基于遗传算法的阶梯流量抽水试验反演水文地质参数方法研究

Study on step pumping test inversion of hydrogeological parameters based on genetic algorithm method

为更好的利用抽水试验实测数据,本文使用遗传算法进行阶梯流量水文地质参数的求算。通过建立理想模型对遗传算法的选择方式、变异方式、交叉方式进行对比研究,得出在随机均匀选择、分散交叉、线性约束变异的参数设置下算法稳定、高效,结果证明遗传算法在求参的精度和稳定度方面均优于传统配线法与试算法。将其应用到北京市应急水源地水文地质参数反演中,求得导水系数T为44.64m2/d,储水系数μ*为9.30×10-4。

In order to make full use of the pumping test data,step pumping hydrogeological parameter is calculated by genetic algorithm.Comparison study on the selection methods,crossover methods and mutation methods is performed through an ideal model.It is proved that the calculation is more precise and stable under the parameter setting of stochastic uniform choosing,scattered crossover,constraint dependent mutation method.The genetic algorithm is proved superior to traditional curve fitting method and trail method.The method is applied in the Beijing emergency water source hydrogeological parameter calculating and derived that T is 44.64m^2/d and μ＊is 9.30 × 10^-4.