Groundwater Flow Simulation And Uncertainty Analysis At Feifeng Mountain Disposal Site

In recent years,with the development of nuclear energy,the disposal of nuclear waste has become an important environmental issue.In the long term,once the engineering barrier is broken,nuclear waste can enter the human environment through the atmosphere,surface water and groundwater.The groundwater transport is the main channel.It is of great significance to evaluate the safety of disposal facilities by studying the dynamic characteristics of groundwater by means of numerical simulation.Due to randomness of model parameter,model simulation results contain large uncertainties.through the analysis of the uncertain model,it is necessary to improve the reliability of the model,however,in the process of uncertainty analysis,groundwater model will spend a lot of time,with a huge computational load,alternative model is an effective way to solve this problem.In this paper,a hydrogeosphere software was used to establish a groundwater numerical model to simulate and predict the temporal and spatial variation of the groundwater level in the Feifeng Mountain low-medium disposal site.Then the local sensitivity analysis,Latin hypercube sampling and support vector machine(SVM)were combined to build a substitute model for the groundwater model,which greatly reduced the running time of the model.Finally,Monte Carlo method is used for parameter uncertainty analysis to analyze the influence of model parameter changes on the prediction results of groundwater model.After the above research,the following conclusions are drawn:(1)The groundwater of the disposal platform mainly collects in the north and east directions,and the fastest time to reach Pingxi River is about 32 years.The model is used to predict the change of groundwater level from 2020 to the end of this century.The prediction results show that the overburden plays a good role in protecting the disposal reservoir and effectively blocks the erosion of the disposal reservoir caused by rainfall.The groundwater level of the disposal unit will decrease by about 0.7m by 2100,which is lower than the elevation of the disposal platform and will not have an impact on the disposal facilities.(2)The sensitivity analysis results showed that the runoff production coefficient and the reciprocal value of the intake of the strong weathered layer were highly sensitive,0.542 and 0.522 respectively,which had the greatest impact on the groundwater level of the model and were selected as random variables for the next uncertainty analysis.(3)Support vector machine(SVM)method is used to construct an alternative model of groundwater model.The verification results show that the fitting determination coefficients of the output values of the simulation model and the alternative model are both above 0.96,and the mean square error is below 0.03.The model can realize the function of groundwater model with high accuracy,and the uncertainty analysis of the next step is carried out using the model,saving about90% of the time in the calculation process.(4)The uncertainty analysis of the prediction results of groundwater model is carried out by Monte Carlo method,and the parameter statistics and interval estimation are carried out.The results show that the groundwater level approximately obeys the normal and skewed distribution.The higher the confidence level is,the larger the range is,and the total water level is between 598-601 m.