The use of genetic algorithms for determining the transport parameters of core experiments

From hydrocarbon reservoirs, brine is produced as a waste material, which m ay be injected into the ground or discharged at the surface. When the waste water is injected into the ground, it may be mixed with fresh-water sources by several processes. Groundwater contamination from leakage, spills, or t he injection of hazardous or toxic materials is widely regarded as one of t he leading environmental problems. This study presents the use of genetic a lgorithms (GAs) as a viable means of estimating the transport parameters su ch as dispersivity, retardation factor, and diffusion coefficient of water- saturated porous media. The unknown transport parameters of advective-dispe rsive contaminant equations for homogeneous, linear, radial, and fractured systems are predicted by the use of GAs coupled with the experimental data. The parameter estimation study is considered as a constrained optimisation problem by minimising the total error between the calculated and the measu red effluent concentrations satisfying state equations, boundary conditions , and limits on parameters. The optimisation formulations combine advective-dispersive pollutant transp ort simulation with GAs optimisation. The calculated concentrations were ob tained from analytical solutions of these models. The results of the simula tion runs were calibrated with those obtained from the experimental tests. A satisfactory agreement between estimated and experimental results was ach ieved. Since all the unknown parameters in the models are found to be in a reasonable agreement with respect to the experimental conditions, GAs can b e said to be an effective optimisation method for estimating the transport parameters in groundwater contamination problems.