VansAirForceForums  
Home > VansAirForceForums

-POSTING RULES
-Advertise in here!
- Today's Posts | Insert Pics

Keep VAF Going
Donate methods

Point your
camera app here
to donate fast.

  #1  
Old 04-19-2021, 04:20 PM
Pilot8 Pilot8 is offline
 
Join Date: Sep 2018
Location: California
Posts: 78
Default CFD software recommendations

Anybody on the list using computational fluid dynamics software? Looking for info on what are the best packages, interfacing to 3D cad, accuracy, cpu power needed; any insight either work related or growing up the learning curve. Thanks in advance!

Sam
__________________
Sam Ray
RV8 fastback: 400 hours
IO-360-M1B
Hartzell composite prop
Reply With Quote
  #2  
Old 04-19-2021, 05:01 PM
krwalsh krwalsh is offline
 
Join Date: Nov 2009
Location: San Francisco, CA
Posts: 376
Default For What?

This depends on lot of things. Based on your question, I am guessing that you are new to CFD analysis, and therefore ease of use is probably the key requirement. You can do a lot of good with the free license to Solidworks that is available to EAA members. It gets you most of the packages you need, including the fluids module, I believe. But CFD is definitely a garbage in, garbage out endeavor. Start small with something you can verify and go from there.

That said, at my work we would never use Solidworks for CFD analysis. The things we do, both internal and external to our vehicles, are very far beyond what we would trust from Solidworks. For the most part we use Simcenter NASTRAN inside of Siemens NX for most of our analyses. I have also used ANSYS Fluent as well as Star CCM for much more complicated fluid analysis (supersonic flows, droplet size determination in two-fluid atomizers, hypersonic vehicle loads, etc). You would need years to become good with each of these packages to fully understand all of the various ways you can tweak the models to understand what you can fully do with it.

I've used Fluent in the past to model various wheel pant shapes, and within the ability of the models to determine it suggested there was basically a 1kt difference between the two geometries. The issue was that in using different turbulence models I could get the 1kt difference to swing between the two models. So knowing which model to believe is not so simple. It is much better at comparing small geometry changes at the same input conditions to predict pressure distributions or total drag or lift.

Wheel Pant Model
__________________
Kevin R. Walsh
Cozy Mk-IV

Last edited by krwalsh : 04-19-2021 at 05:03 PM. Reason: Wrong URL
Reply With Quote
  #3  
Old 04-19-2021, 05:45 PM
plehrke's Avatar
plehrke plehrke is offline
 
Join Date: Sep 2006
Location: Defiance, MO
Posts: 1,756
Default

Quote:
Originally Posted by krwalsh View Post
This depends on lot of things. .............But CFD is definitely a garbage in, garbage out endeavor. Start small with something you can verify and go from there.
Agree with Kevin. In my day job I have created thousands of pieces of geometry using NX that were run in various CFD tools. We continue to evaluate many of the commercially available CFD software packages, including Star CCM and Fluent. We always go back to our own in-house CFD codes. None of those on market are ever as easy to use As they claim that any aero engineer can run them. They usually require a CFD specialist to get reliable results.

So back to original question, there are many out there but not sure the average joe can get reliable results. I am just as interested in finding one since one of my hobbies for after I retire later this year will require CAD and CFD.
__________________
Philip
RV-6A - flying 970+ hours
Based at 1H0 (Creve Coeur)
Paid dues yearly since 2007

Last edited by plehrke : 04-19-2021 at 05:47 PM.
Reply With Quote
  #4  
Old 04-19-2021, 05:57 PM
Pilot8 Pilot8 is offline
 
Join Date: Sep 2018
Location: California
Posts: 78
Default

Hi Kevin
Thank you for your reply. Ease of use is important, something with a reasonable GUI would work, however, results that make sense, win at the end of the day. I hear you, garbage in, garbage out, and taking deltas to understand experiments. With respect to taking years to learn- my old manager used to say: "It takes two years to tell if someone is good at circuit design, and 10 years to get good at it". I'm used to running simulations to answer fundamental questions, and have managed to accumulate some fluids questions: 1) what is going on during the last 5 feet of a landing- a dynamics statement w/r to pitch changing over time- CG, ground effect, wing/HS lift, 2) How is the air flowing through the bottom half of the cowling, 3) what does the pressure look like over time inside the Lycoming crankcase. What is a good path on the tools side to move it that direction? If it takes less time than building my 8, I'm a happy camper!
__________________
Sam Ray
RV8 fastback: 400 hours
IO-360-M1B
Hartzell composite prop

Last edited by Pilot8 : 04-19-2021 at 06:25 PM.
Reply With Quote
  #5  
Old 04-19-2021, 06:32 PM
David Paule David Paule is offline
 
Join Date: Dec 2009
Location: Boulder, CO
Posts: 4,759
Default

There's the old-school approach, which still works. Plan the systems on your RV-8 so that you can obtain pressures at nearly any point you are interested in, and have a method of recording them. And that could be as simple as having a few cameras taking pictures of manometers.

Nothing wrong with actual flight test data. A well-thought out test might not provide as much information as a more general simulation but it has the benefit of being reliable, repeatable and accurate. Plus it gives data at your specific places of interest.

Also, don't neglect such simple techniques as yarn tufts or an oil film.

Lots of ways to handle this and find answers to your questions.

Dave
Reply With Quote
  #6  
Old 04-19-2021, 06:52 PM
Pilot8 Pilot8 is offline
 
Join Date: Sep 2018
Location: California
Posts: 78
Default

Hi Dave
I hear you on the old school ways- they are the most accurate. I'm presently setting up some pressure probes around the cowl exit- it would be nice to have some additional tools to do some what if experiments. Presently on a fact finding mission to find out how far the software side has progressed. It seems the software should be able to reliably do the first item to simulate a landing?
__________________
Sam Ray
RV8 fastback: 400 hours
IO-360-M1B
Hartzell composite prop
Reply With Quote
  #7  
Old 04-19-2021, 07:13 PM
David Paule David Paule is offline
 
Join Date: Dec 2009
Location: Boulder, CO
Posts: 4,759
Default

I would expect that to be the most difficult to achieve and the hardest to verify.

But I'm not a CFD user.

Dave
Reply With Quote
  #8  
Old 04-19-2021, 08:18 PM
Aluminum Aluminum is offline
 
Join Date: Jul 2017
Location: San Jose, CA
Posts: 316
Default

Sam, you didn't specify budget. If it's just for hobby ($0 + negative time value), OpenFOAM/ParaView/Salome will solve everything that the big boys can but the UI will make you wish you had a different hobby. There is plenty of academic code out there to get you started.

For most any PDE solver you'll want to invest in the fastest DRAM, not necessarily the hottest CPU. If you want to be really serious about it, the IBM Power architecture will generally outperform Intel by 2x-3x at these kinds of workloads, but it is not for the faint of heart.

Finally, no software will get you very far unless you have deep understanding of the math. Oleg Zikanov's "Essential CFD" is pretty good for the basics.
__________________
Dan V
'91 Zodiac 2013-2020 retired
RV-14A wiring
Reply With Quote
  #9  
Old 04-19-2021, 08:37 PM
Pilot8 Pilot8 is offline
 
Join Date: Sep 2018
Location: California
Posts: 78
Default

Dan
Great tip on the book- just ordered Oleg's book on Amazon. The cost willingness depends on the payback- to get industrial performance at the price of a redbird flight simulator is reasonable. Starting small also makes sense- have an idea on how to approach that. Plus people have mentioned software packages, so we are running on all cylinders. Thanks! What got me thinking CFD was mainstream, user friendly- Mike Patey running simulations to optimize Scrappy, his modified carbon cub. From guy's posts here, I'm guessing paid technical consultants are supporting his application.

Sam
__________________
Sam Ray
RV8 fastback: 400 hours
IO-360-M1B
Hartzell composite prop

Last edited by Pilot8 : 04-19-2021 at 08:48 PM.
Reply With Quote
  #10  
Old 04-19-2021, 10:58 PM
Pilot8 Pilot8 is offline
 
Join Date: Sep 2018
Location: California
Posts: 78
Default

OK, I'm onboard! Turbulent flow is hugely difficult; solving it will be a Nobel Prize.
__________________
Sam Ray
RV8 fastback: 400 hours
IO-360-M1B
Hartzell composite prop
Reply With Quote
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -6. The time now is 07:06 AM.


The VAFForums come to you courtesy Delta Romeo, LLC. By viewing and participating in them you agree to build your plane using standardized methods and practices and to fly it safely and in accordance with the laws governing the country you are located in.