梯度正则化法在土壤非饱和水分运动参数反演中的应用

Application of gradient regularization method in the parameters identification of soil water movement

基于多孔介质流体动力学理论和参数反演辨识技术,建立了非饱和土壤水分运动反问题的数学模型及最优化模型,并采用梯度正则化方法对非饱和土壤水分运动参数进行反演辨识。数值算例反演结果表明,采用梯度正则化方法对解的先验信息依赖极小,解决了参数初值对反演结果的敏感性问题,且反演结果与测试结果具有很好的一致性。

Based on the theory of dynamics of fluids in porous media and parameters identification technology, an optimal model for describing soil moisture movement in unsaturated zone was developed. The parameters identification of soil water characteristic curve are investgated by adopting gradient regularization method. Numerical inverse results shows that the gradient regularization method has little dependence on the prior information of problem solution, which solves the sensitivity of the initial values to inverse results. The parameter estimation values is in good agreement with experimental data.