| With the rapid development of petrochemical industry, organic pollution in soil and groundwater pollution problems become more prominent, Dense non-aqueous phase liquid, DNAPL, is one of the most common type of organic pollutants underground, and therefore an attractive research field. Current research focuses on DNAPL source identification, DNAPL migration and remediation. Before DNAPL Remediation, the source location and plume distribution are need to be determined, and delineation of the source location can greatly improve the remediation efficiency and reduce costs. After successfully locating sources, sound knowledge of the DNAPL migration underground can provide direct assistance to DNAPL remediation. Among various DNAPL remediation methods, bioremediation stand out and are widely used due to its environment-friendly and cost-effective feature, however, the bioremediation can change the medium’s permeability and dispersivity, which can then affect the remediation efficiency. Based on this, this paper studied three aspects:1) Applied the source identification algorithm based on the stochastic Monte Carlo simulation, Kalman filter and fuzzy set theory to a DNAPL contaminated site in Nanjing and identified the ture DNAPL source with minimum sampling; 2) Based on the previous experimental results of DNAPL migration in 2D sandbox using light transmission method, numerical simulation of multiphase flow in porous medium were then systematicaly studied.; 3) The study finally investigatged the impact of biogenic gas and biofilm on permeability and dispersivity using numerical modelling based on two-dimensional sandbox bioremediation experimental results. The main conclusions were as follows:1) An optimal source search strategy incorporating Monte Carlo method, Kalman filtering and fuzzy set theory, was applied to a contaminated site in Nanjing to define the 6 suspected multiple DNAPL source locations. Using 9 field and sample data, the algorithm identified two true sources (Source#1 located in the waste water basin and #4 located in the PF workshop) which agree with the field investigation and indicate the efficiency of the algorithm. However, in the site application, lack of informative sampling data and confined small plume resulted from the site’s low permeability made the plume can only update locally rather than globally, thus limited the algorithm’s application. Also, existence of multiple true sources, geological heterogeneity and the parallel-to-flow layout pattern of the sources influence the convergence of the algorithm (Source #2 and #3). The algorithm also rily on good intial estimation of the suspected sources.2) Based on the experimental results in 2D sandox using light transmission method, the system carried out numerical simulation of multi-phase flow in homogeneous and heterogeneous porous media, the results show that numerical simulation results generally agree with the experimental results on the movement of multiphase flow. The major difference reflected in water/gas two-phase flow and water/DNAPL two-phase flow were the irregular channels and preferential flow presented by the gas flow and DNAPL migration in experiments, also DNAPL saturation in certain areas increased irregularly, while the simulation results showed uniform and stable fluid upward (gas) or downward (DNAPL) seepage. These differences were caused mainly by "nonuniform sand packing process". In addition, quantitative study of the water/gas two-phase flow in heterogeneous porous media shows that 5316 local saturation data obatain from simulation results and experimental resutls spanned across 12 time series reach an R2 of 0.94, while the total gas volume calculated from simulation results and experimental results also agree well with each other, indicating that the simulation results coincided the experimental results from a statistical point and local point of view. In water/DNAPL/gas three phase flow experiments in homogeneous medium, the vertical migration of TCE got impeded by gas after TCE injection in the gas distribution zone. The modelling results show the TCE lateral plume widened after entering the saturated zone from unsaturated zone, which is against the experimental results. This may be related to TCE volatilization, or related to the limitation of the current light transmission model of three pahse, which cannot tell the light intensity change caused by TCE or the gas.3) Impact of Bioremediation on medium’s permeability were mainly the impact of the biofilm and biogenic gas on the permeability of porous media. Experimental and simulation results show that the Taylor model exaggerated the effect of biofilm blockage. Based on the K field calculated by Taylor Model and VG model, the combined effect of the biofilm and bubbles showing a srip pattern, that the permeability in the upper rigion of the sandbox is significantly reduced by the bubble, while the permeability in the bottom also reduced by mainly the biofilm blockage, for the middle region, the combined effect of both bubble and biofilm are small, so that the tracer is preferentially migrate in central area. Study then set low permeability zones resulted from bio-remediaiton in two group sandbox’s middle region and run the solute transport simulation, showing that the blockage caused by bioremediation can reduce remediaiton efficiency by 94-98%, showing that the permeability reduction caused by bioremediation had a huge impact on remediation efficiency. Though both experimental and simulation results indicate the biofilm and the bubble will increase the dispersivity, they only doubled the original dispersivity. Both biofilm-affected dispersivity calculated by T-J model 2 and biofilm-and-gas-affected dispersivity calculated by T-J model 2 and Sato model are larger than the actual dispersivity. This is mainly due to the T-J model 2 and Sato model were both derived from one-dimensional sand column experiment, where microbial activity can have a direct impact on the flow of water, while the water can flow around the biologically active zone in 2D sandbox thus lowering the impact of the bioremdiaiton zone on the flow. In addition, the breakthrough curve obtained from simulation results that incorapted the sorption effect fit well with the experimental results, indicating the sorption effect of the biofilm and the bubble on tracer. |
PDF Full Text Request