[求助]有做AMR算法的吗？

jiajialily

有做AMR（adaptive mesh refinement）自适应网格算法研究的吗？我是新手，只是知道要用AMR算法求解时，对于二维的问题要用四叉树，3维的用八叉树，但是进行区域划分的网格上的数值怎么再用数组记录了，希望高手能够指点一下迷津，在此先谢过了！

merryfish

^_^，我也在研究amr，不过最近刚开始，还没有完全看明白。你用的是什么code

试写程序

正在学习,希望和大家交流.

jiajialily

我也是初学者，老板没有做过，让我自己摸索，我看国内做这个的也不多，国外倒是很多，觉得应该有作的潜力，于是想尝试，万事开头难啊！希望和大家多交流，我的常用E-mail是jfwang@ipp.ac.cn，不知道谁有能力，我们可以建一个群

jiajialily

我最近发现一个很好的关于AMR算法的网址
http://flash.uchicago.edu/~tomek/AMR/index.html
里面有很多与其相关的资料，还有一些现成的code，大部分都能连接上的，很多错，各位有时间可以好好看看

merryfish

Here is the link in Aoxue forum, I have attached some of my initial results there. If you are interested, please go and have a look.
http://www.aoxue.org/cgi-bin/ut/topic_show.cgi?id=111904&h=1&bpg=1&age=0

zzuhuyan

merryfish兄,你在傲雪论坛上发表的"大作"我已经拜读过了,幸运的是我跟你做的方向好像很一致,我目前也是做3D-MHD模拟,目前重点是三维太阳风模拟,想基于国外一些开源代码如"paramesh"等构建我们自己的自适应三维太阳风模式,目标是做出我国第一代空间天气预报模式,当然是任重而道远啊!
我的email:yqhu@spaceweather.ac.cn
    QQ:51580266
希望能跟感兴趣的朋友多多交流![br][br][以下内容由 zzuhuyan 在 2007年09月11日 04:20pm 时添加] [br]
ps:merryfish能否告知您用的是什么模式,我看你给的图中有太阳风,以及和磁层的相互作用

merryfish

haha, I am not in China at the moment. The code I have used was developed by other people, but I am expected to incorporate the AMR method into my hybrid code to do the global simulation of magnetosphere, and to improve the current code efficiency by using high-order resolution scheme.

merryfish

Haha, if you give me money, I can do it for you, and much better than the current code.

zzuhuyan

I am very sorry to say that i have no enough money, since i am just a graduate student for my doctor degree now. I just want to discuss this interested  topic with people who has the same interest, of course, if you like.

jiajialily

[这个贴子最后由jiajialily在 2007/09/19 02:53pm 第 1 次编辑]

merryfish, where are you, USA? PPPL? I know some peoples in this institute wh o do the AMR. I want to know what AMR, do you have used? That is patch-based or bisection? I see you statement on the website, maybe, it is patch-based? yes, or not?

zzuhuyan

请问merryfish:
1、你的太阳风模式里，如何处理冕旒衍生出的电流片的重联问题，你在傲雪论坛上发的图片，电流片中的磁力线几乎平行，不知是怎么处理的？
2、你的自适应网格应该是tecplot处理的，请问你的数据格式是什么，按每个block输出吗，另外数据是cell-centered，还是在格点上？
  还望赐教！多谢1[br][br][以下内容由 zzuhuyan 在 2007年09月24日 06:47pm 时添加] [br]
3、你做的问题是两维还是三维？如果是三维的话，能否秀一下三维图片，三维可视化用什么做的？Visit？ or other?
[br][br][以下内容由 zzuhuyan 在 2007年09月24日 06:53pm 时添加] [br]
4、你是否用的Michigan';s model or other?

merryfish

Good news for the AMR community are that we have won a project to develop a parallel amr MHD code using adaptive unstructured mesh. Hopefully, we will finish the code development in three years and will open as a free software under GPL2.0. So everybody can download and use that code in the future.

merryfish

We did use BATSRUS code at the moment, but we also use other AMR MHD code at same time. There are a lot of software to do the postprocessing including tecplot. I only posted the figure produced by Tecplot.
Three-dimensional plots are not easy things to do, but look much better than 2D plots, especially the 3D movie looks very nice.

zzuhuyan

It';s a good news for parallel AMR practitioners! I am waitting here for your progress. I, with my parter, are now trying to develop a parallel AMR model for space weather modeling just like BARTRUS. However, we would like to use a more simple but robust higher order Godunov-type scheme. Being the first step, we use the parallel AMR toolkit PARAMESH to do AMR. We are very happy to know something about your new project. One question: what kind of unstructured grid (tetrahedron/hexahedron) will you use?