10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010 Identification of the unknown pollution Identification of the unknown pollution source in the Alsatian aquifer (France) source in the Alsatian aquifer (France) through groundwater modelling modelling and through groundwater and Artificial Neural Networks applications Artificial Neural Networks applications Ing. Maria Laura Foddis University of Cagliari – Italy University of Strasbourg – France Department of Land Engineering (DIT) Laboratory of Hydrology and Geochemistry Section of applied Geology and applied geophysics of Strasbourg (LHyGeS)
INTRODUCTION Pollution may results from contamination whose origins are generated at different times and places where these contaminations have been actually found. Such situations needs to develop techniques that allow identify unknown contaminant sources behaviour in time and space The purpose of this work aims at studying the spreading of a dangerous chemical - carbon tetrachloride (CCl 4 ) - that contaminated, due to a tanker accident in 1970, a part of the largest aquifer in Western Europe: The Alsatian aquifer (France) The exact amount of the chemical infiltrated is unknown and this constitutes the main issue for its individuation and remediation. 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
OBJECTIVE OF THE RESEARCH The objective of the research is to find a solution of the inverse problem for the Alsatian aquifer: � using known contamination concentration data from pumping wells � behaviour and temporal evolution of the unknown pollution source is reconstructed 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
Upper Rhine Graben valley GEOGRAPHICAL LOCATION OF THE ALSATIAN AQUIFER 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
STRUCTURE OF THE AQUIFER Alsatian Region Continental extensive alluvial aquifer Phreatic aquifer fed by meteoric deposits and drained mainly by rivers and human activities. It has a Structure layered with a random superposition of different alluviums Monitoring wells Monitoring wells Pumping wells Pumping wells 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
HYSTORY OF THE AQUIFER POLLUTION BY CCl 4 11 December 1970 - accident An unknown quantity of carbon tetrachloride (CCl 4 ) spreading in the accident area ~ 4m 3 (data SGAL - Service Géologique d’Alsace-Lorraine) 1991 – Erstein – first analyses carried out by BRGM G é ologique (BRGM : Bureau de Recherche et Mini è re) CCl 4 � 62,4 ÷ 56,2 μ g/l These quantities exceeded the safe limits recommended by the World Health Organization � 2 μ g/l This high level of CCl4 concentration has caused serious problems in the region by contaminating the most important drinking water source in the area 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
PHYSICAL AND CHEMICAL PROPERTIES OF CCl 4 Volatile Organic Chemical does not naturally occur in the environment. It is miscible with most aliphatic solvents but has low solubility in water. 153,84 g/mol Molecular weight 0,52 mg/l Concentration limit that permit the perception of its sweet smell in water Boiling point 76,6 °C Solubility at 20°C 800 mg/l Density 1,59 mg/l at 20°C K oc – Soil Sorption Coefficient 71 Vapor pressure 91,3 mm Hg a 20°C 3,04*10 -2 atm-m 3 /mol a 24,8°C Henry’s Law Constant DNAPL – Dense Non Aqueous Phase Liquids Convection In Alsace aquifer CCl 4 and its toxic constituents have actions of: Dispersion Diffusion Volatilization Are insignificant: Sorption 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
DIFFICULTIES OF THE PROBLEM SOLUTION � High uncertainty in the aquifer formation because of its heterogeneity and its different permeable layers. � The exact amount of the chemical infiltrated and the source behaviour and how the pollutant feeds the contamination is unknown. Carbon tetrachloride has low solubility in water, it represents a � continuous source of contamination for the groundwater. � The chemical properties such as the solubility in water, diffusion, volatilization, and degradation coefficients are uncertain. 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
MATERIAL & METHODS The inverse problem for the Alsatian aquifer is solved using Artificial Neural Network (ANN) The ANNs are used in processing information. These have the ability to solve complex problems and the capacity to approximate any input-output relationship. Training set Validation set The Patterns for ANN Test set This patterns are based on flux and transport model in porous media of CCl 4 contamination in the studied domain. 26 different scenarios of source contamination behaviour are performed with a 3D model of the Alsatian aquifer created by Fluid and Solid Mechanics Institute of Strasbourg using TRACES SOFTWARE. TRACES � Transport or Radio Activer Elements in the Subsurface Hoteit et Ackerer, 2003 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
ALSATIAN AQUIFER NUMERICAL MODEL Time of simulation: Tempo di simulazione: 5 anni Tempo di simulazione: 5 anni 20000 days ~ 54 years Time of source activity: 11520 days~ 32 years Distribution of CCl 4 concentration after 5 years of the accident 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
ALSATIAN AQUIFER NUMERICAL MODEL Tempo di simulazione: 10 anni Tempo di simulazione: 10 anni Distribution of CCl 4 concentration after 10 years of the accident 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
ALSATIAN AQUIFER NUMERICAL MODEL Tempo di simulazione: 21 anni Tempo di simulazione: 21 anni Distribution of CCl 4 concentration after 21 years of the accident 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
ALSATIAN AQUIFER NUMERICAL MODEL Tempo di simulazione: 28 anni Tempo di simulazione: 28 anni Distribution of CCl 4 concentration after 28 years of the accident 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
ALSATIAN AQUIFER NUMERICAL MODEL Tempo di simulazione: 54 anni Tempo di simulazione: 54 anni Distribution of CCl 4 concentration after 54 years of the accident 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
PATTERN CONSTRUCTION FOR THE ANN The examples obtained with TRACES consisted of matrix of size [m,n] � Input matrices are composed of 4 columns: one for each layer in the source. � Output matrices are composed of 45 columns: one for each well. � In both matrices, rows represent time. Matrices were too large to be processed through the ANN For each example we calculated: � 2-D discrete Fourier transform (FFT) Among the frequency components of the FFT only the most significant were considered, and the remaining ones were set to zero � patterns are normalized in the range [-1 ; +1] 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
TRAINING OF THE NEURAL NETWORK MODEL To construct the ANN several trials were performed to define: � number of neurons � 11-11-56 � number of hidden layers � 1 � number of training epochs � 100 � activation functions for each layer The ANN are developed with the Neural Network Toolbox of MATLAB 7.1 ⎧ ⋅ + = Input layer W x b y ⎪ 1 1 During the training the connection weights = σ ⎨ Hidden layer h ( y ) were modified in order to minimize the error ⎪ ⋅ + = Output layer W h b u ⎩ on the training set. 2 2 At the same time the error was calculated Algebraic equations systems also on a validation set, independent from of the trained ANN the training set. that realize the relationship As the validation error began to rise, the between input and output training process is interrupted. patterns 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
INVERSION OF THE NEURAL NETWORK MODEL Once the training phase was completed and all the weights were determined, the inversion of the network could be performed. Knowing the output of the ANN, which derived from a set of measurements in the wells, the corresponding input could be calculated. On the basis of the third started by the know output u equation showed: were determined ( ) − the vectors h ( ) 1 = T ⋅ ⋅ T ⋅ − h W W W u b 2 2 2 2 ( ) − = σ 1 the vectors y y h − = ⋅ 1 x W b the inverted ANN input pattern x . 1 1 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
INVERSION OF THE NEURAL NETWORK MODEL The real and the inverted input patterns in frequency domain. There is a strong correspondence between the two patterns. 10th World Wide Workshop for Young Environmental Scientists 31 May - 4 June 2010
Recommend
More recommend
