CFD Simulation of Avalanches and Other Snow Flows

Winter has arrived in the northern hemisphere and with it comes snow. Snow complicates our lives — roads become slick with ice and drifts can block our way. When it piles up high on mountains, avalanches can wreak havoc.

But what I really want to talk about is how snow illustrates a bigger story about CFD. Back in the day, CFD required much time, attention and resources to get right. It was applied to high value targets — what I like to call racing cars and rocket ships. If you didn’t have a deep experience and “big iron” servers, you may as well forget using simulation. How times have changed! We are enjoying what we call the democratization of CFD simulation. Moore’s Law continues to put more horsepower on every desktop. ANSYS developers have rolled out improvements to model prep, meshing and post processing that puts CFD within the reach of every engineer. As a result, CFD is being applied to important, interesting problems that just don’t enjoy the kind of budget and resources of say, an automotive or aerospace company. Snow flows is one such area.

Snow flows can be extremely difficult to measure in the real world. Avalanches don’t just happen on command. And, how do you measure them when any instrumentation just gets swept away? Even more mundane problems can be daunting. Let’s consider snow fences. Snow fences are primarily used to keep roadways or railways clear by minimizing the amount of snowdrift, especially in rural areas that have less snow plow coverage. How does one go about optimizing a snow fence? You could test them in the field. However, snow is usually limited to just a few months out of the year and ideal locations are typically far from urban areas making logistics difficult. Iterating improvements and optimization could take years of field work. What about wind tunnels?

A research team in Norway complains that wind tunnels have limited testing space, present very high costs, require elaborate equipment and still present difficulties in collecting data to reflect the full structure of the wind field. On the other hand, they report that CFD has advantages including being more flexible, efficient, productive and better agreement with real world results.

CFD results of the snow fence model are in good agreement with experimental results. (Xu and Mustafa, Implementing a 3D CFD Model to Study the Performance of Porous Fences under Harsh Climatic Conditions)

CFD results of the snow fence model are in good agreement with experimental results. (Xu and Mustafa, Implementing a 3D CFD Model to Study the Performance of Porous Fences under Harsh Climatic Conditions)

Teams from around the world are making use of CFD. A team in Spain reported that they successfully solved the complex nonlinear problem of the determination of the efficiency of a snow fence with airfoil snow plates. In New Zealand, Sara Jane Wakes found CFD to be an effective tool in the initial design of snow fences. The Oregon Department of Highways is soliciting proposals to use CFD to optimize the design of living snow fences. There are many more than I can list.

Snow volume fraction contour plots around a snow fence (Wakes, Use of CFD in the initial design of a snow fence)

Snow volume fraction contour plots around a snow fence (Courtesy Wakes – Use of CFD in the initial design of a snow fence)

I’ll leave you with this last look at progress towards modeling avalanches and say be safe out there in the snow!

3 thoughts on “CFD Simulation of Avalanches and Other Snow Flows

  1. What module was the CFD analysis run with, CFX or Fluent? And was it post processed in CFD post? Is there any additional information you can send me on this?

  2. I’m glad you found this work interesting!
    Both of the Snow Fence papers were done in ANSYS Fluent. Unfortunately, the avalanche video does not credit the specific CFD solver. The best way to get more detail is to follow the links to the original papers.

Comments are closed.