Some Examples on Cartesian Grids.

boling2000

boling2000

I really like it

boling2000

how about this one.

icemcfd

这些是用Cart3D计算的例子，Cart3D由NASA开发，ICEM CFD下的一个产品：
http://www.icemcfd.com/cart3d/index.html

周华

***** 版主模式 *****
该贴子是管理员从<a href=forums.cgi?forum=12>网格技术</a>转移过来的！

飘风

为什么都是直角坐标的，没有坐标变换吗？
这样在边界处不是要插值？是否影响精度？

zhifawww

坐标的变换同样会影响精度的.

sillybear

来自97年AIAA的论文。Robust and Efficient Cartesian Mesh Generation for
Component-Based Geometry, 有上图的详细说明。(附PDF)
Abstract
This work documents a new method for rapid and
robust Cartesian mesh generation for component-based
geometry. The new algorithm adopts a novel
strategy which first intersects the components to
extract the wetted surface before proceeding with vol-ume
mesh generation in a second phase. The intersec-tion
scheme is based on a robust geometry engine that
uses adaptive precision arithmetic and which auto-matically
and consistently handles geometric degener-acies
with an algorithmic tie-breaking routine. The
intersection procedure has worse case computational
complexity of O(N logN) and is demonstrated on test
cases with up to 121 overlapping and intersecting com-ponents
including a variety of geometric degeneracies.
The volume mesh generation takes the intersected
surface triangulation as input and generates the mesh
through cell division of an initially uniform coarse
grid. In refining hexagonal cells to resolve the geome-try,
the new approach preserves the ability to direc-tionally
divide cells which are well-aligned with local
geometry. The mesh generation scheme has linear
asymptotic complexity with memory requirements
that total approximately 14 words/cell. The mesh gen-eration
speed is approximately 10 6 cells/minute on a
195Mhz RISC R10000 workstation.
I. Introduction
The past several years have seen a large resur-gence
of interest in adaptive Cartesian mesh algo-rithms
for application to problems involving complex
geometries. Refs. [1-12], among others, have proposed
flow solvers and mesh generation schemes intended
for use with arbitrary geometries. Since generating
suitable Cartesian meshes is relatively quick, and the
process can be fully automated, much of the on-going
research focuses on quick extraction of CFD-ready
geometry from the CAD databases to provide easy
access to accurate solutions of Euler equations within
the design cycle.
Viewing configurations on a component basis has
several conceptual advantages over treatments which
work with a single complete configuration. The most
obvious of these is that components can be translated/
rotated with respect to one another without requiring
user intervention or a time consuming return to CAD
in order to extract new intersection information and a
new CFD-ready description of the wetted surface.
Many approaches begin with a surface triangulation
already constrained to the intersection curves of the
components [9] . By starting upstream in the process,
the component-based approach requires only that each
piece of the geometry be described as a single closed
entity. Thus, relative motion of parts may be pre-pro-grammed
or even computed as a result of a design
analysis. The approach offers obvious advantages for
automation through external, macroscopic control.
This flexibility comes at the expense of added com-plexity
within the grid generation process. Since com-ponents
may overlap, the possibility exists that a
Cartesian cell-surface intersection detected during
mesh generation may be entirely internal to the con-figuration,
and thus all such intersections must be
classified as “exposed” (retain) or “internal” (reject).
Even if the vast majority of such intersections are
actually part of the wetted surface, all intersections
have the possibility of being internal, and therefore
must be tested. An analysis of the mesh generation
procedure documented in Refs.[1] and [13] revealed
that up to 60% of the computation was dedicated to the
resolution of this issue.
Although several approaches toward streamlining
the process exist, the most attractive appears to be one
which avoids the issue of intersection classification
altogether. By first intersecting all components
together, one can extract precisely the wetted surface
so that all subsequent Cartesian cell intersections are
guaranteed to be “exposed” and therefore retained.
The remaining mesh generation problem may then be
treated as if it were a single component problem.
While conceptually straightforward, efficient imple-mentation
of such an intersection algorithm is deli-cate.
Each component is assumed to be described by a
surface triangulation and the solution involves a
sequence of problems in computational geometry. The

wangye03219

谢谢!