Uncertainty analysis of groundwater solute transport based on surrogate model
OUYANG Qi1,2, LU Wen-xi1,2, HOU Ze-yu1,2, GU Wen-long1,2, XIN Xin1,2
1. Key Laboratory of Groundwater Resources and Environment, Ministry of Education, Jilin University, Changchun 130021, China;
2. College of Environment and Resources, Jilin University, Changchun 130021, China
In order to analyze the influence that parameter uncertainty has on groundwater solute transport numerical simulation, this study adopted Monte Carlo simulation to conduct an uncertainty analysis on an example and illustrated its results from the perspective of risk assessment. For the purpose of reducing computation load, Sobol' method was used to analyze the sensitivity of model parameters, which helped to select the more sensitive parameters as random variables and to construct the Kriging surrogate model of the simulation model. This surrogate model provided further help in achieving the Monte Carlo simulation. Results showed:when the confidence was 80%, the confidence intervals of well 1, 2, 3 were 23.46~42.06, 47.99~66.73, 69.54~82.94mg/L, respectively. Combined with risk assessment, the risk of groundwater contamination was calculated as 0.54, which could serve as a scientific guide for the prevention and control of groundwater pollution.
Feyen L, Caers L. Quantifying geological uncertainty for flow and transport modeling in multi-modal heterogeneous formations[J]. Advances in Water Resources, 2006,29:912-929.
