High Speed Compressor design CFD

[mihir1310]

Openfoam.  But it requires a Linux platform .

[vochitoken]

is COMSOL any good for this types of simulations?

[camelon]

i was watched that gus presentation one of SpaceClaim webinar but bro your sound so weird soo sory but this is real...

[vangro1]

Cohiba,

The compressor wheel is spinning in the proper direction. See Peng's "Fundamentals of Turbomachiner".

[CohibaSkeeza]

Dude, this desing has two major problems:

1. Your radial turbine is backwards dumping the fluid in the counter-clockwise direction and then you want it to travel in a clock-wise direction - the air is being pushed in the wrong direction

2.The fluid is turning that huge corner and you have separation. The fast moving fluid is due to all the back-pressure rather than steady flow.

Your stator veins work because they redirect the air clockwise. Section-off your volute and flip your turbine.

[CohibaSkeeza]

If you want to leave your compressor you need to take the corner of the volute to eliminate separation.

[sjh7132]

If you assume the intake is at the top then it would be spinning clockwise.

[whotookmynickkokokai]

Good observation. How will it be a better design? Do you think the design need to have a smaller housing(radial boundary) near the turbine?

[vangro1]

@CohibaSkeeza

The compressor wheel is spinning in the CORRECT direction in this simulation. The compressor direction should match the direction of the volute to conserve momentum.

As whotookmynickkokoal mentions, the key to improving this design is to reduce the housing size, especially to force the fluid to travel in the volute's direction.

See Peng's "Fundamentals of Turbomachinery" for more info.

[PNCA1984]

for a compressor or pump simulation is not necesary ALE formulation??

[DanFrederiksen]

what's the software's name????

ahh what do I even ask, it's obvious the software will be massively overpriced.

probably have to implement my own CFD at some point because noone do it right

[nofuckthat]

- Maybe you should work on your English first.

[DanFrederiksen]

actually my english is rather good. just that I apply my mind so much that the speech part is often neglected. it's a common phenomenon in these internet times of volume information. you should try thinking sometime. it's good for you. maybe a less douchy name while you're at it..

[erjohnson227]

how is your english bad??? other than using the word noone as one word, i cant find anything wrong with that sentence. what is with that douche?

[DanFrederiksen]

it should say 'why do I even ask' and 'no one does it right'. you might say no one do it right but it's a little relaxed grammar.

[nofuckthat]

Yeah - I was just being an asshole I'll admit. Looking back at the comment I feel ridiculous - I apologize. I'd probably had a little too much to drink, though I don't blame the alcohol.

More to the subject, after trying CFdesign personally I really prefer ANSYS product solutions such as Fluent 6.2 for CFD analysis, though the learning curve is extreme for someone without a degree in the study.

The Fluent plugin for Catia is particularly good, as is the COSMOS plugin for Solidworks.

[DanFrederiksen]

apology accepted. maybe a new account name too while you're at it? to completely let go of that beastiality.

as for fluent for catia, what's the total price for that?

[asst9999]

ya few k for university edition and up to few M for super computer compatible build

[whotookmynickkokokai]

Good stuff.

Just curious.

what are the inputs you need at pre-simulation (besides the model)?

[Kenjineering]

This is very interesting! 5 STARS~! MORE

[brooke959]

where can you download this simulator simulation

[jaxpower1]

this is obviously from an engineering firm or something and they pay thousands of dollars a year for the licenses of the software so i doubt this is the place to be asking that.

[MetalManiac257]

Can you really get away with not having to make the physical prototypes? I was thinking that because the software is a mathematical model of reality, thus not 100% correct, or is it close enough?

[t2thei]

let me guess:

LAING?

:D

[NguyenJamie]

turbo xD lols

[Tempest20000]

Is there a specific reason that this particular type of semi-open "inductive" impeller is not used on standard (liquid-pumping) centrifugal pumps?

[ValExperimenter]

Abrasive wear of the sealing surface and leakage is easier to manage with a closed impeller.

[encikraju]

anyone know where i can get the CAD drawing of the impeller?

[djillali8722]

you need 32Go of RAM (DDR 3) + 3.2 Ghz of processor (quad) to keep working the animation with flurent

[Particletrace]

OUCH! This was cfdesign. I only used a 32bit XP workstation with a dual core 3.2 GHz processor and 3 GB (using unofficial 3 GB windows switch) of RAM. standard stuff.

[kamran007007]

I am looking to model some few simple geometries in Flunt can you please help me ?

or do you accept clients

[MotoX88294]

woah! 32 Gb?

[alecTHElion]

the only problem is that that much at current speeds would slow a syststem down, as the cycle (not clock) would take conciderably longer, also most computers cant handel high enough numbers to manage the contents of such space (different to harddrive as that uses sectors)

[gearhead364]

I noticed that you incorporated stator vanes into the design in order create a more uniform flow through the scoll/volute, but why did you not take the next step an incorporate diffuser vanes (wedge or cambered) to further increase the total to static pressure ratio?

[kscrosson]

I agree - its pretty standard to have a diffuser at the outlet of the rotor vanes. Otherwise you'll loose more of your total pressure due to free (turbulent) expansion into the volute.

[djdelta777]

have u got anything on an external gear pump, or any links to one ? i'd be interested in seeing that

[kakarot1232001]

What is the program you used for CFD called?

[particleman111]

cfdesign - goodstuff