A testing little question on flow rates

[stealthfti]

Originally Posted by FlowSpecialist

To answer the original question it's easiest to start with incompressible flow. That's more intuitive to visualise and you can measure it with a bucket which is far simpler than ploughing through the flow equations.

With incompressible flow the density of the medium can't alter and therefore the fluid condition upstream and downstream of the restriction is the same. Flow rate is always proportional to the square root of the pressure drop across the restriction. Most importantly it doesn't matter what the pressures upstream and downstream are, only their relative magnitude. It could be 20 psi upstream and 10 psi downstream or 3000 psi upstream and 2990 downstream. The difference is still 10 psi.

With compressible flow things get more complex. The square root law still applies but ONLY when the upstream conditions are held constant. In other words to increase the pressure drop without changing anything else we have to change the downstream conditions i.e suck harder. If we change the upstream conditions then the density of the fluid changes and so does the mass flow rate.

For an airflow bench testing inlet ports the upstream condition is atmospheric pressure. The bench sucks air into the inlet port, through the head and into the bench - or more properly atmospheric pressure pushes air into the depression created by the vacuum pump but let's stick to the word 'suck' for now even though physicists hate it.

If we change the pressure drop the upstream conditions still stay the same which is vital if the square root law is to apply. If the pressure drop doubles then the mass flow rate goes up by root 2.

For a turbocharged engine we are not actually applying the same conditions. We are increasing the pressure drop by changing the upstream conditions not the downstream ones. That brings a second factor into play which is air density.

The first impact of doubling inlet manifold pressure is that inlet manifold air density doubles. That changes the resistance to motion of the fluid. It has twice the mass per unit volume but it's also harder to get moving. Luckily a second square root law comes into play. In fact there are more square roots in flow theory than under square trees.

The total flow into the engine is now made up of two factors.

1) With twice the pressure drop the mass flow rate increases by root 2 for a given upstream air density.

2) For a given pressure drop the mass flow rate increases by root 2 if the air density doubles.

The total flow into the engine is now root 2 x root 2 = 2

In other words it stays exactly proportional to manifold pressure so with 1 atm of boost the 100 bhp engine becomes a 200 bhp one - not a 141 bhp one.

However, now go back to the bit about flow benches testing inlet flow. What happens when you try to test exhaust flow? It might be the reason why your very expensive flow bench doesn't actually work at all if it hasn't been properly designed.

Dave

This made reading the whole thread worthwhile. Thank you!

One thing I have noticed about GFN since I started reading and posting here is that reply time is much slower than on other forums I've visited. Deliberate consideration of the subject at hand requires some time. Although sometimes the deliberation before posting a reply is done in hope of keeping from sticking one's foot in one's mouth, it doesn't always work. I've done it here; and I am fairly sure that I'll do it many more times before I die.

GFN does not have the atmosphere of a chat room for kiddies, which is something that I especially like.

I, for one, am not intimately familiar with flow benches. But I will be correcting that deficiency this winter by building a flow bench. Many thanks to both DV and FS Dave for pointing the way to that.

Dave, please do write that book. I would probably have to go to school in order to understand it all, but it would be worth it.

And if you have the time and inclination, the answer to your question about what happens when one tries to test exhaust flow would also be appreciated.

TF

[sir yun]

''With compressible flow things get more complex. The square root law still applies but ONLY when the upstream conditions are held constant. In other words to increase the pressure drop without changing anything else we have to change the downstream conditions i.e suck harder. If we change the upstream conditions then the density of the fluid changes and so does the mass flow rate.''


l'll take a stab

if you push test exhaust (blow outward) the density/''upsteam'' condition changes thereby affecting mass flow. you are ''supercharging'' the exhaust

according to this reasoning it would be better to test by pulling through the exhaust port.

but what i don't understand is the effect of the adapter..if i just bolt the manifold through the head and use that to pull the vacuum i made the port quite a bit different.

*noodle cooking again*

cheers

[FlowSpecialist]

Originally Posted by sir yun

''With compressible flow things get more complex. The square root law still applies but ONLY when the upstream conditions are held constant. In other words to increase the pressure drop without changing anything else we have to change the downstream conditions i.e suck harder. If we change the upstream conditions then the density of the fluid changes and so does the mass flow rate.''


l'll take a stab

if you push test exhaust (blow outward) the density/''upsteam'' condition changes thereby affecting mass flow. you are ''supercharging'' the exhaust

according to this reasoning it would be better to test by pulling through the exhaust port.

You are correct. In exhaust flow testing the upstream conditions are not atmospheric and will change at different valve lifts. This is not a problem provided the flowbench properly accounts for everything in its calculations but my concern is whether they all do so.

Last year I was testing a Flowquik equiped bench with my orifice test plates and once we had the inlet side dialled in I ran a test by holding the plate down by hand and testing it with the bench blowing not sucking. The flow figures went all over the place. I forget the exact numbers but a plate that was showing the correct flow when tested in the inlet direction was showing something like 20% higher numbers when tested in the exhaust direction.

Clearly something in the system wasn't working at all and all the exhaust flow numbers the guy had accumulated over the previous five years were meaningless. I knew that anyway by looking at the data for a couple of heads and the exhaust CFM was far too high to be correct.

In the end he rigged up an adaptor and now tests exhaust flow with the bench sucking through the exhaust port in inlet flow mode. The flow numbers gathered like that look to be spot on. Whether the problem was in the Flowquik or the design of the rest of the bench I didn't have time to look into.

However any rigorous test of a flow bench with test plates should test in both flow directions to have any validity.

Dave

[rookie]

When you’re dealing with exhaust aren’t you dealing with a pressurized system?

[Lasse]

Originally Posted by rookie

When you’re dealing with exhaust aren’t you dealing with a pressurized system?

You are allways dealing with pressurized system, unless you are in complete vacuum. Meaningless are you "blowing" or "sucking" trought hole, higher pressure pushes air towards lower pressure. What makes difference, when you "blow" or "suck", density of air changes. Also, when you "suck"-test head with different depressions, density of air changes.

So I don't see much difference in
blowing or sucking air, but I really like to know more.

[FlowSpecialist]

Originally Posted by Lasse

So I don't see much difference in
blowing or sucking air, but I really like to know more.

If you look at the flow equations used to calculate flow across a test restriction such as an orifice plate you need to know, amongst other things, the upstream fluid density and the pressure drop across the orifice. The fluid density is calculated from the fluid pressure and temperature. When testing inlet flow you simply measure ambient barometric pressure and temperature and you know those aren't going to change during the course of a single flow test.

The upstream pressure at the orifice i.e. between it and the test piece will then be ambient pressure minus the pressure drop across the head being tested. Temperature can be assumed not to change during the air's brief passage through the flowbench. The upstream fluid density is then a simple calculation.

When testing exhaust flow all that goes out of the window. The blower motors are not only pressurising the air but also heating it, often quite considerably as you can tell by putting your hand over the exhaust of a domestic vacuum cleaner, and the amount of heating is going to change during the course of the test as the motors get hotter or work harder.

So at minimum you now also need a temperature probe between the orifice and the motors and you're going to have to read that at every valve lift. The pressure upstream of the orifice (what would be downstream of it in an inlet flow test) can still be calculated from ambient pressure plus the pressure drop across the head plus the pressure drop across the orifice.

Assuming I've got the numbers right a 40 degree centigrade rise in the air temperature through the blower motors would change the air density by 12% and alter the final corrected flow figures by 7%. That's a pretty big error if it isn't measured and adjusted for.

Finally, although in fairness it's a very small factor, as the temperature inside the flow bench changes the orifice itself will be expanding or contracting and so its flow capability will change.

I have no idea how different commerical flow benches take temperature into account or even if they correctly allow for the fact that upstream air density is not going to be the same as in an inlet flow test. I solved the problem with my own flowbench very easily. I don't have enough interest in exhaust flow to even want to test it so my bench works only in the inlet direction. My opinion about exhaust flow is it's pretty easy to get any exhaust port modified to a high efficiency from basic principles and any extra flow you're missing by not doing flow development is going to make very little difference to power anyway.

For all the above reasons if you are going to test exhaust flow it makes life much easier and removes several possible sources of error if you do it in the inlet flow direction by sucking through the exhaust port.

Dave

[sir yun]

interesting.
I have a FQ. not used yet it arrived today..w a 110v/14v DC adapter.

but was planning to only do vacuum testing anyway. but thank you for reminding me that i should assure calibration before jumping in both feet.


i'm very much interested in exhaust flow as on an A series head there is lots of work there to be done and on the latest cams and small engines the exhaust manifold type affects power by quite a lot (>10%) .

i'm planning on flowing both intake and exhaust both directions as well (but pulling only).

[brucepts]

The static pressure pickup "sees" the same air condition changes as the differential pressure so the Cd of your sharp edge orifice plate does not change. If you see a change flowing the opposite way you have a problem with your plate or bench design.

Plenum design and quality of flow to your orifice plate will play into this. This has been extensively discussed in great details on various forms of flowbench design on my flowbench forum. I welcome anyone who is thinking of building a bench to take the time to read all they can before settling on a flowbench design.