Isogeometric analysis for unsaturated flow problems

Unsaturated flow problems in porous media often described by Richards’ equation are of great importance in many engineering applications. In this contribution, we propose a new numerical flow approach based on isogeometric analysis (IGA) for modeling the unsaturated flow problems. The non-uniform rational B-spline (NURBS) basis is utilized for spatial discretization whereas the stable implicit backward Euler method for time discretization. The nonlinear Richards’ equation is iteratively solved with the aid of the Newton–Raphson scheme. Owing to some desirable features of an efficient numerical flow approach, major advantages of the present formulation involve: (a) numerical oscillation at the wetting front can be avoided or facilitated, simply by using either an h-refinement or a lumped mass matrix technique; (b) higher-order exactness can be obtained due to the nature of the IGA features; (c) the approach is straightforward to implement and it does not need any transformation, e.g., Kirchhoff transformation or filter algorithm; and (d) in contrast to the Picard iteration scheme, which forms linear convergences, the proposed approach can however yield quadratic convergences by using the Newton–Raphson method for solving resultant nonlinear equations. Numerical model validation is analyzed by solving a three-dimensional unsaturated flow problem in soil, and its derived results are verified against analytical solutions. Numerical applications are then studied by considering three extensive examples with simple and complex configurations to further show the accuracy and applicability of the present IGA.     

Modeling of a semi-real injection test in sand

This article presents a model of flow and transport with filtration in porous media which is used to analyze large-scale grouting tests. A program based on the finite element method is developed to solve the model equations; a particular attention is paid to inherent issues of transport problems. The analysis of these tests aims at providing insights on the propagation patterns associated to the injection of cement-based grouts in sand. To apprehend particular features that characterize field injections, the experiments are performed by using a tube-a-manchette and a patented grout. Finally, the role of filtration during the tests is discussed.     

二维浅水明流的一种二阶高性能算法

为了精确求解二维浅水方程组,在由任意三角形构成的无结构网格上建立了有限体积MUSCL算法,这是文献[1]中有限体积一阶Osher格式的一类二阶推广.通过采用预测校正二步时间积分和单元内引入坡度限制,该算法在时空上均可达到二阶精度.跨单元边界的法向数值通量采用通量向量分裂公式(FVS)计算.最后,通过长江口南支潮流计算和一维瞬时溃坝模型算例来说明该格式的优良性能及在浅水流动计算中的应用.     

Tag retention and survival of fish tagged in controlled pond experiments

Fish survival, tag retention and tag loss were evaluated in pond experiments and described by linear regression equations. Tagging of fish had no significant influence on their survival. However, loss of tags caused strong differences between actual fish survival and the survival rate e^–z estimated on recovered tags. Mathematical functions were derived to correct tag loss. Values of those functions depend upon the type of tag and time after tagging. The B-type of tagging, which approximated the Swedish Carlin method, appeared to be most effective. The rate of tag loss was independent of the fish species (common carp, tench) used in the experiments.     

含自相容源的可积系统

含自相容源的可积方程（SESCSs）是一类重要的非线性偏微分方程,研究其可积性和寻求其精确解都具有重要的数学意义,SESCSs描述的是不同孤波间的相互作用,又有重要的物理意义。这类方程也正丰富了Lax可积系统的内容。从可积意义上详细介绍了含自相容源的可积方程,与求解此类方程有关的构造含自相容源的可积方程的方法。给出Lax可积的概念,作为例子,给出了含自相容源的非等谱KdV方程族、含自相容源的非等谱AKNS方程族及其Lax表示。     

应用二阶完全非线性Boussinesq方程模拟破碎波浪

建立了基于高阶Boussinesq水波方程的一维波浪破碎数值模型。基于一组具有二阶完全非线性特征的Boussinesq水波方程,建立了交错网格下的高精度差分格式,推导了适用于该组方程的永形波解析解,其和松弛造波技术相结合实现了数值波浪水槽中(强)非线性波浪的无反射入射。通过模拟封闭容器内水体晃动问题对数值格式进行了验证,通过模拟孤立波在斜坡海岸上的浅化过程说明了将方程从弱非线性发展到完全非线性的必要性。采用涡粘方法处理波浪破碎,利用物理模型实验数据,分析了模型中各波浪破碎参数对数值结果的影响并对参数进行了率定。应用该模型对规则波在斜坡海岸上的传播、变浅以及破碎过程进行了数值模拟研究,数值结果同实验数据吻合良好,验证了模型的有效性。     

泥石流的二维数学模型

泥石流是在重力作用下,由砂粒石块和水等组成的固液混合物,是一种发生于山区的复杂的地质灾害现象。泥石流主要是由暴雨诱发引起的,它沿着复杂的三维地形高速流动,具有流体流动的特性。为了模拟泥石流的运动规律,预测降雨诱发的泥石流的到达距离和泛滥范围,减少和避免泥石流引起的灾害,把泥石和雨水组成的固液混合物假定为遵循均匀、连续、不可压缩的、非定常的牛顿流体运动规律。基于质量守恒方程和Naiver-stokes方程,采用深度积分方法,推导出了一个模拟泥石流运动的二维数学模型。所有方程式可用有限差分法来求解。结合GIS,该模型可用于预测泥石流的流动距离和泛滥范围,以及泛滥范围内的危险房屋和路段,也可以用于泥石流灾害的风险性分析。     

断层防水煤柱可靠度分析

为了合理设计防水煤岩柱的宽度,使用了可靠度分析的“JC”方法,分析了水压力、煤层抗张强度及突水系数的不确定性,建立了防水煤柱可靠度分析的极限状态方程。并计算了与防水煤柱有关的可靠度指标、失效概率和安全系数。参考其它事故的死亡率,防水煤柱的失效概率为1× 10 -5~1× 10 -6比较合适。      

扩展型Boussinesq水波方程的混合求解格式

为高效求解扩展型Boussinesq水波方程,建立了基于有限差分和有限体积方法的混合数值格式。将一维控制方程写为守恒形式,方程中通量部分采用有限体积方法求解,剩余部分采用有限差分方法求解。其中,有限体积方法采用Godunov类高分辨率格式,并结合HLL(Harten-Lax and van Leer)式黎曼问题近似解求界面数值通量,黎曼问题界面左右变量通过高精度状态插值方法(MUSCL)构筑。有限差分方法则采用具有二阶精度的中心差分公式进行。采用具有TVD(Total Variation Diminishing)性质的三阶龙格-库塔多步积分法进行时间积分。对数值模式进行了验证,数值结果同解析解或实验数据吻合良好。     

溃坝水流数值计算的非结构有限体积模型

针对溃坝洪水数值计算面临不规则边界和复杂地形等问题,建立了三角形网格下求解二维浅水方程的高精度Godunov型有限体积模型.空间上,引入变量重构和限制器技术,采用HLLC近似Riemann算子计算数值通量;时间上,采用Hancock预测-校正法.将底高程定义于单元顶点,结合单元水位～体积关系,提高了干湿界面处理能力.采...