Upscaled Modeling Of NAPLs Transport And Remediation In Groundwater

Upscaling is often employed in numerical simulation of groundwater flow because the parameters are measured in the field on a support scale which is much smaller than the discretization scale of the numerical model.A reasonable upscaling method can both improve the calculation efficiency and ensure the simulation accuracy.The main content of the thesis is about a study on upscaling of NAPLs transport and remediation in groundwater,where Laplacian-skin is used to scale up the permeability coefficient field,comparing with the arithmetic mean upscaling method.First,the method is applied to the upscaling of heterogeneous hydraulic conductivity fields in the numerical contaminant transport model of perchloroethylene(PCE),a typical non-aqueous phase liquids(NAPLs),in groundwater on the basis of the introduction of Laplacian with skin.Also,the simulation results are compared with those output from the fine scale model and the coarse scale model upscaled by arithmetic average.The case study shows that the PCE contaminant transport model established by UTCHEM(University of Texas Chemical Compositional Simulator),which is upscaled by Laplacian with skin can describe the spatial moments(including the zero moment,the first moment and the second moment)of the plume over time well,whether for single source situation or dual sources situation.The absolute value of the relative error of the zero moment is less than 0.25%.Moreover,the coarse scale model only takes about 3% of the calculation time of the fine scale,indicating that the methodology of Laplacian with skin is worth promoting in practice.Then,in order to find an appropriate permeability upscaling method which can greatly improve the computational efficiency of surfactant-enhanced aquifer remediation(SEAR)model while ensuring the accuracy of the simulation,two coarse-scale models based on Laplacian with skin method and the arithmetic mean upscaling method were applied in this study,and their performance was also compared with fine-scale models.The results indicated that the maximum calculation error of the residual mass of aquifer pollutants by the model based on Laplacian with skin was better than the model built by the arithmetic mean upscaling method in all cases.The superiority of Laplacian with skin became more significant when the aquifer heterogeneity was stronger.In addition,Laplacian with skin achieved a better simulation effect on the centroid location and the shape of the pollutant.Particularly,the use of coarse-scale model can greatly reduce the computational cost of SEAR,leading to the significant computational cost-savings,e.g.,about 6.5% of the original runtime by using the arithmetic mean upscaling method and 4.5% by Laplacian with skin.The coarse-scale models based on Laplacian-skin method perform well in NAPLs transport and remediation in groundwater.In the subsequent simulation of groundwater transport and remediation,the permeability coefficient upscaling method can be used to save computation costs while ensuring the accuracy of simulation.