You will probably have seen wildwater kayaking on television. More and more this sport tends to fast flowing, straight forward, man made channels. I would like to model some sections of these channels in 2D. There is an enormous amount of software available, but most seems to be unsuitable for this task. Can someone help me which software to choose? (preferably freeware or public domain for even a national sports organisation is almost budgetless; I'm an unpaid volunteer)

charateristics are: free surface, shallow water, incompressible, discontinuities in bottom profile, steady flow, super- and sub critical flow, plenty shock waves.

thanks, Rob

Dutch kayak Union

(1). Why not build a small scale model in the backyard. (2). Or you can try to get a copy of Dr. C.W. Hirt report at Los Alamos National Lab. The program listing in the report is for 2-D problem. (I don't have the report number, but you can check with the library, something like program SOLA.) (3). There is a finite element free program somewhere in Europe which can deal with free surface. You can search through this forum or the resources section. (4). Take a look at http://www.flow3d.com/ of Flow Science.Inc

Actually John, there are three reports!

sola-vof is a 2-d version.

There are also reports for nasa-vof/2d and nasa-vof/3d.

The last is

Torrey, Martin D., Raymond C. Mjolsness, and Leland R. Stein, "NASA-VOF3D, A Three-Dimensional Computer Program for Incompressible Flows with Free Surfaces," LANL report LA1109-MS, July, 1987.

The others are (NASA-VOF2D) LA10612-MS, December, 1985 and (SOLA-VOF) LA-8355, August, 1980.

The NASA versions add the FAVOR option. That's Fractional Area Volume Obstacle Representation. It's a neat way to represent random region shapes and boundaries on a structured Cartesian mesh.

You may want to take a look at:

water.usgs.gov/software/surface_water

Modeling system: FESWMS-2DH (free)

You may also want to look at www.rmanet.com for the RMA2 model. (cost ?)

Most codes are restricted to subcritical flow.

Regards, Rob

Dear Rob,

As somebody who is working in the field of shallow-water flow and has some years of experience with the 2D RMA2 Software I need to discourage you.

A 2D-calculation does not help you. Th assumption that leads from 3D to 2D-open-channel is the hydrostatic pressure assumption. For which you need to assume that your bottom and water surface are fairly flat. The discontinuities in your bottom profile clearly violate this assumption. I investigated the situation of a backward facing step, and found out, that the prediction of the frictional losses is considerably wrong when calculated with a 2D code even if you are able to calibrate with measurements from one discharge.

The opinion of Rob Del Core, that the the code is restricted to subcritical flow, is not true as far as i know. The reason for the frequently observed divergence of 2D-codes when supercritical flow occures, is, as far as i have tested the problem, that the mesh is is not fine enough to resolve the hydraulic jump.

So from the hydraulic point of view your problem is a very hard nut to crack. You need a 3D non-hydrostatic open surface code with a self adapting mesh, to get realistic predictions.

Scale model test will give better results than 2D-computations in your case.

I would try to find a goodwilling sportsmen who is working in good hydraulic laboratory in your place to get some advice and support. You know sombody at WL-Delft?

good luck Jens Wyrwa

Jens,

Re: your comment about hydraulic jumps.

I disagree with your opinion. A finer mesh will not resolve the jump as you suggest. In a 2D code, the hydraulic jumps need special treatment and would most likely be properly handled as an internal boundary condition. (ie. local treatment of vertical accelerations, etc.)

You say that a 3D non-hydrostatic code with adaptive meshing is required. As far as I know, RMA2 is non adaptive, and hydrostatic.

How then does RMA2 treat supercritical flows?

If you think mesh refinement is the answer, then are you always going to know where to refine your mesh?

Rob

Dear Rob Del Core,

We do not disagree in the point that the large vertical accelerations in a hydraulic jump violate the hydrostatic assumption, which is the basis of 2D- verically averaged calculations.

I never said, that RMA2 is something else than 2D-hydrostatic on a fixed mesh.

I mentioned my tests with RMA2, where I got a supercritical flow with a water-level-rise that looks like a hydraulic jump, to point that the calculation does not diverge automatically when Fr>1 and attributed the divergence in those cases to a different phenomena. I did not say that these are realistic predictions, I wanted to warn, that a 2D-code might give convergent solutions for supercritical flows.

I wrote about self adapting meshes. The aquivilant problem in aerodynamics are shocks. There are some proposals how a code can detect and automatically refine the mesh around the shock.

Jens

Thanks gentlemen, very usefull input!

The reports on SOLA-VOF, NASA-VOF2D/3D are available online at the Los Alamos library. http://www-lib.lanl.gov/pubs/la-pubs.htm Go for the online search menu. There is even a predecessor to the VOF programs called Ripple (dated end 1999).

Still I'm curious about the use and results of the VOF programs. The reports are about the internal mathematics and lack a reference to (hydraulic) test results. Even after an extensive library and internet search, I could't find a reference to a case study with bores or hydraulic jumps. Does someone have a good reference or known good results?

Rob