Artificial Neural Networks Based Approach for Identification of Unknown Pollution Sources in Aquifers

This work focuses on groundwater resources contaminations identification. The problem of identifying an unknown pollution source in polluted aquifers, based on known contaminant concentrations measurement in the studied areas, is part of the broader group of issues, called inverse problems. In this field, often pollution may result from contaminations whose origins are generated in different times and places where these contaminations have been actually found. To address such scenarios, it is necessary to develop specific techniques that allow to identify time and space features of unknown contaminant sources. The characterization of the contaminant source is of utmost importance for the planning of subsurface remediation in the polluted site. In this work, such identification is solved as an inverse problem in two stages. Firstly a Multi Layer Perceptron neural network is trained on a set of numerical simulations, and then the case under study is reconstructed by inverting the neural model.