|
|
发表于 2011-3-19 09:58:48
|
显示全部楼层
回复 30# wdlxmzd 的帖子
Dear Wdlxmzd,
Your original posting was modified before I have the chance to respond. However, I thought that your candid comments deserve direct response. Thus, I will respond to your original posting in the following.
> 感谢罗教授的回复。
Likewise, thank you for your willingness to engage this discussion or debate, which may benefit others.
> 我并不是BGK的卫道者，只是一个东西存在或者发展都有它的道理。
Whether you are a defender of the lattice BGK (LBGK) model or not is besides the point. However, it is rather disappointing that this discussion has little technical/scientific substance. I must also distinguish the BGK model and the LBGK model -- it is the latter which I am attacking.
> 在气体动理论中，并没有什么MRT，当人们直接从动理论出发构造相应的LBM模型时，BGK模型几乎是必由之路，但这不是说，到这就终止了，由于BGK 的简单，人们更容易发现一些建模的规律，而后可以将这些规律推广，构建相应的MRT模型。事实上，直接从MRT出发，或许难以发现这些建模规律。既然 BGK有这个用途，为何不用？
You are mistaken.
1) The "MRT" model in the gas kinetics is the Gross-Jackson model (1959), which is a generalization of the BGK model (1954), and Gross was the "G" in the BGK trio. The Gross-Jackson model has infinite number of eigenvalues (or relaxation times) and the BGK model is only a special case of it. The MRT-LBE can be directly derived from the Gross-Jackson model -- I gave lectures on this subject several times. The MRT-LBE is just a discretized and truncated model. What makes you to assert that "在气体动理论中，并没有什么MRT"?
I would also add that I have not seen a single LB model with kinetic underpinning which cannot be derived from kinetic equations or other PDEs. This has been my research effort in the past decade or so.
2) Since the LBGK model is only a special case of the corresponding MRT model, I don't see how one can claim that "事实上，直接从MRT出发，或许难以发现这些建模规律。" Please educate us: what has been easily discovered by the LBGK model which is hard to be achieved by the MRT model? My observation and experience are just the opposite, and that is why I advocate to abandon the LBGK model entirely and once for all.
> 我们为什么要自己束缚自己？为何不可以多一条发现问题，分析问题的思路？
Yes, why do we have to confine ourselves (e.g., to the LBGK or any such models)? This is a very good question, and I would like to know the answer, too.
> 也许没有人要打压BGK，但是夸张的渲染或许会使得国内的BGK使用者，研究者有很大的压力，这在LBM之外的其他数值方法领域里是很少看到的。此外，请问POWERFLOW现在是否升级到MRT？请问您在2000或1999年以前是否也从事过BGK？是否所有从事过BGK的人都要以此为耻？
1) "但是夸张的渲染或许会使得国内的BGK使用者，研究者有很大的压力"
Well, I would like to know where and what are these "夸张的渲染". If there are exaggerations, most of them have been made by those who are preaching the LBGK doctrines: the LBGK model can simulate micro-flows in transition regime with large Kn number, fully compressible flows with shocks, the Burnett equation, and much much more. These claims have been made for quite some time. Now, where is the beef?!
2) 请问POWERFLOW现在是否升级到MRT？
Are you saying that just because PowerFlow uses the LBGK model, therefore the LBGK model is superior?
My question: why should we be concerned with PowerFLOW at all?
3) 请问您在2000或1999年以前是否也从事过BGK？
Yes, I did study and use the LBGK model in my thesis (1993) for I did not know better. I only knew the LBGK model then and I did not have anyone who is an expert in kinetic theory to interact with. Worse, I was ill-advised that the LBGK model is a new wheel and kinetic theory is not much of use. It was this experience that forced me to learn kinetic theory from scratch. I only learned the MRT in late '90s, and I immediately realized how ignorant I was (for I did not know much about kinetic theory when I was a student). I learned my lessons hard way and I am determined not to allow my students to repeat this experience. I always advise them to learn some basic kinetic theory first, before they get into the LBE. For the same reason I am engaging this discussion here.
4) "是否所有从事过BGK的人都要以此为耻？"
Well, it depends. for those innocent students who have not been properly advised, they have nothing to be shamed of and I feel sorry for them. I only hope I could help them. However, for those "experts" who simply ignore mounting evidence and advocate certain models/methods for whatever reason, they ought to be shamed!
> SIMPLE系列算法发展近半个世纪，最原始的SIMPLE算法有不少缺陷，其后有很多新算法高级算法提出来，但在FVM领域从来没有见到学者呼吁大家看看这种算法是多么愚蠢。相反，最原始的SIMPLE算法仍然是从事FVM研究或者FLUENT的使用者必须掌握的基本知识。
I know nothing about "SIMPLE", but I assume that "SIMPLE" evolves and no one would insist to use the most primitive version of "SIMPLE" when better ones exist. However, this is not the situation in the field of LBE -- Some do insist the LBGK model has "advantages" without any evidence.
I view popularization of the LBGK model as a pandemic and would like to see it eradicated completely and entirely.
> 我并不赞同您BGK没有优点的说法， 更不赞同去否定BGK的历史贡献。
We can agree that we disagree. To this day I have yet to see ONE single technical advantage of the LBGK model -- Please offer one concrete example if you have one. As for the "historic contributions" of the LBGK model, I never try to deny them. However, my view is that there is ABSOLUTELY nothing positive about the LBGK model, except that it does popularize the LBE method in such a way that it makes the LBE as whole as a laughing stock in the CFD community.
> "同时，MRT也不是万能的。"
Who claims it is?!
Finally, I hope this discussion can be limited within the realm of science -- Let's speak facts, not faith. If you do believe the LBGK model has "优点", please share them with us.
-- LSL |
|
[复制链接]
IP卡
狗仔卡
显身卡
