首页> 外文期刊>Applied mathematics and computation >Numerical methods for nonlinear second-order hyperbolic partial differential equations. I. Time-linearized finite difference methods for 1-D problems
【24h】

Numerical methods for nonlinear second-order hyperbolic partial differential equations. I. Time-linearized finite difference methods for 1-D problems

机译:非线性二阶双曲型偏微分方程的数值方法。 I.一维问题的时间线性化有限差分方法

获取原文
获取原文并翻译 | 示例
           

摘要

A three-parameter family of finite difference methods for one-dimensional hyperbolic equations with damping and non-linearities based on the introduction of a new dependent variable and three implicitness parameters for the time discretization and the diffusion and reaction terms, is presented and shown to result in two-time level semi-discrete equations. Second-order accurate spatial discretizations and time linearization of the nonlinear source terms are shown to result in three-time level linear finite difference equations which are second-order accurate in time whenever the reaction and diffusion terms are allocated in the proportion 1:2:1 to the past, current and future time levels. Second-order accurate in time, time-linearized, three-point compact operator methods for the spatial derivatives which only involve three time levels and three grid points and are fourth-order accurate in space, are also presented, and their linear stability analyzed. An exhaustive comparison between analytical and numerical results indicates that the accuracy of time-linearized methods increases as the time step and grid spacing are decreased, and is nearly independent of the order of the spatial discretization when there is a sufficient number of grid points to resolve the wave structure, and the best accuracy is achieved when the three implicitness parameters are equal to 1/2. It is also shown that the accuracy of time-linearized methods is a strong function of the wave speed; as the wave speed is increased and the wave equation tends to a parabolic one, the methods are still three-time level techniques which may not preserve the positivity of the solution, whereas a Crank-Nicolson discretization of the corresponding parabolic equation does preserve this property provided that there are no mathematical in-compatibilities between the initial and the boundary conditions. It is also shown that, for one-dimensional wave equations with monotonic kinetics and time-dependent forcing, there are very few differences between second- and first-order time-accurate discretizations of the diffusion terms provided that the time step and grid size are sufficiently small, and the discretization of the reaction terms plays a less important rule than that of the diffusion terms. (c) 2007 Elsevier Inc. All rights reserved.
机译:提出了一种新的因变量和三个隐式参数,用于时间离散化,扩散和反应项,并给出了具有阻尼和非线性的一维双曲方程组的三参数有限差分方法。得出两级半离散方程。非线性源项的二阶精确空间离散化和时间线性化显示为可生成三级线性有限差分方程,只要将反应和扩散项按比例1:2分配,它们在时间上都是二阶精确的: 1到过去,当前和将来的时间级别。提出了时间导数的二阶精确度,时间线性化的三点紧凑算子方法,该方法仅涉及三个时间水平和三个网格点,并且在空间上具有四阶精度,并对它们的线性稳定性进行了分析。分析结果和数值结果之间的详尽比较表明,时间线性化方法的精度随时间步长和网格间距的减小而增加,并且在有足够数量的网格点可解析时几乎与空间离散化的顺序无关。波形结构,当三个隐式参数等于1/2时,可以获得最佳精度。还表明,时间线性化方法的准确性是波速的重要函数。随着波速的增加和波动方程趋于抛物线,这些方法仍然是三级技术,可能无法保持解的正性,而相应抛物线方程的Crank-Nicolson离散化却保留了这一性质。前提是初始条件和边界条件之间不存在数学上的不兼容性。还表明,对于具有单调动力学和时变强迫的一维波动方程,假设时间步长和网格尺寸为零,则扩散项的二阶和一阶时间精确离散化之间的差异很小。足够小,并且反应项的离散度不如扩散项重要。 (c)2007 Elsevier Inc.保留所有权利。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号