Study On Parameter Stochastic Groundwater Numerical Simulation Based On Improved Latin Hypercube Sampling

Numerical simulation is an important method of studying the law of groundwater flow and solute transport. There is a limitation in deterministic numerical simulation for describing the complicated hydro-geological structure due to spatial variability of porous media. Groundwater flow stochastic numerical simulation and solute transport stochastic numerical simulation are established here to simulate the distribution of head and solute concentration on the condition of exploitation in the vicinity of river.The analysis of groundwater stochastic numerical simulation methods is presented in the thesis firstly. Then, a new sampling method is achieved by incorporating Dual variable into Latin Hypercube Sampling, sampling process of the new method is realized by the computer program written in c# language. Through the analysis of sample statistics and distribution characteristics,it is concluded that there is an increase of improved Latin Hypercube Sampling in degree of dispersion.On the basis of detailed data of Hun river alluvial and flood fan, hydro-geological conceptual model is established, samples of stochastic parameters, hydraulic conductivity, porosity and dispersion are obtained from LHS and improved LHS, a parameter stochastic flow model and a parameter stochastic solute transport model are established sequentially, through analysis of sample statistics and distribution characteristics of simulated head and simulated solute concentration, there is a decrease of simulated head and simulated solute concentration in standard deviation, the difference of two methods in simulated head has relationships with sample distribution of two methods and variance ration between head and hydraulic conductivity, the difference of two methods in simulated solute concentration has relationships with sample distribution of two methods and variance ration between solute concentration and porosity.The coverage area of solute concentration of exceeding certain standard in quality standard for groundwater is predicted by means of solute transport stochastic numerical simulation, thus providing essential principles for the remediation of groundwater contamination.