An artificial neural network (ANN) model is proposed for the simultaneous determination of transmissivity and storativity distributions of a heterogeneous aquifer system. ANNs may be useful tools for parameter identification problems due to their ability to solve complex nonlinear problems. As an extension of previous study—Karahan H, Ayvaz MT (2006) Forecasting aquifer parameters using artificial neural networks, J Porous Media 9(5):429–444—the performance of the proposed ANN model is tested on a two-dimensional hypothetical aquifer system for transient flow conditions. In the proposed ANN model, Cartesian coordinates of observation wells, associated piezometric heads and observation time are used as inputs while corresponding transmissivity and storativity values are used as outputs. The training, validation and testing processes of the ANN model are performed under two scenarios. In scenario 1, all the sampled data are used through the simulation time. However, in the scenario 2, there are data gaps due to irregular observations. By using the determined synaptic network weights, transmissivity and storativity distributions are predicted. In addition, the performance of the proposed ANN is tested for different noise data conditions. Results showed that the developed ANN model may be used in simultaneous aquifer parameter estimation problems.RésuméUn modèle de réseau neuronal artificel (ANN) est proposé pour la détermination simultanée des distributions des transmissivités et coefficients d’emmagasinement dans un système aquifère hétérogène. Les ANN peuvent constituer des outils utiles pour les problèmes d’identification de paramètres, grâce à leur capacité à résoudre des problèmes complexes non-linéaires. Dans la continuité de l’étude précédente—Karahan H, Ayvaz MT (2006) Forecasting aquifer parameters using artificial neural networks [Estimation des paramètres des aquifères par un réseau neuronal artificiel], J Porous Media 9(5):429–444)—les performances du modèle ANN proposé sont éprouvées sur un système aquifère bidimensionnel hypothétique en régime transitoire. Dans le modèle ANN proposé, les coordonnées cartésiennes des piézomètres, leurs niveaux piézométriques et les temps d’observation sont utilisés comme entrées, et les valeurs de transmissivité et coefficients d’emmagasinement correspondants comme sorties. Les procédures de mise en œuvre, de validation et de test du modèle ANN suivent deux scénarios différents. Dans le scénario 1, toutes les données acquises sont utilisées lors de la simulation, tandis que les données présentent des lacunes dues à l’irrégularité des observations dans le scénario 2. Les distributions des transmissivités et coefficients d’emmagasinement sont estimées à partir des pondérations du réseau synaptique. Les résultats ont démontré que le modèle ANN développé peut être utilisé dans le cas de problèmes simultanés d’estimation des paramètres des aquifères.ResumenSe propone un modelo de red neural artificial (RNA) para la determinación simultánea de las distribuciones del coeficiente de almacenamiento y de transmisividad de un sistema de acuífero heterogéneo. Las RNA pueden ser herramientas útiles en problemas de identificación de parámetros debido a su capacidad para resolver problemas complejos no lineales. Como parte de la ampliación de un estudio previo—Karahan H, Ayvaz MT (2006) Forecasting aquifer paramters using artificial neural networks [Predicción de parámetros de acuífero usando redes neurales artificiales], J Porous Media 9(5):429–444—se evalúa el desempeño del modelo propuesto RNA en un sistema acuífero hipotético de dos dimensiones en condiciones de flujo transitorio. En el modelo RNA propuesto se usan como entradas las coordenadas Cartesianas de pozos de observación, presiones piezométricas asociadas y tiempo de observación mientras que los valores correspondientes del coeficiente de almacenamiento y de transmisividad se usan como salidas. Se utilizan dos escenarios para los procesos de evaluación, validación y entrenamiento del modelo RNA. En el escenario 1 todos los datos muestreados se usan a través del tiempo de simulación. Sin embargo, en el escenario 2 existen brechas en los datos debido a observaciones irregulares. Mediante el uso de pesos de redes sinápticas determinados se predicen distribuciones de coeficiente de almacenamiento y de transmisividad. Además se evalúa el desempeño de la RNA propuesta para distintas condiciones de datos con ruido. Los resultados muestran que el modelo RNA puede ser usado en problemas de estimación simultánea de parámetros de acuíferos.
[1]
Lefteri H. Tsoukalas,et al.
Fuzzy and neural approaches in engineering
,
1997
.
[2]
Srinivasa Lingireddy,et al.
AQUIFER PARAMETER ESTIMATION USING GENETIC ALGORITHMS AND NEURAL NETWORKS
,
1998
.
[3]
M. Tamer Ayvaz,et al.
Forecasting Aquifer Parameters Using Artificial Neural Networks
,
2006
.
[4]
M. Tamer Ayvaz,et al.
Groundwater Parameter Estimation by Optimization and Dual Reciprocity Finite Differences Method
,
2005
.
[5]
Luis A. Garcia,et al.
Using neural networks for parameter estimation in ground water
,
2006
.
[6]
Ashish Sharma,et al.
A study of optimal model lag and spatial inputs to artificial neural network for rainfall forecasting
,
2000
.
[7]
Mustafa M. Aral,et al.
Aquifer parameter and zone structure estimation using kernel-based fuzzy c-means clustering and genetic algorithm
,
2007
.
[8]
Ferenc Szidarovszky,et al.
Artificial Neural Network Approach for Predicting Transient Water Levels in a Multilayered Groundwater System under Variable State, Pumping, and Climate Conditions
,
2003
.
[9]
Enrico Zio,et al.
Approaching the inverse problem of parameter estimation in groundwater models by means of artificial neural networks
,
1997
.
[10]
J. C. Ramírez,et al.
Estimation of aquifer parameters under transient and steady-state conditions
,
1984
.
[11]
Giovanna Castellano,et al.
An iterative pruning algorithm for feedforward neural networks
,
1997,
IEEE Trans. Neural Networks.
[12]
Keith Beven,et al.
A Bayesian approach to stochastic capture zone delineation incorporating tracer arrival times, conductivity measurements, and hydraulic head observations
,
2003
.
[13]
A. Cheng,et al.
Aquifer parameter estimation by extended Kalman filtering and boundary elements
,
1997
.
[14]
Jagath J. Kaluarachchi,et al.
Application of artificial neural network and genetic algorithm in flow and transport simulations
,
1998
.
[15]
Richard L. Cooley,et al.
A Comparison of Several Methods of Solving Nonlinear Regression Groundwater Flow Problems
,
1985
.
[16]
Bithin Datta,et al.
Optimal Management of Coastal Aquifers Using Linked Simulation Optimization Approach
,
2005
.
[17]
Luis A. Garcia,et al.
Parameter Estimation in Groundwater Hydrology using Artificial Neural Networks
,
2003
.
[18]
K. El Harrouni,et al.
Groundwater parameter estimation by optimization and DRBEM
,
1997
.
[19]
Simon Haykin,et al.
Neural Networks: A Comprehensive Foundation
,
1998
.
[20]
Yoshio Hirose,et al.
Backpropagation algorithm which varies the number of hidden units
,
1989,
International 1989 Joint Conference on Neural Networks.
[21]
S. P. Neuman.
A statistical approach to the inverse problem of aquifer hydrology: 3. Improved solution method and added perspective
,
1980
.
[22]
K. Lakshmi Prasad,et al.
Estimating net aquifer recharge and zonal hydraulic conductivity values for Mahi Right Bank Canal project area, India by genetic algorithm
,
2001
.
[23]
Nikola IL Kasabov.
Adaptable neuro production systems
,
1996,
Neurocomputing.
[24]
Alex S. Mayer,et al.
Development and application of a coupled-process parameter inversion model based on the maximum likelihood estimation method
,
1999
.
[25]
T. Ulrych,et al.
A full‐Bayesian approach to the groundwater inverse problem for steady state flow
,
2000
.
[26]
Richard L. Cooley,et al.
Incorporation of prior information on parameters into nonlinear regression groundwater flow models: 1. Theory
,
1982
.
[27]
S. P. Neuman,et al.
Estimation of aquifer parameters under transient and steady-state conditions: 2
,
1986
.
[28]
S. P. Neuman,et al.
A statistical approach to the inverse problem of aquifer hydrology: 1. Theory
,
1979
.
[29]
Paulin Coulibaly,et al.
Groundwater level forecasting using artificial neural networks
,
2005
.
[30]
M. Eppstein,et al.
SIMULTANEOUS ESTIMATION OF TRANSMISSIVITY VALUES AND ZONATION
,
1996
.
[31]
A Dirac-δ Function Notation for Source/Sink Terms in Groundwater Flow
,
2005
.
[32]
Holger R. Maier,et al.
Neural networks for the prediction and forecasting of water resource variables: a review of modelling issues and applications
,
2000,
Environ. Model. Softw..
[33]
P. Kitanidis,et al.
An Application of Bayesian Inverse Methods to Vertical Deconvolution of Hydraulic Conductivity in a Heterogeneous Aquifer at Oak Ridge National Laboratory
,
2004
.