基于PEST的地下水模型参数优化及不确定性分析

Parameter Optimization and Uncertainty Analysis of Groundwater Model Based on PEST

由于地下水系统中相关参数的时空变异性,利用地下水模型得到的计算值与真实值之间通常会出现一定的差异。介绍了参数优化(parameter estimation,PEST)程序的基本原理,并以建立的某地浸采铀单元井场地下水流动模型作为研究实例。结果表明,以水头实测值作为校准数据,PEST能够对渗透系数和贮水率取得很好的优化效果;目标函数的变化趋势总体是下降的,但并不总是不断减小的;参数的灵敏度在优化过程中总体处于较低水平;参数优化幅度与参数不确定性之间呈正相关关系。研究结果可为地下水模拟不确定性分析提供参考。

Because of the spatiotemporal variability of the relevant parameters in the groundwater system,there is often some deviation between calculated result of groundwater model and real value. The basic principle of parameter estimation (PEST) which is a parameter optimization tool was introduced. And established the groundwater flow model of a unit well site of in-situ leaching uranium mining as an example. The results show that PEST can achieve good optimization effect for the hydraulic conductivities and the specific storativities when the measured water head as the calibration data. The change trend of the objective function is generally decreasing,but it is not always decreasing. The sensitivity of parameters is roughly at a low level in the optimization process. There is a positive correlation between the amplitude of parameter optimization and the uncertainty of parameters. The research results can provide reference for the uncertainty analysis of groundwater simulation.