Groundwater Pollution Source Identification Using Genetic Algorithm Based Optimization Model

Abstract

Groundwater is an important natural resource available on the earth. Contamination of groundwater resources has become a major problem today due to some artificial and natural activities. Identification of groundwater pollution sources is a major step in groundwater pollution remediation. A pollution source is said to be known only when its source characteristics (location, strength and duration of pollution activity) are known. Identification of unknown groundwater pollution source is an inverse problem, which is generally ill posed due to existence of local minima. This problem becomes more complex for real field conditions, when the lag time (defined as the time difference between the first reading at the observation well and the time when source becomes active) is not known. Genetic Algorithm (GA) based simulation optimization methodology has been used in this study for complete identification of unknown groundwater pollution source. GA is non-gradient based search technique and it is capable of finding the global optimum. A contaminant transport model, which can simulate the concentrations at the observation well location, is combined with the GA based optimization simulation model. The performance of the developed methodology is evaluated for one and two dimensional cases with error free and erroneous concentration measurement data. Performance results show the capability and practical applicability of proposed methodology. Main advantage of the proposed methodology is that complete identification of unknown groundwater pollution source is possible with the help of only one observation well when the lag time is also not known.