negative time step, possible to set?

 User Name Remember Me Password
 Register Blogs Members List Search Today's Posts Mark Forums Read

 LinkBack Thread Tools Display Modes
 March 26, 2014, 06:39 negative time step, possible to set? #1 Member   Tom Join Date: Jan 2012 Location: France Posts: 43 Rep Power: 7 hi there, Do you know if it possible to introduce in OF, icoFoam algorithm in particular, a time step (adjusted or not) which is negative? I mean, can I run a simulation in OF that start from t=0 and it goes till t=tstar with tstar<0 with a deltaT<0? Thank you in advance Regards

 March 26, 2014, 12:50 #2 Senior Member     Marco A. Turcios Join Date: Mar 2009 Location: Vancouver, BC, Canada Posts: 735 Rep Power: 21 I don't believe you can do that with the default time classes, as there is only has increment operator: http://foam.sourceforge.net/docs/cpp/a02579.html I'm not sure in what situation these would be useful. If you really wanted it you would have to have your own time counter and decrement it, although this would only mean that your simulation was still evolving forward in time and you've flipped your time axis. That, or you could find a flux capacitor and drive 88 mph...

March 26, 2014, 12:57
#3
Member

Tom
Join Date: Jan 2012
Location: France
Posts: 43
Rep Power: 7
Quote:
 Originally Posted by mturcios777 I don't believe you can do that with the default time classes, as there is only has increment operator: http://foam.sourceforge.net/docs/cpp/a02579.html I'm not sure in what situation these would be useful. If you really wanted it you would have to have your own time counter and decrement it, although this would only mean that your simulation was still evolving forward in time and you've flipped your time axis. That, or you could find a flux capacitor and drive 88 mph...
You got me. This was just the theoretical plot draft for "back to future IV"!

Anyway my question was bad posed. Indeed the equation I want to solve are not NS classic ones but a motification of them with the temporal term which is going to be -ddt(U)...OF icofoam solver comes to be just the skeleton of the algorithm...

How can I create "my own time counter"?

Thank you in advance to support my idea

 March 26, 2014, 13:06 #4 Senior Member     Marco A. Turcios Join Date: Mar 2009 Location: Vancouver, BC, Canada Posts: 735 Rep Power: 21 I don't think you would need a negative time-step to do what you want. Is ddt(U) negative because it absolutely needs to be, or just because its on the RHS of the equation you are solving (a reference to the equations you are solving would be helpful)?

March 26, 2014, 13:15
#5
Member

Tom
Join Date: Jan 2012
Location: France
Posts: 43
Rep Power: 7
Quote:
 Originally Posted by mturcios777 I don't think you would need a negative time-step to do what you want. Is ddt(U) negative because it absolutely needs to be, or just because its on the RHS of the equation you are solving (a reference to the equations you are solving would be helpful)?
No, it is not just a change of hand side. It is the prescription of the equation, indeed I am dealing with the Adjoint Navier Stokes equation wich prescribe this procedure...

March 27, 2014, 05:40
#6
Member

Tom
Join Date: Jan 2012
Location: France
Posts: 43
Rep Power: 7
Quote:
 Originally Posted by mturcios777 I don't think you would need a negative time-step to do what you want. Is ddt(U) negative because it absolutely needs to be, or just because its on the RHS of the equation you are solving (a reference to the equations you are solving would be helpful)?
I was thinking about it...could it work if I would just change the sing in the time schemes discretization that OF does in CrankNicholson (for instance)?

I mean leaving the time flows as positive but introducing at the place of Dt the term -1*Dt in the CrankNicholson schemes...

My question at this point could be. If it is possible, where is that OF does compute the CrankNicholson scheme, where is the exact file where it does the discretization?

 November 6, 2014, 05:36 #7 New Member   Join Date: Apr 2012 Posts: 8 Rep Power: 7 Dear Tom, Dear Foamers, In my point of view, the problem here is to set the right boundary conditions in time for the adjoint NSE. To make sure, that the problem is well posed, one definitely is forced to compute the solution of ANSE backward in time. Tom, did you find a way to realize a computation in openFOAM with negative timeSteps? I tried a lot so far, but haven't find a satisfying answer yet, so even short hints might be very useful ( I hope you're still familiar with the subject posted above...). Kind regards Tom

 Tags negative time step

 Thread Tools Display Modes Linear Mode

 Posting Rules You may not post new threads You may not post replies You may not post attachments You may not edit your posts BB code is On Smilies are On [IMG] code is On HTML code is OffTrackbacks are On Pingbacks are On Refbacks are On Forum Rules

 Similar Threads Thread Thread Starter Forum Replies Last Post mbcx4jc2 OpenFOAM Running, Solving & CFD 12 August 4, 2015 02:20 karasa03 OpenFOAM 7 December 12, 2013 04:41 xiuying OpenFOAM Running, Solving & CFD 8 August 27, 2013 15:33 hsingtzu OpenFOAM Native Meshers: blockMesh 2 March 14, 2012 10:56 bookie56 OpenFOAM Installation 8 August 13, 2011 04:03

All times are GMT -4. The time now is 13:31.

 Contact Us - CFD Online - Privacy Statement - Top