包气带水分特征曲线参数模型及算法分析

目的]为包气带水分特征曲线参数寻找一种合理有效的拟合方法。方法]结合试验实测数据，利用非线性无约束算法、非线性最小二乘法和粒子群优化算法对模型中的参数进行拟合。结果]在对Brooks．Corey模型采用非线性回归分析、优化算法拟合的参数中，拟合的θr分别为0．830和0．349，显然脱离了物理背景；非线性回归分析拟合VanGenuchten模型中的参数α显然脱离物理背景。不同算法对Gardner模型中参数影响不大。VanGenuchten模型能完整地刻画整个脱湿曲线，Brooks—Corey模型和Gardner-Russo模型在含水率接近饱和状态时不适用或在进气值处存在不连续。VanGenuchten模型在不同算法之间的差别较小，Brooks-Corey模型和Gardner-Russo模型在不同算法之间的差别较大。非线性回归分析拟合的误差最小，优化算法次之。结论]非饱和带水分特征曲线参数模型的选用不能脱离参数的物理背景。

Parameter Models of Water Retention Curve in Unsaturated Zone and Analysis on Their Algorithms

Objective] The purpose was to find a reasonable and effective fitting method for the parameters of water retention curve in unsaturated zone.Method] Combining with experimental measured data,the parameters of the model were fitted by nonlinear unconstrained algorithm,nonlinear least square method and particle swarm optimization algorithm.Result] Among the parameters of fitting by nonlinear regression analysis and optimization algorithm of Brooks-Corey model,θr values of fitting were 0.830 and 0.349 resp.and they were obviously divorced from physical background.The parameter α of fitting by nonlinear regression analysis of Van Genuchten model was obviously divorced from physical background.The influences of different algorithms on the parameters of Gardner model were little.The whole drainage curve could be completely depicted with Van Genuchten model.Brooks-Corey model and Gardner-Russo model were not applicable when the water content was close to saturated state or they had discontinuity at air-entry value.Among different algorithms,Van Genuchten model had less difference and Brooks-Corey model and Gardner-Russo model had more difference.The error of fitting by nonlinear regression analysis was least and that of fitting by optimization algorithm was secondary.Conclusion] The selection of parameter model of water retention curve in unsaturated zone could not be divorced from physical background.