首页> 外文学位 >Interpolation and error control schemes for algebraic differential equations using continuous implicit Runge-Kutta methods.
【24h】

Interpolation and error control schemes for algebraic differential equations using continuous implicit Runge-Kutta methods.

机译:使用连续隐式Runge-Kutta方法的代数微分方程插值和误差控制方案。

获取原文
获取原文并翻译 | 示例

摘要

It is well known in the literature that standard implicit Runge-Kutta methods applied to algebraic differential equations exhibit order reduction in both differential and algebraic components. As a result, interpolation and error control schemes are also affected, especially for higher index problems. This thesis presents alternative interpolation and error controls schemes based on the defects and the algebraic residuals of a corresponding perturbed equation for particular classes of semi explicit index 1, index 2 and index 3 problems.;Due to its explicit nature, the proposed scheme can be implemented efficiently and at the same time yields results comparable with those from the literature. At each level of this "boostrapping" procedure, the differential interpolants and algebraic interpolants can be constructed separately using explicit evaluations of the interpolants constructed from the previous level.;As in the case with any continuous approximation, the perturbed equation corresponding to these interpolation schemes can be converted to an equivalent underlying initial value problem which is independent of the algebraic component. This allows one to apply established results in the literature to a defect-based error control scheme for the differential component. The same idea can be applied to derive an error control scheme for the algebraic component.;Based on the above ideas, we develop the three schemes--namely, Scheme 1, Scheme 2 and Scheme 3--corresponding to semi explicit algebraic differential equations of index 1, index 2 and index 3 respectively. For each of these schemes, we also derive a corresponding error control scheme based on the defect and algebraic residual of an underlying perturbed initial value problem.;The test problems we use to illustrate our approach are the classical pendulum problem, the seven body problem and the driven cavity problem. While the driven cavity problem is of index 2, the other problems are of index 3 or index 3 reformulated as index 2 problems. These problems arise from two important classes of application: constrained mechanical systems and fluid dynamics.;The three implicit continuous Runge-Kutta formulas used as examples in our investigation are SDIRK(2), Gauss(2) and Gauss(3).
机译:在文献中众所周知,适用于代数微分方程的标准隐式Runge-Kutta方法在微分和代数成分中均表现出阶次减少。结果,插值和错误控制方案也受到影响,尤其是对于高索引问题。本文针对半显式指数1,指数2和指数3问题的特定类别,基于相应扰动方程的缺陷和代数残差提出了替代插值和误差控制方案。高效地实施,同时产生的结果与文献中的结果相当。在此“ boostrapping”过程的每个级别上,可以使用从上一级别构造的插值的显式评估分别构造微分插值和代数插值。可以转换为与代数成分无关的等价的基础初始值问题。这使人们可以将文献中已建立的结果应用于基于差分的差分组件的错误控制方案。可以将相同的思想应用于代数分量的误差控制方案。基于上述思想,我们开发了三种方案-方案1,方案2和方案3-对应于半显式代数微分方程。分别位于索引1,索引2和索引3中。对于这些方案中的每一个,我们还基于潜在的扰动初值问题的缺陷和代数残差推导相应的错误控制方案。我们用来说明方法的测试问题是经典的摆问题,七体问题和驱动腔问题。从动腔问题的指数为2,其他问题的指数为3或重新定义为指数2的指数3。这些问题来自两个重要的应用类别:受约束的机械系统和流体动力学。;在我们的研究中用作示例的三个隐式连续Runge-Kutta公式是SDIRK(2),Gauss(2)和Gauss(3)。

著录项

  • 作者

    Nguyen, Pham Hung.;

  • 作者单位

    University of Toronto (Canada).;

  • 授予单位 University of Toronto (Canada).;
  • 学科 Computer Science.
  • 学位 Ph.D.
  • 年度 1995
  • 页码 111 p.
  • 总页数 111
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

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