Belgian startup AirShaper makes CFD a matter of a few clicks. Engineering degree not required.
If the thought of doing a fluid flow simulation makes you sweat, worry if you are up to task, try to remember what Navier-Stokes equations were all about, you will rejoice at using AirShaper. A tiny team of five in Antwerp, Belgium has managed to create a flow simulation tool so easy to use that ordinary product engineers can use it. You don’t have to be a computational fluid dynamics (CFD) analyst.
The procedure is this simple:
- You upload an STL file to the AirShaper site.
- You orient the part relative to the ground and the wind. It takes a click or two.
- You set the wind speed.
- Tell AirShaper how accurate you want it to be (low, medium, high).
- Enter your credit card.
That’s it! The next day, you get an email that your solution is available. No ten-thousand dollar program to buy and weeks to learn it. No mesh to create. No technical questions that remind you that you should have listened and retained more from your fluids courses.
Currently, only STL files are accepted through the automated interface, though the company can be sent STEP and IGES files by email.
The model position is verified on the screen, and you can rotate it about the axes if it does not come in properly. Fake streamlines function as a visual aid showing the direction of air flow.
If you are starting to whine about having to wait 24 hours for a canned report, note that the report is over 30 pages, with many images of airflow over your model, tables of data and a well-written narrative that ties it all together (you cannot tell it was written in Belgium). How long would it have taken you to produce that?
No Mesh No Fuss
AirShaper makes meshing completely invisible to the user, taking pains to remove any mention of meshing, mesh size, mesh refinement or adaptive meshing from its interface. It’s perfect for users who can come to the simulation with a total lack of simulation background.
You essentially have to know nothing about CFD to use AirShaper.
The ease with which you will have completed a CFD simulation will no doubt have you wondering if you’ve been tricked, if AirShaper is a game or toy, if the results can be believed, or if they are the creative result produced by CG, like the so-called “physics-based” special effects you see in movies. Rest assured. AirShaper is based on the open source CFD code OpenFOAM, which has many professional adherents. The solution type is RANS (Reynolds-averaged Navier-Stokes) Two-Equation Model: Standard k-omega and SST k-omega. It is not child’s play. This is real CFD.
In addition, AirShaper uses its own techniques to make the mesh adaptive, i.e., increase the mesh density in critical areas as the OpenFOAM code does not have that capability.
To protect users from themselves, the program limits the speeds of airflow, allowing a maximum of Mach 0.3, comfortably in the compressible, subsonic range for which Navier Stokes equations are valid.
Test Drive with a Bike
AirShaper takes pains to explain basics, aware that many who are reading the report or even doing the analysis, may not be fluid flow analysts, or even engineers. Strangely enough, in our test case, a bike rider is shown with a Cd (drag coefficient) closer to a bus than what is expected of a road biker. (Image from AirShaper report.)
AirShaper calculates that the cyclist will need about 150 Watts to go 10 m/s (22 mph) which seems reasonable and, thus, serves as a rough check of accuracy of its results. This accuracy was achieved even with the simplest mesh option (Concept).
AirShaper generates a 31-page report full of streamlines, stagnation plots, even noise analysis that will make it look like you know your way around CFD. The report will have enough basic theory that a CEO with no engineering background will understand it. It takes pains to explain what a Cd (coefficient of drag) is in general and how your particular Cd relates to objects anyone can identify (buses and buildings).
AirShaper may have no equal in ease of use among CFD programs. Autodesk Force Effect Flow, which was simpler (you could upload a photo, it was only 2D) has been discontinued along with all its Force Effect products.
To find out how a small startup can succeed in creating such an easy-to-use CFD program, when the main simulation vendors have only managed to keep it complicated, as well as to find out what is behind the software, we tracked AirShaper Founder and CEO Wouter Remmerie to his home in Antwerp. He called us after he had put his other baby to bed.
Wouter’s program was developed from his attempt to create a race car, called the Aquilo, and control it in a most unusual way—with lateral aerodynamic forces. It is customary for race cars to use airflow for vertical forces (downward force from airfoils to increase cornering speed and prevent lift-off), but lateral flow was a first. Wouter received a patent for it. But despite nibble from both Ferrari and Toyota, neither company took the idea to the racetrack. Air passages required for significant lateral forces take up too much of the width of the car, they said, and the Aquilo was abandoned, according to Wouter.
However, the interface to OpenFOAM developed for the Aquilo and the need for an alternate for wind tunnel testing for concept cars, led Wouter to found AirShaper.
We Wish…
With ANSYS Discovery Live able to do simulations in real time, we’re curious why users have to wait a day for results from AirShaper, especially when a coarse mesh is used.
AirShaper’s 31-page report is chock full of pictures. For ease of generating a standard format, they use preset side, top, front and back views of flow results. Certain results, like streamlines, are shown in 2D section views. But flow around 3D objects is never 2D. Even symmetry planes, including the vertical plane that would cut through the length of the bike and rider would be intersected by flow, such as with turbulent flow.
Airshaper promises a 3D model of the results which would allow users to move around the object and zoom in and out. As the data has already been generated, 3D viewing of it incurs no additional cost for AirShaper—unlike changing the flow or the model, which takes processing time and cost. We wish the 3D PDF would be standard, not the special request.
American users may wonder if AirShaper speaks their measurement language. The program seems to have no settings to change the results to Imperial units (lbs, feet). While we can do those conversions, it would be nice if we didn’t have to. Furthermore, buying the AirShaper service is done in Euros, drawing even more attention to the fact that your precious data not only has to leave your walls, but has to leave your country.
How Much
One run costs as little as 50 Euros (coarse mesh) and 2,500 Euros for detailed models. The detailed models use a much finer mesh.
Should you have to change your geometry after seeing the results, you have to pay the same amount again.
With the Detailed Plan, AirShaper does allow a 50% discount on “sweeps,” what it calls changing the orientation of the model relative to the airflow.
The Final Plan allows for some “tweaks” in the design shape and the finest mesh.
Both Detailed and Final Plans include some phone support, perhaps even with Wouter, himself.
Fresh Air
But to criticize AirShaper is to look the proverbial gift horse in the mouth. It’s an injustice to criticize Airshaper’s results when, if it wasn’t for Airshapre, many engineers might not have the means or the will to accomplish a CFD simulation. Airshaper avoids the normal CFD barriers, the training, model clean-up and preparation, meshing and more. This example, admittedly simple, took 10 minutes to submit, more than half of that for the CAD model to be imported and for entering the credit card information.
All told, AirShaper is drop-dead easy to use, the first of its kind for CFD. We applaud AirShaper in bringing a new level of access to a class of product notorious for being held tightly by the most elite practitioners. While every engineering software now promises democratization, AirShaper actually delivers.
