Sand Dune Formation Due to an Object

Hi there,

I am currently looking for a way to simulate how an object introduced to a sandy plane could create new sand dunes behind it due to its interference with a constant wind vector. The principle is used on beaches to create embankments as shown in the following diagram, where the fence slows the wind leading to a deposition of sand and the creation of a dune:

What I would like to do is simulate possible dune formations behind not just a linear fence but a variety of solid forms. This simulation would of course need to be heavily simplified. I imagine this could be done using vector fields or CFD, similar to that seen in one of Generative Landscape's wonderful tutorials:

The difference being that I would like to show possible areas of sand accumulation. My end game is to be able to design several obstructing masses that will produce different dune patterns behind it given a constant supply of sand and wind, similar to how this bag is generating a particular dune formation behind it:

If anyone is able to point me in the right direction I would be eternally grateful!

- Adam

Replies to This Discussion

Permalink Reply by David Rutten on September 7, 2016 at 5:52am

I have no genuine advice for you as I've never done anything like this (I always chickened out before actually starting to code anything up). However I think my data-structure would be a bitmap, or perhaps a grid of floating point values if you need more z-details. You'd need a second bitmap which represents wind strength (does the wind direction also change due to obstructions? if so, it'll get a lot more complicated), then you start iterating small time-steps until you reach some sort of equilibrium.

However my immediate misgivings:

You have no idea whether your simulation is in any way correct.
There's no guarantee that you'll reach an equilibrium.
You'll have problems with sand blowing away from the leading edge of the simulation area and accumulating at the trailing end.
You'll have problems maintaining sand volume over time.

Permalink Reply by Adam Roggero on September 8, 2016 at 5:43pm

Hi David,

Thanks for your thoughts. Just an update for anyone who might still be interested in the topic. I have decided to try and tap into the source code of a dune simulation similar to this one:

There is also one that utilises Python scripting that might be more easily applied. My hope is to introduce NURBS obstructions into the heat map from which I can project into a 3D surface like this:

I will update how this goes over the coming days. It can be done, as it seems this is the technique used in this project (woohoo!)

Permalink Reply by mark zirinsky on September 7, 2016 at 7:12am

Also for snow fences.

An approach come to mind: a) in the cross section view above, the deposition of the sand dune appears to be a wavelength relationship to the height of the fence, and b) in the plan view above, the flow of the sand dunes looks very much like an equipotential map, or isoflow lines. As for various shapes, you can probably simplify, using a cross section area normal to wind direction, and get past full aerodynamic modeling

Some of the basic work in this area approaches this from a geological modeling point of view, as in sand deposition, so you may also want to look at some of the Colorado School of Mines library material, sorry I do not have a specific reference for you.

Let us know how it turns out.

PS I just recalled that the mechanism of dune formation is a phenomena called saltation; the mechanism involves bouncing particles within about 1 foot of the surface, at least for sand, might be three feet for snow.

Permalink Reply by Adam Roggero on September 8, 2016 at 5:46pm

Thanks Mark,

I'm now trying to tap into a dune simulator that can replicate the realistic movement of sand and dunes over a plane. Hopefully I'll end up with something that can be played around with in 3D. I'll let you know how it goes!

Permalink Reply by Kim hauer on September 8, 2016 at 7:18pm