CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Help! how to give initial VOF (http://www.cfd-online.com/Forums/main/1520-help-how-give-initial-vof.html)

Ray November 11, 1999 00:03

Help! how to give initial VOF
 
Dear friends

I'm simulating fluid flow with free surface using Fluent v.5

(1) It is difficult to give the initial volume of fluid (VOF) using PATCH command since the shape of free surface is complex

(2) I tried to write user defined function. C_T(c,t) is used to define/change the temperature distribution. But how can we give the initial values of VOF? Is it C_F(c,t)?

(3) If we first write a Profile file (with x,y,z,VOF in turn), we do not know how the x,z,z,vof are put into the profile file.

Could anyone help me? Your help will be greatly appreciated.

Ray

John Chien November 11, 1999 03:24

Re: Help! how to give initial VOF
 
(1). Your experience about the PATCH command should be directed to the vendor of the code. Tell them about your difficulties, and wait for the user-friendly version next time. (it is possible that they already got the answer, you just have to talk to them) (2).You are trying to define something without knowing how to do it. You can't make it to work by trial-and-error. (you don't have to take my words. it is all right to keep trying.) I would say that somewhere along the line, you need to get a copy of the users guide, or a smaple case. (3). To specify the conditions, you need to know two things, one is the format of the input file to the code, the other is the condition itself. You will have to know your conditions before you can define your problem. (4). By the way, a patch is normally referred to a piece of surface in CAD geometry.

T.S.Dhanasekaran November 11, 1999 04:32

Re:Help velocity profile bc
 
How to give boundary profile at inlet of turbulent jet

John C. Chien November 11, 1999 10:13

Re:Help velocity profile bc
 
(1). It is a good question. But I think, it has been discussed way back many months ago. (2). I hope you are not looking for the code implementation of the user defined condition. So, you are not going to get answers in that area. I myself have not even touch that advanced topic. (3). My approach is: if it is not in one of the sample case, don't try it. In this high tech research field, one always get emotionally depressed if one does not get the answer. (4). So much for the introduction. I guess, historically, at the begining, the jet velocity was routinely represented by one or a few cells (or nodes). In this case, one can use uniform velocity profile. To make the profile compatible with the wall, one can modify the wall velocity or next-to-wall velocity.(this is good for condition with more than one cell case) This is fine in the days when the computer memory is counted in how many K bytes. (5). Later on, when the computer memory was counted in terms of how many Mega bytes, it became possible to specify the profile. I think, most of the available codes were developed in this era. (6). On the surface, it is user-friendly options. But in reality, it is hard for the user to come up with a good profile or any profile at all, unless one happens to be working in a test laboratory where test data can be obtained easily. So, we can call this era a "pseudo user-friendly era". That is, an added option without positive answer. (7). One has to remember that, a jet is a boundary layer type, and the jet development has to follow the boundary layer theory, meaning that the downstream development depends on the upstream solution. That's why marching methods are normally used to solve parabolic boundary layer equations. (8). So, in this "pseudo user-friendly era", one can use either the test data, the user-defined semi-analytical profile, or any profile he think is suitable. (9). As you have already expected, at the meeting, everyone will have different answers because their initial jet profiles are different. This is fine because no one in the scientific community would like to be the clone of another person. (10). Now we are outside that pseudo era, the current approach is to include the upstream region of the jet in the computational domain. This has two advantages: one is that he no longer has to worry about the jet exit condition, it is now a part of the solution. There is no need to specify the condition at the jet exit. This is the true user-friendly era. The other advantage is that the upstream condition usually is much simpler to specify. (11). In this "true user-friendly era", we let the computer to do more work, and the inlet is moved further upstream of the jet exit, and the outlet is moved further downstream. That's progress. When a hard disk capacity grows, the application program size also grow. (just remember that growth is the right to everybody. To prohibit the growth by others can become a legal problem.) (12). My suggestion is: in the true user-friendly era, the jet boundary condition should be specified upstream of the jet eixt. In this case, a simple velocity or total pressure condition can be used. The added advantage is that, the jet exit solution sometimes has very important effect on the jet development. (13). Well, now you have three approaches to take, the "pre-user-friendly era approach", the "pseudo-user-friendly era approach", and the "modern-user-friendly era approach". The choice is yours. (In case of hardware limitation, the up-stream region can be solved separately. But the computed profile should be used in the region still away from the jet exit for the main problem.) The answer may not be the one you are waiting for, but this is Internet era. You can't get everything for nothing.


All times are GMT -4. The time now is 00:51.