建立与真实状态相近的初始渗流场是正确分析边坡降雨入渗的重要前提。鉴于目前FLAC^3D中的饱和-非饱和渗流场初始化方法寥寥无几,提出并详细介绍了饱和-非饱和渗流场初始化的设置方法。根据5种分布形式的基质吸力,按照其基质吸力线性分布和非线性分布规律提出了2种对应的初始化方法:对于线性分布的基质吸力运用MIDAS建模和FLAC^3D内置FISH语言进行渗流场初始化设置;对于非线性分布的基质吸力,用函数来表示其复杂几何形态的地下水位面具有一定的难度,不适合大面积推广应用,将地下水位面简化成一平面后,可利用FISH语言编程进行渗流场初始化设置。通过与常规渗流场初始化方法的计算结果对比,初始渗流场对边坡降雨入渗最终计算结果有较大影响,初始渗流场与真实状态越相近,其计算精度越高。

The establishment of initial seepage field close to the real state is an important prerequisite for the correct analysis of slope rainfall infiltration. In view that methods of initializing saturated-unsaturated seepage field in FLAC^3D are rarely reported, in this paper we put forward and introduce in detail the setup method of initializing saturated-unsaturated seepage field. According to the five distribution forms of matric suction, two initialization methods are proposed corresponding to the linear and non-linear distribution laws of matric suction. For the linear distribution of matric suction, MIDAS modeling and FLAC^3D built-in FISH language are employed to initialize the seepage field. For non-linear distribution, since it is difficult to represent the complex geometry of groundwater level surface by using function, the groundwater level is simplified into a plane for the initial setting of seepage field by using FISH programming language. Comparison with conventional initialization method demonstrated that the initial seepage field has a great influence on the final calculation result of slope rainfall infiltration. When the initial seepage field is closer to the real state, the calculation accuracy is higher.