I'm looking for a simple internal flow simulation software that I can use to help design better flowing parts. Anyone have any ideas?
Flow simulation software
- Thread starter RickDLance
- Start date
The CFD programs I use regularly are ANSYS and Comsol. Unfortunately, neither are cheap. If you get the flow simulation package with Solidworks it will give you a decent model, but it pretty limited on configurations and doesn't support multiphysics. However, the Solidworks flow simulation might work the best for you because it is easy to use and gives accurate solutions.
A package that I've always wanted to try is GT-SUITE, but I suspect it is priced along the lines of other multiphyiscs programs.
A package that I've always wanted to try is GT-SUITE, but I suspect it is priced along the lines of other multiphyiscs programs.
Make sure whatever you get has programmable cyclic restriction. There is a trick to high flow on car engines. The valves open and shut, spider manifolds have flow redirection, and the exhaust has multiple pulses, pressures and temperatures.
Some things that look good, flow like shit, and other things that look like shit flow well.
Some things that look good, flow like shit, and other things that look like shit flow well.
Example would LS1 engines. Long thin intake ports actually flow less than the same sized oval on paper. In real life, they flow more. Simple log manifolds flow better than shorty headers.
If you know the speed of sound at the temperature and pressure involved, and know the cam timing, you can make bigger gains than actually porting.
In general, GM ain't dumb. Stuff that might look like crap is actually pretty good.
In general, GM ain't dumb. Stuff that might look like crap is actually pretty good.
Star CCM is another good tool. It's relatively easy to learn, but won't be as powerful as a more expensive tool. If you are looking for a tool to model static flow (i.e. steady state), Solidworks will be fine (if you're looking to improve the flow of your manifolds, pedestals, etc.). If you want to model how each exhaust pulse influences the next, you probably want a more sophisticated tool.
If you want a really crazy model (flow as temperature increases while being acted on by each exhaust pulse), you'll need a multiphysics program.
I've found that Solidworks' FEA is very accurate. Their CFD, on the other hand, leaves much to be desired. I tend to treat Solidworks' CFD tools as a quick sanity check before moving on to a more powerful analysis tool.
For my own curiosity, what do you work on that sees 51k feet? The flow simulations do get complicated when you cross from subsonic to mach.
I'm looking to start out very simple.
For example..
I'd like to play with the radius of bends and see the differences.
I'd also like to see what a sharp protruding surface at a joint does to airflow.
I'd like to see how the airflow slows or speeds up as the size of the passage increases and decreases.
Stuff like this. Then I will work my way up.
For example..
I'd like to play with the radius of bends and see the differences.
I'd also like to see what a sharp protruding surface at a joint does to airflow.
I'd like to see how the airflow slows or speeds up as the size of the passage increases and decreases.
Stuff like this. Then I will work my way up.
My bad, I interchange FEA and CFD at work, but yea FEA for stress and normal stuff I have good luck predicting stress concentrations etc. Not to drag this wayy off topic but DO160 has been the go to for companies these days and we have had great luck with vibe analysis in then designing units that don't fail. 20g's break a lot of stuff :rofl:
Techincally speaking I work for zodiac aerospace, but in reality we are a little holding and we do electrical power generation (rotating machines/controllers/rectifiers) for airplanes. [hence the altitude]
Back on topic, you can also do casting modeling with SW, its not the easiest but its doable.
If that is all you are looking to do, the CFD package in SolidWorks will do everything you are looking for. There might be some free tools that will work too, but I haven't tried any.
Other CAD packages also have CFD tools, so you'd probably be better off using your package's built in one of you don't use SolidWorks.
IIRC, the "perfect" curve tries to keep the lateral acceleration constant. Since the air is accelerating linearly when the valve opens, it is picking up speed, which alters the lateral acceleration. Any time you accelerate anything, it consumes power, so the trick is to keep the losses at a minimum.
Picture driving into a corner while your foot is on the gas. If the corner is straight radius, you will scrub hard on entry due to instantly going into max lateral acceleration. Then you will hold the lateral acceleration, then it goes away instantly. You want a cubic curve that brings the acceleration up while the speed is increasing, then for it to relax.
The speed at which the air is accelerating governs the perfect shape. Constant speed airflow will generate a curve that is a different shape.
This is how some heads and exhaust systems show good flow numbers yet pieces with lower bench flow make more HP.
Look closely at a modern turbocharger blade. There are no true radii on either the blade or even the hub. And certainly the snail. They are designed for a certain acceleration. It's why they suck when the airspeed is not in the target range.
Most of my experience is from 30 years ago, as it pertained to 2 stroke racing engines.
At first, it gets tighter (decreasing radius), then it gets looser (increasing radius). Because velocity goes up, and the same radius creates higher lateral losses. It's pretty similar to the line you take on a racetrack for a corner you can accelerate through.
Anything protruding into flow will most likely produce a vortex. The same phenomenon is said to begin after 15 degrees of angle change, and do not forget that your flow can separate from the surface as well.