Software-Computational fluid dynamic/thermodynamic
Software-Computational fluid dynamic/thermodynamic
"Molly" Jan. '84 black/tan 300d 180k miles, two owner car
Silberpfeil
01-17-2011, 01:07 PM #1I need a recommendation for professional level (top of the line-PC based) computational fluid dynamic/thermodynamic software. This would be used to design custom turbo/intake and exhaust systems. Also possibly used for internal engine modification but primarily for intake and exhaust.
"Molly" Jan. '84 black/tan 300d 180k miles, two owner car
(01-17-2011, 01:07 PM)Silberpfeil I need a recommendation for professional level (top of the line-PC based) computational fluid dynamic/thermodynamic software. This would be used to design custom turbo/intake and exhaust systems. Also possibly used for internal engine modification but primarily for intake and exhaust.
I take it your new to this field. Be prepared for a steep learning curve.
expensive option http://www.ansys.com/ (masters degree helpful)
cheap option http://www.openfoam.com/ (phd helpful)
You'll also want some serious computer hardware. (Maybe more than one machine in a cluster)
(01-17-2011, 01:07 PM)Silberpfeil I need a recommendation for professional level (top of the line-PC based) computational fluid dynamic/thermodynamic software. This would be used to design custom turbo/intake and exhaust systems. Also possibly used for internal engine modification but primarily for intake and exhaust.
I take it your new to this field. Be prepared for a steep learning curve.
expensive option http://www.ansys.com/ (masters degree helpful)
cheap option http://www.openfoam.com/ (phd helpful)
You'll also want some serious computer hardware. (Maybe more than one machine in a cluster)
Time to see if my hunch about it taking 3D wireframes/solids is correct...
1990 Power Ram 50 V6 SOHC 24V 6g72
I can be wrong, don't take everything I say as verbatim, please fact-check first.
My posts are my personal opinions and thoughts, unless otherwise noted.
Oooh! OpenFOAM looks awesome!
Time to see if my hunch about it taking 3D wireframes/solids is correct...
1990 Power Ram 50 V6 SOHC 24V 6g72
I can be wrong, don't take everything I say as verbatim, please fact-check first.
My posts are my personal opinions and thoughts, unless otherwise noted.
If you can get your hands on the Premium package it can do almost anything
I wouldn't mind getting an STD..
Spaceraver
01-18-2011, 04:28 AM #4Solidworks does do CFD http://www.solidworks.com/sw/products/cf...ftware.htm
If you can get your hands on the Premium package it can do almost anything
I wouldn't mind getting an STD..
1990 Power Ram 50 V6 SOHC 24V 6g72
I can be wrong, don't take everything I say as verbatim, please fact-check first.
My posts are my personal opinions and thoughts, unless otherwise noted.
I've got solidworks with COSMO at the college, need to bug my brother on how to use sw though... I'm Inventor, he's SW, leads to some interesting dynamics when we try to share ideas, haha.
1990 Power Ram 50 V6 SOHC 24V 6g72
I can be wrong, don't take everything I say as verbatim, please fact-check first.
My posts are my personal opinions and thoughts, unless otherwise noted.
Edit: Also setting up a good mesh before hand is critical to getting good results
The problem with doing CFD simulation is you often have to calibrate the conditions of your simulation to a known test verified in a wind tunnel to make sure what your getting out is sane. Its often far more than setting up a case and just hitting the run button.
Edit: Also setting up a good mesh before hand is critical to getting good results
Robert_UAH
01-18-2011, 04:43 PM #7Open Foam is a serious pain in the a$$. It seems to be mainly for research to me. Solidworks will do internal flow if you get the right package. I am working on getting CFDesign software for my FSAE team.
Most of those solvers require the user to make significant assumptions about the boundary layer properties, the more accurate of which require good knowledge of the gas properties.
But with combustion products, you're going to be lucky if you guess the viscosity within an order of magnitude. Then there's viscosity temperature dependence.. I have no idea how you'd measure this property, but I would expect it to be a significant factor since you're dropping a couple hundred degrees from the turbine exit to the exhaust tip.
There's some literature out there published either by CFD software makers or some university researchers about the accuracy of CFD in predicting friction drag losses in internal flows. I've seen some good papers done by Fluent on properties of combustion products, but they were all gas not diesel. Maybe you could use them as a starting point and then calibrate.. yeah, um.. I'm not really going anywhere with this.
My point is that you'll spend more time determining gas properties and adjusting boundary layer models in an attempt to get a non-absurd solution than you will spend designing the piping. Check my reasoning with a pipe loss hand calc. Start with 10 feet of 3" straight pipe and see what 3 or 4 bends do to head loss and pumping power.
1983 300SD - 350k miles, disassembled for a head R&R
2010 Golf TDI - The wife's car
ConnClark is being too nice to you guys about using CFD in a amateur/semi-pro project.
Most of those solvers require the user to make significant assumptions about the boundary layer properties, the more accurate of which require good knowledge of the gas properties.
But with combustion products, you're going to be lucky if you guess the viscosity within an order of magnitude. Then there's viscosity temperature dependence.. I have no idea how you'd measure this property, but I would expect it to be a significant factor since you're dropping a couple hundred degrees from the turbine exit to the exhaust tip.
There's some literature out there published either by CFD software makers or some university researchers about the accuracy of CFD in predicting friction drag losses in internal flows. I've seen some good papers done by Fluent on properties of combustion products, but they were all gas not diesel. Maybe you could use them as a starting point and then calibrate.. yeah, um.. I'm not really going anywhere with this.
My point is that you'll spend more time determining gas properties and adjusting boundary layer models in an attempt to get a non-absurd solution than you will spend designing the piping. Check my reasoning with a pipe loss hand calc. Start with 10 feet of 3" straight pipe and see what 3 or 4 bends do to head loss and pumping power.
1983 300SD - 350k miles, disassembled for a head R&R
2010 Golf TDI - The wife's car
I'll start messing with more this weekend.
1990 Power Ram 50 V6 SOHC 24V 6g72
I can be wrong, don't take everything I say as verbatim, please fact-check first.
My posts are my personal opinions and thoughts, unless otherwise noted.
1 Guest(s)
1 Guest(s)