使用基于物理信息的卷积神经网络求解Navier–Stokes方程的物理合理且守恒解

基于物理信息的神经网络方法(PINN)是一种使用神经网络有效求解偏微分方程(PDEs)的新兴方法。基于物理信息的卷积神经网络方法(PICNN)是一种由卷积神经网络(CNNs)增强的PINN的变体。由于卷积神经网络的参数共享特性可以有效地学习空间依赖关系,因此PICNN在一系列偏微分方程的求解问题上取得了更好的结果。然而,应用现有的基于PICNN的方法求解Navier–Stokes方程时会产生振荡的预测解,这违背了物理定律和守恒特性。为了解决这一问题,我们提出了一种将PICNN与有限体积法相结合的新方法,以获得Navier–Stokes方程的物理上合理且具有守恒特性的预测解。我们使用有限体积法推导了Navier–Stokes方程的二阶迎风差分格式。然后我们使用所推导的格式来计算偏导数并构造基于物理信息的损失函数。我们对以稳态Navier–Stokes方程作为控制方程的不同场景进行了实验以评估所提出的方法,包括对流传热问题和顶盖驱动流问题等。实验结果表明,我们的方法可以有效地提高PICNN预测解的物理合理性和准确性。

具p-双调和算子的非局部椭圆方程Navier边值问题的广义解

利用变分方法和相应的临界点定理研究一类具有p-双调和算子的非局部椭圆方程Navier边值问题,在非线性项满足超线性条件时,得到了两个非平凡广义解的存在性定理.

分数阶Navier-Stokes方程的格子Boltzmann方法

对于分数阶Navier-Stokes方程的数值求解问题,首先将该方程进行离散化处理,然后构造D1Q3格子Boltzmann模型,并采用Taylor展开和Chapman-Enskog多尺度展开等技术恢复宏观方程,同时推导出该模型平衡态分布函数的表达式。最后根据一维的两个数值算例对方程进行数值模拟以及误差分析,并将得到的数值解与精确解进行比较,从而验证格子Boltzmann方法的准确性与有效性。

THE NAVIER-STOKES EQUATIONS IN STREAM LAYER AND ON STREAM SURFACE AND A DIMENSION SPLIT METHODS

In this paper,we proposal stream surface and stream layer.By using classical tensor calculus,we derive 3-D Navier-Stokes Equations(NSE)in the stream layer under semigeodesic coordinate system,Navier-Stokes equation on the stream surface and 2-D Navier-Stokes equations on a two dimensional manifold. After introducing stream function on the stream surface,a nonlinear initial-boundary value problem satisfies by stream function is obtained,existence and uniqueness of its solution are proven.Based this theory we proposal a new method called"dimension split method"to solve 3D NSE.

A class of fully third-order accurate projection methods for solving the incompressible Navier-Stokes equations

In this paper, a fully third-order accurate projection method for solving the incompressible Navier-Stokes equations is proposed. To construct the scheme, a continuous projection procedure is firstly presented. We then derive a sufficient condition for the continuous projection equations to be temporally third-order accurate approximations of the original Navier-Stokes equations by means of the localtruncation-error-analysis technique. The continuous projection equations are discretized temporally and spatially to third-order accuracy on the staggered grids, resulting in a fully third-order discrete projection scheme. The possibility to design higher-order projection methods is thus demonstrated in the present paper. A heuristic stability analysis is performed on this projection method showing the probability of its being stable. The stability of the present scheme is further verified through numerical tests. The third-order accuracy of the present projection method is validated by several numerical test cases.

求解不规则区域上椭圆方程的一种Cartesian网格方法及其在Navier-Stokes方程中的应用

提出了一种求解不规则边界上有Robin边界条件的椭圆方程的Cartesian网格方法。该椭圆方程经重写后转化为定义在矩形区域上的椭圆界面问题,进而采用水平集浸入界面方法(IIM)对其进行求解。特别地,Robin边界条件采用单边三次插值离散。随后,利用该方法求解定义在不规则区域上的Navier-Stokes程。Navier-Stokes方程的解法器由求解速度方程的虚拟流体方法(GFM)和辅助变量方程的IIM耦合而成。数值测试表明,椭圆方程的解法器能够产生二阶精度的数值解和梯度,而且能够快速收敛,Navier-Stokes方程的解法器产生了二阶精度的速度及一阶精度的压力。圆柱绕流的仿真验证了Navier-Stokes方程解法器的鲁棒性。

TWO-LEVEL MULTISCALE FINITE ELEMENT METHODS FOR THE STEADY NAVIER-STOKES PROBLEM

In this article, on the basis of two-level discretizations and multiscale finite element method, two kinds of finite element algorithms for steady Navier-Stokes problem are presented and discussed. The main technique is first to use a standard finite element discretization on a coarse mesh to approximate low frequencies, then to apply the simple and Newton scheme to linearize discretizations on a fine grid. At this process, multiscale finite element method as a stabilized method deals with the lowest equal-order finite element pairs not satisfying the inf-sup condition. Under the uniqueness condition, error analyses for both algorithms are given. Numerical results are reported to demonstrate the effectiveness of the simple and Newton scheme.

FICTITIOUS DOMAIN METHODS FOR NAVIER-STOKES EQUATIONS BASED ON PENALTY FACTOR ε

Abstract In this article we briefly studied the fictitious domain methods for steady and nonsteady Navier-Stokes equations based on Penalty factorεon the extended domain. The convergence u?→u in H01（Ω）d and L2（Ω）d（d=2,3） is given as well as p?→p in L02（Ω）.

A Two-Grid Technique for the Penalty Method of the Steady Navier-Stokes Equations

A two grid technique for solving the steady incompressible Navier Stokes equations in a penalty method was presented and the convergence of numerical solutions was analyzed. If a coarse size H and a fine size h satisfy H=O(h 13-s )(s=0(n=2);s=12(n=3), where n is a space dimension), this method has the same convergence accuracy as the usual finite element method. But the two grid method can save a lot of computation time for its brief calculation. Moreover, a numerical test was couducted in order to verify the correctness of above theoretical analysis.

The proper orthogonal decomposition method for the Navier-Stokes equations

The proper orthogonal decomposition (POD) method for the instationary Navier-Stokes equations is considered. Several numerical approaches to evaluating the POD eigenfunctions are presented. The POD eigenfunctions are applied as a basis for a Galerkin projection of the instationary Navier-Stokes equations. And a low-dimensional ordinary differential models for fluid flows governed by the instationary Navier-Stokes equations are constructed. The numerical examples show that the method is feasible and efficient for optimal control of fluids.

A DETERMINISTIC VORTEX METHOD FOR SOLVING THE NAVIER-STOKES EQUATIONS

In this paper, a new 2-D vortex method is developed, which treats the vorticity diffusion in a deterministical way. The Laplacian operator, which describes vorticity diffusion, is approximated by a contour integral. The numerical results of two model problems show that this method has a good accuracy. A primary error estimation is given, and the self-adaptive vortex blob and the boundary conditions are discussed.

Quadrature-free spline method for two-dimensional Navier-Stokes equation

In this paper, a quadrature-free scheme of spline method for two-dimensional Navier- Stokes equation is derived, which can dramatically improve the efficiency of spline method for fluid problems proposed by Lai and Wenston（2004）. Additionally, the explicit formulation for boundary condition with up to second order derivatives is presented. The numerical simulations on several benchmark problems show that the scheme is very efficient.

RESIDUAL A POSTERIORI ERROR ESTIMATE TWO-GRID METHODS FOR THE STEADY (NAVIER-STOKES) EQUATION WITH STREAM FUNCTION FORM

Residual based on a posteriori error estimates for conforming finite element solutions of incompressible Navier-Stokes equations with stream function form which were computed with seven recently proposed two-level method were derived. The posteriori error estimates contained additional terms in comparison to the error estimates for the solution obtained by the standard finite element method. The importance of these additional terms in the error estimates was investigated by studying their asymptotic behavior. For optimal scaled meshes, these bounds are not of higher order than of convergence of discrete solution.

Central Discontinuous Galerkin Method for the Navier-Stokes Equations

Central discontinuous Galerkin（CDG）method is used to solve the Navier-Stokes equations for viscous flow in this paper.The CDG method involves two pieces of approximate solutions defined on overlapping meshes.Taking advantages of the redundant representation of the solution on the overlapping meshes,the cell interface of one computational mesh is right inside the staggered mesh,hence approximate Riemann solvers are not needed at cell interfaces.Third order total variation diminishing（TVD）Runge-Kutta（RK）methods are applied in time discretization.Numerical examples for 1D and2 D viscous flow simulations are presented to validate the accuracy and robustness of the CDG method.

Application of the Hierarchical Functions Expansion Method for the Solution of the Two Dimensional Navier-Stokes Equations for Compressible Fluids in High Velocity

This work presents a new application for the Hierarchical Function Expansion Method for the solution of the Navier-Stokes equations for compressible fluids in two dimensions and in high velocity. This method is based on the finite elements method using the Petrov-Galerkin formulation, know as SUPG (Streamline Upwind Petrov-Galerkin), applied with the expansion of the variables into hierarchical functions. To test and validate the numerical method proposed as well as the computational program developed simulations are performed for some cases whose theoretical solutions are known. These cases are the following: continuity test, stability and convergence test, temperature step problem, and several oblique shocks. The objective of the last cases is basically to verify the capture of the shock wave by the method developed. The results obtained in the simulations with the proposed method were good both qualitatively and quantitatively when compared with the theoretical solutions. This allows concluding that the objectives of this work are reached.

NONLINEAR GALERKIN METHOD FOR NAVIER-STOKES EQUATIONS WITH STREAM-VORTICITY FORM

A nonlinear Galerkin finite element method is presented for the two dimensional incom- pressible Navier-Stokes equations with stream-vorticity form.the scheme is based on two finite ele- ment spaces XH and XH for the approximation of the stream and vorticity function ,defined respec- tively on a coarse grid with grid size H and a fine grid with grid size h<<H.We prove that the difference between the new nonlinear Galerkin method and the standard Galerkin method is of the order H2both in stream function and vorticity.

Fully discrete Jacobi-spherical harmonic spectral method for Navier-Stokes equations

A fully discrete Jacobi-spherical harmonic spectral method is provided for the Navier-Stokes equations in a ball. Its stability and convergence are proved. Numerical results show efficiency of this approach. The proposed method is also applicable to other problems in spherical geometry.

The Boundary Layer Equations and a Dimensional Split Method for Navier-Stokes Equations in Exterior Domain of a Spheroid and Ellipsoid

In this paper, the boundary layer equations (abbreviation BLE) for exterior flow around an obstacle are established using semi-geodesic coordinate system (S-coordinate) based on the curved two dimensional surface of the obstacle. BLE are nonlinear partial differential equations on unknown normal viscous stress tensor and pressure on the obstacle and the existence of solution of BLE is proved. In addition a dimensional split method for dimensional three Navier-Stokes equations is established by applying several 2D-3C partial differential equations on two dimensional manifolds to approach 3D Navier-Stokes equations. The examples for the exterior flow around spheroid and ellipsoid are presents here.

TWO-LEVEL METHOD FOR UNSTEADY NAVIER-STOKES EQUATIONS IN STREAM FUNCTION FORM

Two-level finite element approximation to stream function form of unsteady Navier-Stokes equations is studied.This algorithm involves solving one nonlinear system on a coarse grid and one linear problem on a fine grid.Moreover,the scaling between these two grid sizes is super-linear.Approximation,stability and convergence aspects of a fully discrete scheme are analyzed.At last a numrical example is given whose results show that the algorithm proposed in this paper is effcient.

A streamline diffusion nonconforming finite element method for the time-dependent linearized Navier-Stokes equations

A nonconforming finite element method of finite difference streamline diffusion type is proposed to solve the time-dependent linearized Navier-Stokes equations. The backward Euler scheme is used for time discretization. Crouzeix-Raviart nonconforming finite element approximation, namely, nonconforming （P1）2 - P0 element, is used for the velocity and pressure fields with the streamline diffusion technique to cope with usual instabilities caused by the convection and time terms. Stability and error estimates are derived with suitable norms.