I would be interested in hearing what others are using for post-processing software and their degree of satisfaction with it.

We have switched to Matlab, and I have a number of scripts for stripping data from phi-type output files for use in Matlab. It is not ideal, but it works.

What are others using? How expensive is it? ease of use?

Thanks

(1). For general post-processing programs, I have used two of them, one is the tecplot, and the other is fieldview. (2). Both are highly professional. (3). They are relatively affordable. (relative to the commercial cfd solver) You can try to visit their home webpage to get more information. There are also several other codes available. All you need to do is to take a look at the sponsors of this forum first.(or search the old messages)

Thank you- I am familiar with the variety of general purpose programs. I am asking PHOENICS users if there was one product that was especially useful in the post-processing of PHOENICS output files. Writing routines to strip the data from the output files is tiresome, and usually won't work after every upgrade sadly.

Of additional interest is software that will do things like calculate countour line-lengths which is rather tricky but useful for chaotic statistics.

Any other stories out there?

I've been using Phoenics in combination with Iris Explorer from Numerical Algorithms Group (www.nag.com) for several years now. Years ago it was bundled with SGI machines, but now its a commercial product.

I've been quite satisfied with it. Explorer has a user interface based on wiring together modules. It comes with a module for reading PHI files. This could be connected to a module to generate an isosurface, which could then connect to a module to render it on screen, for example. Their website has examples and screen shots, I'm sure.

And, like Phoenics, you're free to write your own modules in C or Fortran to extend it.

Email me if you'd like to know more details.

Is Iris Explorer some flavor of the IBM product Data Explorer which is now freeware? The module format sounds like Data Explorer as well. That product can be found at:

http://www.opendx.org

Thanks for the tip.

No, it is entirely separate, just a similar name.

I've been following the OpenDX project, but haven't had the time to try it out, have you?

No, I haven't yet tried to use the OpenDX tools but others in my department have, but not with PHOENICS. The install on Linux is apparently pretty easy. I don't know anyone who has doen an NT or Windows install.

But I'm pretty much all Linux anyway.

I do have MATLAB scripts for stripping data out of phi files if anyone is interested.

For many years we used FAST on SGI workstations. Now we are exploring building our own graphics capabilities using VTK. Both are free. For me currently the main issues are (a) converting the cell centred/staggered/colocation data sets to nodally averaged sets in such a way that conjugate or multiple values of properties (e.g from the PROPS files), and quantities like pressures which may not exist in solid regions are properly addressed. (b) the interpolation algorithms are seldom conservative, so errors result e.g.in streamlines "disappearing" and meaningless critical point (eigenvalue) analyses result. (c) Ease in converting objects generated in CAD programs and subsequently read by the VR viewer into an appropriate form to be read by the visualisation packages. (d) Calculation of non-standard derived properties such as entropy or Gibbs free energy in the light of (a) above. The FAST calculator was a step in the right direction in this regard. There's clearly a need for standards of commonality for data structures between different software packages.