| The prediction of solute transport is significant for groundwater resource evaluation and groundwater contaminant remediation.Dispersivity characterizes the hydrodynamic dispersion strength of solute in porous media and exerts a significant influence on the solute transport process.The identification accuracy of dispersivity directly affects the prediction accuracy of solute transport.However,the inherent complexity and heterogeneity of the underground system leads to the scale effect of dispersivity.The dispersivity change significantly with travel distance,which greatly increases the uncertainty of solute transport prediction.Due to the difficulty in characterizing the heterogeneous structure of sediments,there is no complete theory to explain the correlation between the heterogeneous structure parameters of sediments and the dispersivity scale effect.How to accurately predict the change mechanism of dispersivity from small to large scale in heterogeneous sediments has become an urgent problem to be solved.In this paper,the sedimentary structure of the target aquifer in the field area was investigated based on borehole data and radar data.Then,a laboratory scale heterogeneous structure was designed according to the characteristics of the sedimentary structure.A monitoring optimization scheme based on information theory and data worth analysis is proposed.The heterogeneous structure and monitoring optimization scheme will be used in the subsequent heterogeneous solute transport experiments.Solute transport experiments at different scales were conducted in single media,layered media,and heterogeneous media,respectively,to investigate the variation of dispersivity with solute transport distance.The impact of media particle size,spatial dimension,and heterogeneity on the scale effect of dispersivity was analyzed.On the basis of the above studies,we carried out a theoretical study on the upscaled model of dispersivity.The multi-level sedimentary structure framework was used to describe the process of sediments from simple to complex.The spatial correlation structure between different facies types was characterized by the transition probability theory.The upscaled models of dispersivity for bounded and unbounded media were developed based on facies volume proportion,mean facies length,indicator correlation scale,integral scale,length-width ratio of flow velocity field,mean and variance of conductivity.Finally,the prediction accuracy of the upscaled model at the laboratory scale was verified by solute transport experiments.The feasibility of the upscaled model at the field scale was analyzed by numerical simulation.Through the above research,the following conclusions are obtained:(1)The geostatistical method based on transition probability is an effective means to conduct analogy simulation of solute transport experiments.This method can construct the heterogeneous structure with similar geostatistical characteristics to target aquifers according to heterogeneous structure indexes such as volume proportion,mean facies length,and transition probability function in sediments.It can be used to guide heterogeneous structure design in solute transport experiments.(2)By balancing the number of monitoring points,the amount of monitoring data,and the worth of monitoring data,a monitoring scheme optimization method based on information theory and data worth analysis is proposed,which can be used to guide the monitoring network optimization and duration selection.The optimized monitoring network not only characterizes the distribution of contaminant plumes with high performance but promotes an accurate estimation of flow and transport parameters.With the extension of experiment time,the total information of the monitoring network increases to the maximum,while the uncertainty of estimation parameters decreases to the minimum.The suggested experimental duration is the critical value when the total information is stable,and the parameter uncertainty is reduced effectively.(3)Solute transport experiments indicate that the dispersivity scale effect is common in single media,layered media,and heterogeneous media.The dispersivity scale effect is influenced by solute transport space dimension,mixing effect in the well,and sediment heterogeneity.The impervious boundary limits the solute transport direction and weakens the variance of transport velocity,which inhibits the hydraulic dispersion and dispersivity scale effect.The mixing effect in the well not only enhances the hydrodynamic dispersion,but also significantly increases the growth rate of dispersivity with travel distance.Sediment heterogeneity leads to spatial variability of hydraulic conductivity and uneven velocity field distribution,significantly enhancing the hydrodynamic dispersion(4)The upscaled models of dispersivity for bounded and unbounded media were derived based on multi-level transition probability theory.This model can characterize the pre-asymptotic behavior of dispersivity with the increase of solute transport distance.The dispersivity upscaled model indicates that the scale effect of dispersivity is affected by facies volume proportion,mean facies length,anisotropy ratio,indicator correlation scale,integral scale,length-width ratio of flow velocity field,and mean and variance of conductivity.Among them,the indicator correlation scale and volume proportion have the most significant influence on the upscaled models of dispersivity.(5)The upscaled model can accurately capture the variation trend of dispersivity with solute transport distance at both the laboratory and field area scales.At the laboratory scale,the dispersivity at different distances is calculated by the solute transport experiment observation data and the upscaled model.The results show that the relative differences of the model for unbounded and bounded media are 9.8% and6.5%,respectively.At the study area scale,the dispersivity is calculated by particle random walk simulation and moment method.Two upscaled models slightly overestimate the dispersivity,with relative differences of 15.9% and 9.8%,respectively.In three indicator correlation scale ranges,the prediction results of the bounded upscaled model are generally consistent with the simulation results.In summary,the upscaled models of dispersivity derived in this paper based on transition probability theory can better predict the trend of dispersivity from small scale to large scale in 3D heterogeneous sediments,which is of great theoretical value for studying the contaminant transport behavior in site.It can provide a theoretical basis for formulating prevention and control strategies for groundwater contaminant. |
