CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Main CFD Forum

Help me design a conv-div nozzle for race engine!

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Display Modes
Old   May 21, 2009, 19:23
Default
  #21
New Member
 
Join Date: May 2009
Posts: 11
Rep Power: 9
Null is on a distinguished road
The reason why you got flak here was , for starters , what you are trying to accomplish is not a "nozzle", its an intake diffuser.
second, you have no clue of the nature of pressure waves and their total priority in internal flow. to put it in the typical american inches-of-water-cubic-foot-per-minute: you can flow bench your head all you want but without pressure differentials no air is gonna go from one side to the other.
in your corvette you have the pressure variation in the plenum, which its influenced by the primaries, and althought is not as radical at the boundaries at the intake valves, it is still non-steady.
the problem is to find the maximum velocity at which the air will come in and crash with the restriction of your restrictor plate ( vena contracta by Dr Pradtl sounds familiar?) and come to the other side in a mixture of attached bubbles , turbulence, and crappy momentum.
the job of your diffuser is to keep a pressure gradient so as to not let the air separate as it slows down-that increases the pressure in the plenum which incleases the pressure in the primaries which allows a big pressure differential at your intake valve which-if you IVO and IVC has been properly designed- gives you a lot of volumetric efficiency.
trick is, as the pressure pulses come and go, the velocity at the restrictor goes up and down.
and yor shape goes-in THE RIGHT STUFF terminology- out of the envelope.

you have to find the pressure variations really close to where the restrictor is supposed to be so you can figure out the flow and therefore the velocity and therefore the gradient of your diffuser and the gradient of the intake trumpet. and then you can design- thats the pretty part : HOW- come up with a shape that works well at the variations that its gonna encounter.
summary:
the air has to come with out separating at the restriction.
it has to slow down close to zero at the end of its diffusion ( the INFAMOUS PRESSURE RECOVERY where its kinetic energy transforms into pressure adiabatically-you didn't put a refrigerator inside the plenum)

and CFD for that is way down the road . first you need to understand the design of pressure gradients.
I give u some magic dust: Potential Theory.
Null is offline   Reply With Quote

Old   May 21, 2009, 19:32
Default
  #22
New Member
 
Join Date: May 2009
Posts: 11
Rep Power: 9
Null is on a distinguished road
almost forgot,

never talk about recomended ratios in a scientific forum. those famous recomended ratios like Dm/d* and all that bullshit always comes from a monkey with a saw that started welding and cutting again and again until he came up with "an empirical formula " out of his ass.

and, its an oximoron to say the you are gonna optimize it in steady state when you have just said the problem is unsteady.

just trying to help here.
Null is offline   Reply With Quote

Old   May 22, 2009, 10:00
Default
  #23
New Member
 
Digital_Junkie
Join Date: Apr 2009
Posts: 22
Rep Power: 9
MaxCFM is on a distinguished road
Null, Thank you for your observations and advice. I would have appreciated it more without all the condescending cynicism. I know that I asked for a thrashing when I created such a stereotypically uneducated screen name for a scientific forum! You made some good points and I will heed your direction. The problem is unsteady as you pointed out but "I have to crawl before I learn to walk" using CFD. For now I'll use steady state testing for basic flow analysis and when I get more experience I'll try unsteady using combustion data as boundary conditions. We use Kistler 4005 sensors in the intake port near the valve and Kulite sensors in the plenum and airbox. They have given very good results in spite of the heat and severe vibration caused by aggressive valve closing at high RPM. Not like production cars.

The picture I posted was not of my corvette. I don't have a Corvette and if I did it wouldn't have a 4-cylinder in it.
MaxCFM is offline   Reply With Quote

Old   May 22, 2009, 11:08
Default
  #24
New Member
 
Join Date: May 2009
Posts: 11
Rep Power: 9
Null is on a distinguished road
Hey Man,

Sorry about it-wasn't ill intentioned. I've been watch steward and colbert for too long.
Null is offline   Reply With Quote

Old   May 22, 2009, 11:37
Default
  #25
New Member
 
Join Date: May 2009
Posts: 11
Rep Power: 9
Null is on a distinguished road
Yes , I read later of the four cilinder.

you have to come up with a way to introduce a pressure sensor close to the STANDARD UN-ADULTERATED plenum intake so you don't disturb too much the basic pattern. and with good resolution-frequency of sampling as big as you can get.
You realize you are thinking of CFD still as drawing tons of diffusers until one kinda works with out knowing what is the THEORICAL maximum, right?

the pressure pulse in the plenum is different and way milder than in the primaries. Thats the reason you have a plenum: cause the system acts as a spring. but since you are not modifying primaries or valve throats, well, measure the plenum . Also, the plenum is a place where you want to shift from the 1D analysis of waves in pipes to 3D so you can find all kinds of weird stuff. but that is adding to your original question and putting the problem a notch up.

You can choose between "the lets try 300 shapes" in the computer or "lets figure out calculation of pressure gradients".
I don't know how inclined are you to puzzles, but for now, I think that trying different shapes knowing the pressure at the plenum will give you a good aproximation.
the other way implies learning a lot of stuff from different places that don't seem to be related.
remember that everything affects everything -the shape you choose will affect the pressure at the other side, something you'll find out when you actually machine all trumpets and install them and measure, which defies the purpose of using a computer to save that much work. But , a good shape that doesn't provoke detachment of the flow over a reasonable range will do the trick.
I'll be more worry about your valve timings, valve shapes, intake and exhaust ports shapes and primaries. you get twice the volumetric efficiency playing with the intake than with the exhaust.
Null is offline   Reply With Quote

Old   May 22, 2009, 13:14
Default
  #26
New Member
 
Join Date: May 2009
Posts: 11
Rep Power: 9
Null is on a distinguished road
The name wasn't an issue, but when you started talking about the 'recomended' ratio of the throat or any of that empirical rule of a thick sore thumb is NASCAR all over. ...gime a piece of titanium angle IIIIIron and some duct tape.....

there is actually a paper on the design of steam locomotive extractor by a really famous engineer. its revered as the most complete design method for steam extractors ( which in this case is a nozzle but fundamentally the general problem of internal flows: how to accelerate and decelerate with the least losses while the flow goes wherever its wanted in the shape its wanted) all ratios and more ratios , equations and more equations with ratios, but it seems he never bother to find what really makes the shape tick.
there are two types of engineers of course, The Thomas Edison type and the Stanley Hooker/Harry Ricardo type.
If you like ratios check this:

http://www.trainweb.org/tusp/lempor/lempor.html

you don't use combustion data. you use the field out of the boundary between the primaries and the plenum. put a sensor in there or model the primaries and put a sensor as close as you can to the inlet port with the required resolution and noise resistance. correct for the distance to the valve plane itself. combustion data is meaningless to the intake unless you're simulating the chamber and the intake valve-exhaust valve boundary.
Making a time dependent simulation is not that complicated. the same time you are gonna use understanding the mathematics of discretization for a nozzle in the anderson cfd book can be used to understand any comercial software time dependant parameters including how to input the pressure field variation at the boundary. ( aka the pressure profile over time from your pressure sensor)
forget about the math for now. just learn what is it that the software are you using needs to know for unsteady time dependent runs .
Null is offline   Reply With Quote

Old   May 22, 2009, 13:20
Default
  #27
New Member
 
Join Date: May 2009
Posts: 11
Rep Power: 9
Null is on a distinguished road
last thing, if you go for simulating from the intake valves , therefore including the primaries, make sure to put your intakes in the appropriate spot at the plenum. a lot of weird stuff happens in the plenum and you might be getting flow coming AT THE INTAKE when what you want is the least possible pressure where the diffuser empties, in which case baffling is awesome but affects the geometry that the pressure wave sees. so baffle carefully.
Null is offline   Reply With Quote

Old   May 24, 2009, 23:48
Default
  #28
New Member
 
Digital_Junkie
Join Date: Apr 2009
Posts: 22
Rep Power: 9
MaxCFM is on a distinguished road
Thanks Null, You have provided much food for thought. I have a habit of referring to data acquired using the combustion system as "combustion data". Sorry for the mix-up. I never thought of using a baffle but I suppose it makes sense if it can actually improve airflow by attenuating deleterious wave tuning characteristics. The plenum that houses our intake trumpets is a variation of a log style manifold. The trumpets are raised from the plenum walls about 20mm and angled slightly toward the inlet. I may be able to couple with engine simulation in the future for dynamic analysis.
MaxCFM is offline   Reply With Quote

Old   May 25, 2009, 07:15
Default
  #29
New Member
 
Join Date: May 2009
Posts: 11
Rep Power: 9
Null is on a distinguished road
Hey Max, just to make sure, I meant baffling affects the geometry the wave sees in the strict sense of expansion/contraction interactions ( the mechanics of the spring system-the whole thing acts as weight plates attached to springs). the baffling is to control the kinetic pressure at the plenum intake ( air coming at it at speed due to turbulence inside)
inside the plenum, airflow and wave tunning are two different things. a pressure wave will cross a turbulent field , bounce on the wall and come back barely affecting the flow.
to quote you " improve airflow by attenuating deleterious wave tuning characteristics"
it tailors your airflow, without caring about the waves BUT it can affect the wave.
You know how Porsche uses plenums with interconnecting valves with variable areas? or Ferrari uses 3 plenums with several valves?
well, a badly thought baffle acts not so much but close to a valve: it changes the geometry wave wise.
Null is offline   Reply With Quote

Old   May 30, 2009, 18:11
Default
  #30
meb
Member
 
Marco Evangelos Biancolini
Join Date: May 2009
Location: Rome - Italy
Posts: 77
Rep Power: 9
meb is on a distinguished road
The aim of the restrictor is to limit the power limiting the air introduced in the engine. At high RPM the flow is chocked and to achieve the maximum charging the losses need to be minimised.
There is a trade off in component length because increasing the overall length the friction at the walls is increased, reducing the length produce stepper area changing and subsequent vena separation.
Typical angles are 30 deg for the convergent cone and 6 deg for the divergent one. The cleaner is the flow the better is the performance, this means that also the flow condition at the begin of the convergent are very important (but this is of major concern for FSAE where the restirctor is located downstream after the throttle).
If you have already optimised the layout considering a steady flow you are close to the correct result because at the restriction the pulsation is smoothed by the plenum.

MEB
meb is offline   Reply With Quote

Reply

Thread Tools
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

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Axisymmetric nozzle design using the MOC Will James Main CFD Forum 22 May 24, 2012 01:25
compressible flow in a counterflow nozzle d.vamsidhar FLUENT 0 November 24, 2005 02:45
CFD and Engine design varun Main CFD Forum 3 June 29, 2005 16:52
Industrial Nozzle Design Bharath Main CFD Forum 0 December 4, 2002 18:14
Info: Short Course On Thermal Design of Electronic Equipment Arnold Free Main CFD Forum 0 August 10, 1999 10:18


All times are GMT -4. The time now is 03:15.