节点余量下降法求解稳态流动边界值问题
崔玉红 , 王善举 , 董泰郎 , 程冠华 , 尹钊
天津大学学报(自然科学与工程技术版) ›› 2026, Vol. 59 ›› Issue (11) : 1138 -1152.
节点余量下降法求解稳态流动边界值问题
Node's Residual Descent Method for Solving Steady-State Flow Boundary Value Problems
节点余量下降法(NRDM)是在广义有限差分法基础上发展起来的一种数值方法,目前已应用于求解线性弹性问题、超弹性、热弹性和电弹性等固体力学问题.本文将其拓展求解以强非线性为特征的稳态流动边界值问题.非线性流动边界值问题的核心挑战在于强对流效应带来的数值稳定性问题,为克服传统显式迭代格式和同位节点方案遭遇的稳定性难题,研究提出了一种基于压力修正框架的改进节点余量下降算法.通过在节点余量下降法的迭代框架中引入压力修正思想与 Rhie-Chow 插值技术,成功解决了压力-速度耦合与数值振荡难题.研究通过一系列代表性算例验证了节点余量下降法求解稳态流动边界值问题的计算精度和稳定性.结果表明,节点余量下降法具有较好的计算精度与稳定性,其数值解与解析解及经典基准数据高度一致,并且在较高雷诺数下的强非线性算例中依然能够稳定收敛得到精确数值解.研究证明,改进的节点余量下降法是一种兼具精确性与稳定性的数值方法,其既无需构建和计算大型全局矩阵,也无需划分网格,不需要插值函数,不需要积分;同时精确求解了不可压稳态流动边界值问题这一非线性偏微分方程组,进一步拓展并验证了节点余量下降法作为一种通用计算框架的能力.研究为未来继续拓展求解诸如物质扩散与流固耦合等同样遵循基本守恒规律并最终可由偏微分方程描述的复杂物理场问题,提供了一条可靠的求解途径.
Node's residual descent method(NRDM)is a numerical method developed based on the generalized finite difference method. Previous research successfully applied NRDM to solid mechanics problems,including linear elasticity,hyperelasticity,thermoelasticity,and electroelasticity. This research extended NRDM to solve nonlinear problems in fluid mechanics,specifically incompressible steady-state flow boundary value problems. The core challenge in solving these nonlinear Navier-Stokes equations lies in the numerical instability induced by strong convective effects. Traditional explicit iterative schemes and collocated node schemes are often prone to stability problems. To address these challenges,this research proposed an improved NRDM algorithm that incorporates a pressure-correction framework. The pressure-correction concept and the Rhie-Chow interpolation technique were introduced into the NRDM iterative framework to successfully solve pressure-velocity coupling and numerical oscillation problems. This research systematically verified the accuracy and stability of the proposed method through a series of representative examples. Results demonstrate that NRDM achieves high accuracy and numerical stability,with solutions closely matching analytical solutions and classic benchmark data. The method exhibits stable convergence to accurate solutions even for nonlinear problems at high Reynolds numbers. Owing to its simple and direct formulation,the improved NRDM avoids large global matrices,mesh partitioning,interpolation functions,or numerical integration. It effectively solves nonlinear partial differential equations governing incompressible steady-state flows. These findings confirm NRDM as a robust and general computational framework,providing a reliable basis for future extensions to complex multiphysics problems,such as mass transfer and fluid-structure interaction,governed by similar fundamental conservation laws and described by partial differential equations.
| [1] |
|
| [2] |
|
| [3] |
|
| [4] |
|
| [5] |
|
| [6] |
|
| [7] |
|
| [8] |
|
| [9] |
|
| [10] |
|
| [11] |
|
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
|
| [16] |
|
| [17] |
|
| [18] |
|
| [19] |
|
| [20] |
|
| [21] |
|
| [22] |
|
| [23] |
|
| [24] |
|
国家自然科学基金资助项目(12672450)
/
| 〈 |
|
〉 |