gene expression models for prediction of dam breach Parameter. Journal of Hydroinformatics, 16(3), pp. 550-571.
2- Dehghani, N., Pirmoradian, N., Azimi, V. and Khanmohammady, S., 2013. Evaluation of MLP and RBF for estimating of monthly evaporation, case study: Rasht meteorological station. In 2th National Conference on Sustainable Agriculture and Sustainable Environment, (In Persian).
3- Derin, U. and Mert tolon, S. M., 2008. Slope Stability during Earthquakes: A Neural Network Application. Geo Congrss. Characrerization, Monitoring and Modeling of Geosystems, pp. 878-2008.
4-Ferreira, C., 2001. Gene expression programming: A new adaptive algorithm for solving problems. Complex Systems, 13(2), pp. 87-129.
5-Fu, Q., Jiang, R., Wang, Z. and Li, T., 2015. Optimization of soil water characteristic curves parameters by modified firefly algorithm. Nongye Gongcheng Xuebao/Transactions of the Chinese Society of Agricultural Engineering, 31(11), pp. 117-122.
6-Gocic, M., Petkovic, M., Trajkovic, S., Shamshirband, S. H., Moetamedi, S. H. and Roslan, H., 2015. Determination of the most influential weather parameters on reference evapotranspiration by adaptive neuro-fuzzy methodology. Computers and Electronics in Agriculture, 114(1), pp. 277–284.
7-Haykin, S., 1999. Neural Networks: A Comprehensive Foundation. Prentice-Hall Inc.NJ.
8-Kavousi, A., Samet, H. and Marzbani, F., 2014. A new hybrid modified firefly algorithm and support vector regression model for accurate short term load forecasting. Expert Systems with Applications. 41(2), pp. 6047-6056.
9-KazemzadehParsi, M., 2014. A modified firefly algorithm for engineering design optimization problems. Iranian Journal of Science and Technology, 38(1), pp. 403-421.
10- Khalili Shayan, H. and Amiri Tokaldany, E., 2015. Effects of blanket, drains, and cutoff wall on reducing uplift pressure. Seepage, and exit gradient under hydraulic structures. International Journal of Civil Engineering, 13(4), pp. 486-500.
11- Khalili Shayan, H. and Amiri Tokaldani, E., 2012. Experimentally and numerically investigation of Bligh and Lane theory for estimating uplift under diversion dams. In 10th Hydraulic conference, Gilan University, Gilan, Iran, (In Persian).
12- Khan, M. S. and Coulibaly, P., 2006. Bayesian neural network for rainfall-runoff modeling. Water Resources Research, 42, doi: 10.1029/2005WR003971. issn: 0043-1397, pp. 1-18.
13- Khatibi, A., Pourebrahim, S. H. and Danehkar, A., 2015. Application of Genetic Algorithm for Simulation of Land Use and Land Cover Changes; Case of Karaj City, Iran. Journal of Tethys, 3(4), pp. 286–296.
14-Nourani, V. and Babakhani, A., 2013. Integration of Artificial Neural Networks with Radial Basis Function Interpolation in Earthfill Dam Seepage Modeling. Journal of Computing in Civil Engineering. 27(1), pp.183-195.
15-Rahimi, H., 2004. Embankment Dams, Tehran University Press, (In Persian).
16-SEEP/W., 2012. Seepage Modeling with SEEP/W. Geo-Slope International Ltd, Calgary.
17-Talatahari, S., Hosseini, A., Mirghaderi, S. R. and Rezazadeh, F., 2014. Optimum Performance-Based Seismic Design Using a Hybrid Optimization Algorithm. Hindawi Publishing Corporation Mathematical Problems in Engineering. Volume 2014, Article ID 693128, 8 pages.
18-Tayfu, G., Swiate, D., Wita, A. and Singh, V., 2005. Case Study: Finite Element Method and Artificial Neural Network Models for Flow through Jeziorsko Earth fill Dam in Poland. Journal of Hydraulic Engineering, 131(3), pp. 431-440.
19-Taylor, KE., 2001. Summarizing multiple aspects of model performance in a single diagram. Journal of Geophysical Research, 106 (7), pp.7183–7192.
20-USACE., 1986. Seepage analysis and control for dams. Department of the US army corps of Engineers, Washington, D.C. 20314-1000.
21-USBR., 2014. Embankment dams, chapter 8, seepage, phase 4. U. S. Department of Interior Bureau of Reclamation.
22-Vapnik, VK., 1999. An Overview of Statistical Learning Theory. IEEE Transactions on Neural Network, 10(1), pp. 988-998.
23-Yang, X. S. 2010. Firefly Algorithm, Stochastic Test Functions and Design Optimisation. International Journal of Bio-Inspired Computation, 2(2), pp. 78–84.