抽水试验中两种直线图解法的对比

Comparison on Two Kinds of Linear Graphic Approaches in Pumping Test

应用定流量非稳定流抽水试验获取含水层水文地质参数时,常用的求解方法是Jacob直线图解法,而拟合直线的方法一般是基于最小二乘理论,但该方法易受异常值影响,拟合结果常常不理想,从而导致求解的参数值误差较大。相比最小二乘法,基于信息损失最小化的最频值算法对异常值不敏感,是一种更为稳健的算法。将最频值算法与Jacob直线图解法耦合,推导出了求解导水系数和贮水系数的新公式。最后,通过MATLAB环境下自行编制的程序,以受到不同水位影响效应的3个抽水试验为例,对两种直线图解法进行了对比分析。结果表明,基于最频值算法的直线图解法求参结果更准确可靠。

In order to acquire aquifer hydrogeologic parameters by unsteady flow pumping at a constant rate,the commonly used way is Jacob linear graphic approach,and the method of fitting line is basing on the least squares principle.But this method is easily influenced with outliers,and the fitting result is usually not satisfying,thus it leads to the errors of parameters becoming large.Compared with the least squares principle,the most frequent value method（MFV）,which is theoretically derived from the minimization of information loss,is insensitive to outliers,so it＇s a robust and resistant algorithm.In this paper,the MFV algorithm is coupled with linear graphic approach,and then derived the new formulae of transmissivity and storage coefficient.At last,the authors compiled the linear graphic program under MATLAB,and took three pumping tests,which suffered from interference factors in different state,as examples to compare the two kinds of linear graphic approach.The results showed that it＇s more reliable to get hydrogeologic parameters with linear graphic approach based on the most frequent value method.