oops

 

Earl3

I checked you numbers and agree that at a flow of 2Kg/S you will see an approximate average velocity of 50.3 M/S.

By my calculations this leads to about 6"H20 pressure drop.

However you made some assumptions that may not be accurate and I wonder about your flow map. The map does not provide much indentifying information in the borders. Nothing to identify conditions. Additionall the mass flow you assumed may not be accurate. Consider the following.

3.6 Liter Displacement
6000 RPM
No Boost
No inlet duct loss
100% Volumetric Efficiency.

Then flow through the system is:

Q=(3.6/2)(6000)(1/60)=180Liters/S
180L/S(1/1000)1.225=.225Kg/S Approximately what you show but at no boost.

Correct flow to more actual at boost:
Boost =1.6@ 6000
Intercooler disch temp= 200F
Vol Eff=.95
(For the moment don't niggle at these assumed numbers just follow through with the calc. Minor adjustments are not going to change the conclusions.)

Q=.225(1.6/1)(95)(70+460)/(200+460)=.28Kg/S Flow @ boost.

Average velocity in inlet by ratio 50.3M/S(.28/.2)=70M/S=230Ft/S

Velocity Pressure
P=pv**2/(2gc)=.076(230)**2/(2x32.2)=62Lb/Ft**2=.43psi=12"H20.

12" H20 is greater than the 3-4" you mention. I doubt that the friction losses in the short section of inlet duct are higher than say 2-3"H20. So the velocity pressure is not negligble as you suggest. It is greater than or at minimum equal to friction losses.

Consider the following:

1. You mention Reynolds number in your post. Check your Re, the charcteristic length and distance required to achieve fully developed flow. Given the short length of inlet duct the flow is not likely to be fully developed. As a result there is likely no uniform velocity profile.

2. At the noted air velocities the incompressible flow assumption may not be appropriate.

3. As the inlet duct is in an area where ambient temperatures are high say 200F your assumption of an adiabatic system may not be accurate.

One of the static taps is near the outside of a bend where velocities are likely higher. This will lead to low static measurements.

Given that you don't have fully developed flow or a predictable or uniform velocity profile the static readings you make not provide the information that you want. Calculated above is an average velocity of 230F/S but actuall local velocities will be much higher. If these high velocities occur at a static tap you will measure very low static pressure. And it won't be representative of the entire section of the duct.

Can you maintain constant flow conditions over a long enough period of time so that transient affects do not affect our measurements? The actuall analysis of this is beyond the scope of this post. It involves the physical conditions and your instruments.

Someone above mentioned a flow bench. The reason these are used is so that the test can control and evaluate all of these other affects.

Have fun with this. You will learn alot. However the measurements you are making may not provide meaningful results.

Check over my numbers. I have been wrong before; more times than I want to think about.

 

David,
I'm with ya for the most part -one question: since its a 4 stroke, would your mass flow numbers be cut in half? i.e. you only move air in on 1 out of every 2 revs. I noticed you divided the displacement by 2, but assumed that was because there are 2 turbos (assuming they provided equal flow to each bank). That would put your calcs at 0.14 kg/s (per turbo) -similar to whats on the comp map.

Ex: Q = (3.6/4)*6000*(1/60)=90 ltr/s =>0.115 kg/s

Add boost:
Q=.115(1.6/1)(0.95)(70+460)/(200+460)=0.14 kg/s (per turbo)

..now we're back down to 3" H2O of dynamic pressure (in an ideal long tube)

 

Earl3

You caught an error.

I divided the displacement by 2 due to the 4 cycle engine characteristics.

I am embarrased to say that I neglected the two TC's. That certainly makes the flow rate shown on your compressor map make more sense.

At a uniform average velocity due to the .15Kg/S mass flow I get a delta P of 3.5" H20.

The delta P due to friction in the inlet duct is very likely on the same order as this velocity pressure. Vp can not be negligible.

Check the Re and the number of characteristic lengths required to get fully developed flow. Consider that entrance disturbances, where the flow transitions from the filter box to the duct, will likely propogate down stream into the measuring taps.

I believe the observations and conclusions in the earlier post, still apply.

I like it when someone looks closely enough to catch me at an error. Thank you for the time you put into your review and response.

 

David, for the reason I've already posted doing this on a flow bench is not likely to be viable IMO. That being the case how exactly would you do this better?

I think his testing is fine for the point of the matter which is to find restrictions. Swap a component and look for a lower pressure drop. If pressure drop = lower then new part = improvement else swap something else. To really do it right swap one thing at a time until an improvement is noted then rinse and repeat until you have eliminated most of the restriction in the system. Short of spending cubic dollars this seems to be practical to me - do you agree?

 

David, make no mistake, I'm very glad you bring these things up. I wondered a bit why the Autospeed folks didn't talk about total pressure and the effect of changing diameters and materials along the way -figured I was thinking too hard Maybe they neglected it, didn't consider it, ...who knows.

About the only thing I can think to do with it on the car is to fit a section of straight pipe in front of the turbo and add a little extra flex ducting to make it connect back to the stock stuff. This should make a static reading more stable and at least get in the better ballpark for theoretically getting the dynamic contribution. If we can show that the dynamic pressure at some point in the pipe is likely in some range (say 5-10") of total pressure and the static has dropped, say, 40", a before and after of upstream changes should at least show up if enough losses are eliminated.

..deep thoughts.

 

BLKMGK

HTML Code:

That being the case how exactly would you do this better? 

I don't know. Let Earl try and see what he finds.

Earl,

My opinion, as this is not a calculation, is that you aren't going to get what you want.

You have the energy and enthusiasm, give it a try.

Can you can maintain a steady state load, RPM, fuel supply, etc condition, long enough to even make a measurement.

Can you get repeatable results. Try the test several times with no changes to your installation.

See if you can get some measurable difference when you change to a better inlet system. See if these are repeatable.

I've seen a lot of money wasted on tests that weren't thought out. I seen alot of decisions made on bad tests.

On the other hand no one ever accomplished anything that they didn't try.

David

 

Great info! Subscribed...

 

David do you mind me asking what you do for a living. I would venture to guess Engineer,Professor or statistics analyst. You seem to have a higher understanding of problem solving.

 

Paulie,

It couldn't have been much of a guess and you are correct. I am an engineer.

I paid a lot of attention to IC engines in school. I know that this will be a big surprise because no one else on the board ever had similar ideas, but I thought I would be a race car driver. I thought the engineering would help.

It didn't work out; the race car fantasy.

I have been at it, engineering, for awhile and have made alot of mistakes; and have learned from them too. Some of that may have shown up in the posts as well.

I forgot to than Paulie for the compliment. There are a lot of smart people posting to be noticed and praised is appreciated.

David

 

Earl
I would be interested to see something similar on the charged inlet side

Specifically, I would like to see what the pressure readings are:
* After the air leaves the turbo
* After the air leaves the intercooler
* At the throttle body

Is this anything you would be prepared to do at any time?

(BTW - I edited my post from yesterday regarding intercoolers and y-pipes
I'd be keen to hear yours and David Cs thoughts on this if you ever have time to read it)

 

Very interesting. Considering to do the fenderwell intake mod. Look forward to your findings, thanks to OP and others for contributing knowledge.

I'm down for sharing cost on the gauge, if anyone organizes, PM me. thanks.

 

Great info. subscribed :-)

 

Anyone have a diagram of the various air pathways around the turbo engine? I'm trying to visualize airflow, obstructions, and areas of possible improvement.

 

No diagram but as I understand it air flows into the MAF from the airbox, splits into two paths with the passenger side having to cross over the front of the motor to the passenger turbo and the driver's side taking a less tortured path down to the turbo. After exiting the turbo both go up, through intercoolers, up through the fenders into the engine bay, merge in a Y, then enter the throttlebody, where it again splits in half and goes ot either side of the flat motor.

The OEM intake piping, especially going to the passenger side, flattens, twists, and bends all over the place. The Y where airflow first splits is something that it seems some port for flow and even on the driver's side the piping flattens a good bit - possibly for tire clearance? You can pretty easily see where there might be restrictions trying to move the CFM it takes to support say 600FWHP in the OEM setup! How the BOV divert from intake right to the turbo intakes is a slight mystery to me but they must all pass pretty close there or there's piping, when I swapped my BOV it was too tight for me to get a good view of where things went but that's certainly some air routing too

Does that help at all? Pictures of the OEM intake piping have been posted as has some of the aftermarket stuff and you can see that in total it's a ton of pipe to flow so much air though!