Numerical Model Of Groundwater Flow And Stochastic Parameter Simulation

Development of groundwater source field provides a rich source of water supply for national economy which is more important for dry areas. Making accurate assessment of groundwater resources and designing proper exploitation plans to ensure normal water supply and avoid environmental geological problems are very important issues before running a sourcefield. Numerical simulation method is more widely used for its advantages in convenient and flexibility for complex hydrogeological conditions.Taking Haolebaoji source field located in Ordos Basin for example, the thesis uses groundwater simulation software GMS to simulate groundwater flow. Based on the situation of study area, numerical model is constructed and used to predict the groundwater flow field under designed exploitation plans. What’s more, we also evaluate the amount of groundwater resources, exploitable quantity and analysis the environmental geological problems may arise under exploitation; Based on the determinate model, combined with the uncertainty of hydraulic conductivity, stochastic simulation of parameter is estimated. The random number is generated by Latin Hypercube Sampling (LHS) method and the number of stochastic realizations is 243. With the stochastic model, risk analysis of the exploitable quantity given by determinate model is performed and we get the probability distribution of study area when water level drawdown is greater than allowable.Results show that: Average annual recharge of groundwater of study area is 12.20×104m3/d (1985-2007). Under designed exploitation schemes, water level drawdown is lower than allowable during prediction period and water level trends stable at the end of prediction. Scheme II is more reasonable between two designed schemes and more beneficial to sustainable use of groundwater. Exploitable quantity of groundwater is 8×104m3/d. Large-scale exploitation of source field would cause some ecological damages, but also the phenomenon of salinization of low-lying areas will be improved.The stochastic parameter model, combines determinate model and uncertainty of parameter together, can reflect the impact of parameter randomictity on simulation results. The distribution of mean water level drawdown and standard deviation at the end of prediction (2014.10.31) are consistent with actual situation and the maximum risk reaches 70% when drawdown is greater than 35 meters.