Landracing Forum

Tech Information => Aerodynamics => Topic started by: deanroberts on March 01, 2022, 01:53:47 PM

Title: Looking for a decent CFD "wind tunnel" program
Post by: deanroberts on March 01, 2022, 01:53:47 PM
I'm curious:  What CFD programs have you used as a "virtual" wind tunnel?  Would you recommend it to others?  Here's what I know so far:

I've been taking a course in Autodesk's Fusion 360 CAD program at my local community college and the instructor has suggested I look at Ansys.
SimScape is used by numerous colleges and universities for their Formula SAE projects to design as much low speed downforce into their vehicles as possible.
Woody@ddllc uses SolidWorks? Flow Simulation CFD and is clear about why he does on his website.
kustombrad mentions that Sage Cheshire Aerospace is doing CFD work for him, but I can't tell what they use.
simspeed has at least considered using Air Shaper.
Someone (I've lost track of who) suggested using a combination of SimScape and ParaView would be the cat's meow.

What I'm trying to do is probably the same as everyone else:  reduce drag to the extent possible, eliminate lift front and rear, and reduce pitch and yaw sensitivity (again) to the extent possible ALL before building the body.  I don't necessarily need accuracy, but I DO need reliable, repeatable information about the effect changes in body shape have on critical aero characteristics.  (In engine terms I don't need to know absolute HP, I just need to know whether I've made it better or worse.)

I have to say:  I've read all of bratfink's postings and am a bit in awe of his knowledge, experience and resources.  He's clearly at the far end of the LSR spectrum.  Or maybe beyond it!  But I can't begin to afford him.  And it's going to take me a while to read everything woody@ddllc has posted (!), and he is probably more accessible to me financially and time-wise but I doubt that I can afford him either.  Besides, like most LSR folks I'm a DIYer at heart and curiosity and fascination with the LSR puzzle is a big part of my attraction to it.

Do you have anything you think might help?  Thanks!
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: fissionspeed on March 01, 2022, 02:30:47 PM
I used a cloud platform called SimScale. You only get a certain number of core hours for free, but it's plenty imo. Only potential drawback is that free accounts are open, everyone can see what you're doing.
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: WOODY@DDLLC on March 01, 2022, 06:55:37 PM
I get that question and this is my response: The salesman says, "You can go anywhere with 4-wheel drive!", until you get stuck!  :?

30 day free trial here: https://www.symscape.com/

Study these public projects: https://www.simscale.com/projects/#search=F1&viewMode=thumbView&sortBy=project_likes/#search=test&viewMode=thumbView&sortBy=project_likes/&time=null
If you do a public project, folks will chime in to try & help.

https://airshaper.com/videos/how-to-run-an-aerodynamic-simulation/O_DNDxCtI_I
https://airshaper.com/courses
https://airshaper.com/pricing

If that doesn't work out for you - you know where I'm at - but there are three ahead of you - so far!  :-o
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: deanroberts on March 01, 2022, 09:20:36 PM
thanks fissionspeed!  thanks Woody!
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: Bratfink on March 02, 2022, 01:06:21 AM
Hi Dean, thanks for the mad props :cheers:

This is a deep rabbit hole you are peering over the precipice of.

Unfortunately in the world of aero there aren?t many short cuts you can take that yield good results. But making good aerodynamic data available for reasonable money is something a couple of us have been thinking about for a while.

I?ve used many of the CFD codes professionally and I?ve tested at a number of the wind tunnels too. And almost without fail they all produce a different number for the exact same car.

So which one is right? The answer, all of them and none of them.

In order to answer the question of what code/solver/mesh/boundary conditions you should use first we need to know what answer are you looking for?

If you are just trying to measure performance changes to upper body surfaces then a steady state solver with fixed ground might be enough. The mesh dependency really hinges on the local surface geometry at the areas of concern.

If you are looking for safety critical high speed balance. You probably want to consider full transient with rotating wheel and moving ground boundary conditions AND validate it to tunnel data. You wouldn?t use the cheapest shop to weld your safety critical roll cage would you?

Personally I am partial to OpenFoam. It?s free... but the flip side of that is you need to know what you?re doing to avoid poor results.

PM me and we can chat details. Hopefully I can connect you with the right people or software. My suggestion is a blended approach, a couple of big dollar runs for the safety critical knowledge, and cheaper methods for general tuning development.

The last thing I?ll add. No big money solver will get around crap CAD. Your CFD results will only be as good as the initial model you build. More detail (inside and out) always yields more realistic results.
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: deanroberts on March 02, 2022, 12:48:28 PM
Thanks bratfink.  Just what I needed.  And suspected.  But it's a rabbit hole that fascinates me, and has since I started racing many years ago.  It's time for me to scratch that itch.
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: Doc B. on March 02, 2022, 03:47:27 PM
Quote
The last thing I?ll add. No big money solver will get around crap CAD. Your CFD results will only be as good as the initial model you build. More detail (inside and out) always yields more realistic results.

This. I had grand plans to scan my bike with a Revo Pop and/or photogrammetry with a DSLR to do CFD analysis this year. Scanning a 3D model quickly and inexpensively has turned out to be a real can of worms, the results of which are going to require some major CAD rework before I can get to analysis.
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: fissionspeed on March 03, 2022, 12:08:56 PM
Quote

This. I had grand plans to scan my bike with a Revo Pop and/or photogrammetry with a DSLR to do CFD analysis this year. Scanning a 3D model quickly and inexpensively has turned out to be a real can of worms, the results of which are going to require some major CAD rework before I can get to analysis.

I also found this out the hard way. Amazing for knocking out shmedium sized parts that fit in hard to measure areas. Esoteric black magic for defining large freeform surfaces.
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: WOODY@DDLLC on March 03, 2022, 03:05:11 PM
Scanning salesman said, "You can go anywhere with your scanner!" Until you get stuck!  :cry:
Guess that's why the F1 folks have super computers!  lol8
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: Doc B. on March 03, 2022, 06:12:29 PM
Yup. I used the $50 a can matting spray and tried the little dots. The full color model looks maybe decent enough for gaming. But turn off the color and man those smooth surfaces look like lumpy, melted holey crap. I think a component by component redo in CAD is probably the only way to get it good enough for accurate CFD. I did also try a scan with a rider. Really messy.

https://www.instagram.com/p/CZXVL1wJ1U4/ (https://www.instagram.com/p/CZXVL1wJ1U4/)
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: deanroberts on March 04, 2022, 03:24:48 PM
Point well taken about the usefulness of scanners.  I'm hoping I can dodge that particular bullet altogether.  I'm starting with a relatively clean sheet of paper and "only" want to use CFD to inform the shape of the bodywork that'll contain all the parts and pieces.  Time will tell whether I can to that, but so far I'm looking forward to the challenge.
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: Doc B. on March 05, 2022, 11:53:47 AM
Yeah my comments are probably out of context if you are designing something like a streamliner from a clean sheet of paper. The problem is that I run a naked bike with all sorts of odd shaped parts out in the wind that would take a lot of work to model piece by piece in CAD.
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: Bratfink on March 08, 2022, 09:21:17 AM
As you guys seem to have found out, for cfd you typically need sealed volumes to have a successful run. There are programs that will solve single sided surfaces, but no program will solve incomplete patchy scan data.
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: 71GSSDemon on November 28, 2023, 06:23:05 AM
I see there are apps for new smart phones for scanning items into 3d.  Mine is too old but I am looking to try it out.  Then there is a place called Airshaper.  Cfd program to use. I planned on trying this out.  Here is a link.  https://airshaper.com/
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: Rex Schimmer on November 28, 2023, 01:03:51 PM
You are all adventuring into a field that is in a constant change mode. New methods of CAD, new programs for CFD, just trying to keep up with what is available can be a challenge. I am an engineer, I have designed some pretty interesting machines BUT I AM OLD! which means I pretty much did not become adept at CAD, I did all of my designing on a drafting board with a pencil and velum. Which brings me to why my son and I became customers for Woody. Even though my son, Duke, (also and engineer) does CAD and did the the CAD model of our car for Woody, neither of us felt we were qualified to do CFD. The money we spent with Woody was very well spent, we learned so much about the aero of our car and the Woody designed air scoop was worth every cent of what we paid Woody.

deanroberts keep us informed on your journey into CFD and designing your car, and the other people that have commented on this thread, thanks for your valuable input and keep posting as we all can learn from your experiences.

Rex
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: bubruins on November 28, 2023, 07:05:52 PM
I've played with airshaper and there's some information about it on my build thread Midship Runabout 1 - lakester. Basically - it's hard. The scanned models will have many holes in them that have to be repaired. Automation is helping things. Airshaper does seem to give drag and lift numbers within the realm of reason and you can guess if the changes you're making are helping or hurting. There are several restrictions within airshaper that you have to work around. Ex: intake and exhaust gases. It's fun to play with. My goal is to at least get to a couple of quick configurations I can test in a 1:1 wind tunnel in NC by using CFD and have enough of the general shape right that the whole car won't have to be rebuilt.
Title: Re: Looking for a decent CFD "wind tunnel" program
Post by: deanroberts on November 30, 2023, 02:06:12 PM
What a nice surprise to see this thread pop up again. It's good to hear what others are doing.   Me?  I'm in the process of prioritizing 10 shapes to evaluate in the f :cheers:ree version of SimScale.  That may satisfy my curiosities,  but I doubt it. But based on what I learn from those runs I expect I'll have a much better idea where I wat to go next.  I'll give you an update then. In the meantime,  good luck with whatever you're working on.