Improvement And Application Of Groundwater Flow Stochastic Simulation Method Based On T-PROGS

The groundwater system is a complex system with multiple factors,in which the hydrogeological structure directly determines the storage condition of groundwater.It is necessary to explore the aquifer situation before studying the flow of groundwater.However,the uncertainty of hydrogeological structure will be caused by the sedimentation,climate,and limited geological exploration condition.Therefore,the establishment of a reliable hydrogeological structure model to simulate the 3D distribution of lithology is of great significance to the study of groundwater flow.In a region with complex geological conditions,a simple deterministic model can't fully characterize the structural features,so establishing stochastic models can provide better choice.In this paper,a geostatistical method based on transition probability(TPROGS)is used to develop stochastic model of hydrogeological structure.And the models are coupled with the numerical model of groundwater.(1)The important parameters in each module of T-PROGS are explained and discussed in detail.It shows that the lag spacing,data searching control parameters,grid size,simulated annealing parameters,etc.can be determined according to the specific region.For the lateral Markov chain,the lithological mean length is uncertain,and the simulation results are sensitive to it.The larger the mean length,the stronger the structurality of the model and the weaker the randomness;the smaller the mean length,the weaker the structurality of the model and the stronger the randomness.(2)A method of establishing stochastic hydrogeological structure model based on boreholes and profiles is proposed.The profiles are important data to characterize the lateral lithology distribution,from which virtual boreholes are extracted,and uncertainty analysis is carried out according to the distance between virtual boreholes and adjacent controlled boreholes in profiles.So the virtual boreholes are added to the hydrogeological structure modeling as “soft data” with probability.Finally,an example is given to show that the reliability is increased by about 146% and uncertainty is reduced by about 9.8% after adding virtual boreholes,which proves the reliability and feasibility of this method.(3)The coupling of stochastic hydrogeological structure model and numerical simulation of groundwater is realized.The relationship between lithology and hydrogeological parameters is quantitatively represented by quadratic programming method,so the stochastic hydrogeological parameter field with practical geological significance can be established to better represent the characteristics of aquifer variability.(4)A stochastic hydrogeological structure model and a stochastic groundwater model are established in the West Liaohe Plain.The aquiclude medium or aquitard medium in the West Liaohe Plain distributes in the north of the region with a probability of 0.6.The uncertainty analysis of groundwater level predicted by stochastic groundwater model after ten years shows that the average variation of groundwater level is between positive and negative 5m,and the confidence interval width of 90% confidence water level is about 0.292 m..