摘要
Contamination of groundwater poses serious threat to the human health and environment. It is difficult and expensive to clean up contaminated aquifers. Identification of unknown pollution sources is vital for adopting any remediation strategy. Groundwater flow and transport simulation model is used to generate necessary data for Artificial Neural Networks (ANN) model building processes. Breakthrough curves obtained for specified pollution scenario are characterized to reduce the inputs to ANN model. The characterized breakthrough curves parameters serve as inputs to ANN model. Unknown pollution source characteristics, flow parameters and transport parameters are outputs for ANN model. Identification of sources is performed with considerations of three cases—simultaneous estimation of unknown sources and flow parameter; simultaneous estimation of unknown sources, flow and transport parameters; and simultaneous estimation of unknown sources and boundary head. Characterization of uncertainty in source identification due to uncertainty in flow parameter, uncertainty in transport parameter and uncertainty in constant head boundary estimation is performed using fuzzy vertex alpha-cut techniques. Keywords Uncertainty characterization Groundwater flow and transport Pollution source identification Parameter estimation ANN Fuzzy vertex method