使用激波捕捉法求解以数学间断作为初始条件的无黏可压缩流场时,间断结构会逐渐演变为包含多个网格节点的数值过渡区,在此过程中会产生两个平行于间断且规则分布的非物理波动。使用捕捉法计算激波流场时,流场参数应满足修正方程,但作为...使用激波捕捉法求解以数学间断作为初始条件的无黏可压缩流场时,间断结构会逐渐演变为包含多个网格节点的数值过渡区,在此过程中会产生两个平行于间断且规则分布的非物理波动。使用捕捉法计算激波流场时,流场参数应满足修正方程,但作为初始条件进行计算的初始激波满足Euler方程,两者之间的矛盾是产生初始激波诱导非物理波动的原因。鉴于激波等间断是由特征型Euler方程定义的,从该方程出发构造了一种基于特征值迎风特性的通量计算格式(upwind flux scheme based on characteristics,UFSC),并采用若干种常规守恒型通量分裂格式作为参考。数值计算结果表明,UFSC可以消除Steger-Warming、Van Leer等矢通量分裂格式的初始接触间断诱导非物理波动,还可以减小初始激波诱导扰动的幅值。在光滑流场区域,UFSC和矢通量分裂格式的计算结果相近,但是在强激波附近会出现较大的压力尖峰。为克服这一缺陷,进一步构造了在激波区域和其他区域分别采用Steger-Warming格式和UFSC格式计算的混合通量计算格式UFSC+S,可以有效抑制初始激波诱导误差,提高计算精度。展开更多
黏声波方程常被用于描述地下介质的黏弹性及波的传播现象,频域有限差分(finite difference frequency domain,FDFD)方法是黏声波和黏弹性波波场模拟的常用工具.目前FDFD黏声波模拟常用的二阶五点方法和优化九点方法在一个波长内的网格...黏声波方程常被用于描述地下介质的黏弹性及波的传播现象,频域有限差分(finite difference frequency domain,FDFD)方法是黏声波和黏弹性波波场模拟的常用工具.目前FDFD黏声波模拟常用的二阶五点方法和优化九点方法在一个波长内的网格点数小于4时误差较大.通过令FDFD系数随一个波长内的网格点数自适应从而提高FDFD方法的精度,本文针对黏声波波场模拟发展了一种适用于不同空间采样间隔之比的通用格式自适应系数FDFD方法.同时,为了验证自适应系数FDFD方法对一般黏声波模型的有效性,本文针对三个典型的黏声波模型,分别采用解析解和基于高阶FDFD的参考解验证了所提出方法的有效性.本方法的FDFD格式通过在传统的二阶FDFD格式的基础上引入相关校正项得到,其中校正项按网格点与中心点的距离进行分类选取,同时校正项对应的自适应FDFD系数不仅和空间采样间隔之比相关,还和一个波长内的采样点数相关.所需的自适应FDFD系数可通过声波方程的数值频散关系和查找表高效给出.数值频散分析表明,在空间采样间隔相等或不等的情况下,以相速度误差不超过1%为标准,通用格式自适应系数FDFD方法所需的一个波长内的采样点数均小于2.5.数值模拟实验表明,对于不同的空间采样间隔之比,相对于常用的二阶五点FDFD方法和优化九点FDFD方法,通用格式自适应系数FDFD方法均可在相似的计算量和内存需求下,有效提高黏声波模拟的精度.展开更多
Two central schemes of finite difference (FD) up to different accuracy orders of space sampling step Dx (Fourth order and Sixth order respectively) were used to study the 1-D nonlinear P-wave propagation in the nonlin...Two central schemes of finite difference (FD) up to different accuracy orders of space sampling step Dx (Fourth order and Sixth order respectively) were used to study the 1-D nonlinear P-wave propagation in the nonlinear solid media by the numerical method. Distinctly different from the case of numerical modeling of linear elastic wave, there may be several difficulties in the numerical treatment to the nonlinear partial differential equation, such as the steep gradients, shocks and unphysical oscillations. All of them are the great obstacles to the stability and conver-gence of numerical calculation. Fortunately, the comparative study on the modeling of nonlinear wave by the two FD schemes presented in the paper can provide us with an easy method to keep the stability and convergence in the calculation field when the product of the absolute value of nonlinear coefficient and the value of u/x are small enough, namely, the value of bu/x is much smaller than 1. Several results are founded in the numerical study of nonlinear P-wave propagation, such as the waveform aberration, the generation and growth of harmonic wave and the energy redistribution among different frequency components. All of them will be more violent when the initial amplitude A0 is larger or the nonlinearity of medium is stronger. Correspondingly, we have found that the nonlinear P-wave propagation velocity will change with different initial frequency f of source wave or the wave velocity c (equal to the P-wave velocity in the same medium without considering nonlinearity).展开更多
文摘使用激波捕捉法求解以数学间断作为初始条件的无黏可压缩流场时,间断结构会逐渐演变为包含多个网格节点的数值过渡区,在此过程中会产生两个平行于间断且规则分布的非物理波动。使用捕捉法计算激波流场时,流场参数应满足修正方程,但作为初始条件进行计算的初始激波满足Euler方程,两者之间的矛盾是产生初始激波诱导非物理波动的原因。鉴于激波等间断是由特征型Euler方程定义的,从该方程出发构造了一种基于特征值迎风特性的通量计算格式(upwind flux scheme based on characteristics,UFSC),并采用若干种常规守恒型通量分裂格式作为参考。数值计算结果表明,UFSC可以消除Steger-Warming、Van Leer等矢通量分裂格式的初始接触间断诱导非物理波动,还可以减小初始激波诱导扰动的幅值。在光滑流场区域,UFSC和矢通量分裂格式的计算结果相近,但是在强激波附近会出现较大的压力尖峰。为克服这一缺陷,进一步构造了在激波区域和其他区域分别采用Steger-Warming格式和UFSC格式计算的混合通量计算格式UFSC+S,可以有效抑制初始激波诱导误差,提高计算精度。
文摘黏声波方程常被用于描述地下介质的黏弹性及波的传播现象,频域有限差分(finite difference frequency domain,FDFD)方法是黏声波和黏弹性波波场模拟的常用工具.目前FDFD黏声波模拟常用的二阶五点方法和优化九点方法在一个波长内的网格点数小于4时误差较大.通过令FDFD系数随一个波长内的网格点数自适应从而提高FDFD方法的精度,本文针对黏声波波场模拟发展了一种适用于不同空间采样间隔之比的通用格式自适应系数FDFD方法.同时,为了验证自适应系数FDFD方法对一般黏声波模型的有效性,本文针对三个典型的黏声波模型,分别采用解析解和基于高阶FDFD的参考解验证了所提出方法的有效性.本方法的FDFD格式通过在传统的二阶FDFD格式的基础上引入相关校正项得到,其中校正项按网格点与中心点的距离进行分类选取,同时校正项对应的自适应FDFD系数不仅和空间采样间隔之比相关,还和一个波长内的采样点数相关.所需的自适应FDFD系数可通过声波方程的数值频散关系和查找表高效给出.数值频散分析表明,在空间采样间隔相等或不等的情况下,以相速度误差不超过1%为标准,通用格式自适应系数FDFD方法所需的一个波长内的采样点数均小于2.5.数值模拟实验表明,对于不同的空间采样间隔之比,相对于常用的二阶五点FDFD方法和优化九点FDFD方法,通用格式自适应系数FDFD方法均可在相似的计算量和内存需求下,有效提高黏声波模拟的精度.
基金Project of Knowledge Innovation Program from Chinese Academy of Sciences (KZCX2-109).
文摘Two central schemes of finite difference (FD) up to different accuracy orders of space sampling step Dx (Fourth order and Sixth order respectively) were used to study the 1-D nonlinear P-wave propagation in the nonlinear solid media by the numerical method. Distinctly different from the case of numerical modeling of linear elastic wave, there may be several difficulties in the numerical treatment to the nonlinear partial differential equation, such as the steep gradients, shocks and unphysical oscillations. All of them are the great obstacles to the stability and conver-gence of numerical calculation. Fortunately, the comparative study on the modeling of nonlinear wave by the two FD schemes presented in the paper can provide us with an easy method to keep the stability and convergence in the calculation field when the product of the absolute value of nonlinear coefficient and the value of u/x are small enough, namely, the value of bu/x is much smaller than 1. Several results are founded in the numerical study of nonlinear P-wave propagation, such as the waveform aberration, the generation and growth of harmonic wave and the energy redistribution among different frequency components. All of them will be more violent when the initial amplitude A0 is larger or the nonlinearity of medium is stronger. Correspondingly, we have found that the nonlinear P-wave propagation velocity will change with different initial frequency f of source wave or the wave velocity c (equal to the P-wave velocity in the same medium without considering nonlinearity).
