|
[Sponsors] |
How to map a domain changing in length with time |
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
November 26, 2011, 15:37 |
How to map a domain changing in length with time
|
#1 |
New Member
amgotsikmeme
Join Date: Nov 2011
Posts: 6
Rep Power: 14 |
Hi all;
I have a problem with my CFD project and need help. The problem is I have a 1-D domain which is given by L=L_0+delta(sin(wt)) and I try to map it into a domain that has one fixed length by using fixed number of grid points at all times. I need to find a mapping function. Thanks for your concern Last edited by bababanamzdaalma; November 26, 2011 at 16:14. |
|
November 26, 2011, 19:13 |
|
#2 |
New Member
Carlos
Join Date: Nov 2011
Location: The Netherlands
Posts: 17
Rep Power: 14 |
If I understand this correctly, you need a 1D domain of fixed length where only the nodes change position in time? So in your simulation you would get that some cell faces are elongating while others or shortening, yet the sum of lengths is constant in time.You could start with a simpler example, by having an uniform motion of nodes:
for i = 0:N loc_0(i) = L * i/N + L_0 end for t = 0:T_end for i = 1:N-1 loc(i,t) = loc_0(i) + L/N*sin(w*t) end end This would get your nodes to move uniformly to the right and back. If this is not want you meant, please post some additional information!
__________________
"Do not worry about your difficulties in Mathematics. I can assure you mine are still greater." - Albert Einstein |
|
November 26, 2011, 19:32 |
|
#3 |
New Member
amgotsikmeme
Join Date: Nov 2011
Posts: 6
Rep Power: 14 |
the domain can be considered as 1-D engine piston. I have to figure out an algebraic function that maps the current domain in a constant length one by using constant number of grid points.
|
|
November 27, 2011, 05:30 |
|
#4 |
Senior Member
cfdnewbie
Join Date: Mar 2010
Posts: 557
Rep Power: 20 |
maybe i misunderstand you, but that should be easy:
you know your domain size (physically) at each time t, so just map that your unit interval as a function of t .... your domain length is some fct of t: L(t), with a=X(x=0) and b=X(x=L(t)) so your projection is just: [0,1]-> [a,b(t)] with equidistant nodes.... |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Domain Imbalance | HMR | CFX | 5 | October 10, 2016 05:57 |
Full pipe 3D using icoFoam | cyberbrain | OpenFOAM | 4 | March 16, 2011 09:20 |
Convergence moving mesh | lr103476 | OpenFOAM Running, Solving & CFD | 30 | November 19, 2007 14:09 |
DPM: Particle Tracking | Madhukar | FLUENT | 1 | July 24, 2007 03:51 |
time averaged heat transfer in oscillating flow | Matthieu Ubas | Main CFD Forum | 2 | November 5, 1999 14:20 |