La Modélisation Numérique Dans L’Identification de Sources Inconnues de Pollution

Abstract

Groundwater pollution is an increasing problem in water supply.The various kinds and large number of pollutants associated with drilling costs put a great emphasis on the strategies developed to search the unknown position and intensity of some pollution sources.The proposed approaches consist of comparing the pollutant dispersion field resulting from a supposed pollution source with real values taken in observation wells, and to iterate the process until the most probable source is located.Three models are based on the iterative calculus of the differences between computed values of concentration by a dispersion simulation program and observed values, this difference being supposed linear and respectively given under absolute, relative or standardized form.The scattering of the resulting most probable source positions is summarized in a confidence index. The results are compared with tests conducted on a large experimental porous medium tank and show acceptable diagnosis of the pollution source character...