# Air flow through a 90 degree bend

1. Sep 18, 2007

### Kenny.s

Hi,

Firstly i dont really know anything about fluid dynamics so go easy on me :shy:

In simple terms how does air flow through a 90 degree bend? ie where is it moving faster? I initially thought it would be along the outside wall as it has further to travel.
If it makes and odds it's an ilnet to a turbo and the I.D is about 60mm

Many thanks

Kenny

2. Sep 19, 2007

### stewartcs

Fluid flows at different velocities within a pipe’s cross-section. For well-developed, ideal flow, the fluid velocities will be highest near a pipe’s center, decreasing towards the pipe walls in symmetrical fashion. In the case of a 90 degree elbow or similar union, the velocity profile is more complicated.

The attached photo is a typical velocity profile for a similar type of union. The blue areas indicate a lower velocity and the red areas are higher velocities.

CS

#### Attached Files:

• ###### flow through tee.jpg
File size:
12.8 KB
Views:
1,503
3. Sep 19, 2007

### Kenny.s

Thanks for the reply, the 90deg bend i'm looking at is swept so it doesn't have a square edge on the inside as in the T-shape union shown in the photo you attached. Do you mean that the velocity through this type of bend is more complicated? From the photo you attached i can see appreciate how the square edge causes the area of high velocity.

Cheers

Kenny

4. Sep 19, 2007

### stewartcs

Modeling the flow is complicated for either so long as they are indeed 90 degree bends with short raidi. With a "sweep", instead of the square edge, you will see a very similar but, more uniform velocity profile. Similar in the sense that the velocity on the inside of the bend will be greater than the outside of the bend. As the sweep increases in length the velocity profile will tend to act even more uniform, much like that of a straight section of pipe. Hence, longer sweeps will tend to reduce flow losses.

5. Sep 19, 2007

### Kenny.s

I'm trying to get my head round something that happens when you fit different inlets to a particular turbo.
If you fit a straight inlet pipe with a fairly large bore, possibly 50-60mm the velocity of the air entering the compressor is relatively slow. This causes the turbo to go into whats know as surge, (appologies if you know what this is) basically the pressure at the outlet of the turbo is higher than it can sustain, if you restrict the size of the inlet down to 45mm and subsequently increase the velocity of the air this problem does not occur. I dont fully understand the reasons for it or how to read compressor maps but it's got something to do with the pressure ratio across the compressor, if the air is going in faster it's easier for the turbo to apply the required effort to reach the desired boost level. An analogy that was used to describe it was the wing of a plane, if you ask it to climb to steeply whilst not traveling fast enough the wing will stall. So by reducing the inlet size you increase the air speed you effectively increase the speed of the air over the wing, rather than moving the wing faster.
This problem can also be eliminated by fitting 2 90 deg bends of around 60mm I.D one after the other right before the turbo inlet.
From what you've said 90 degree bends will hinder flow but in the instance i'm looking at it appears to help it in some way, is this possibly just because of the area of high velocity in the centre of the elbow? I'm not looking for exact figures or calculations, i'm just trying to get the basic idea of whats happening.

Cheers

Kenny

Last edited: Sep 19, 2007
6. Sep 20, 2007

### stewartcs

Well, I'm not a turbo expert, but from what you are describing, the velocity is increased due to the tandem 90 degree bends, which is what you are trying to achieve (same as having the velocity increased by a smaller diameter pipe inlet). The increased velocity results in a pressure drop at the inlet of the compressor. This keeps the pressure differential across the compressor blades high enough to sustain compression and not stall. That is to say that the blades will not slip through the air and will actually compress it, thus keeping the flow in the correct direction. If the pressure differential is too low across the blades, then they will slip through the air and not sustain compression. The previously compressed air in the boost line will then be able to reverse its direction and flow back through the compressor creating what most people call surge. It flows backward due to the higher pressure in the boost line versus the lower atmospheric pressure at the inlet (once compression is lost). The compression ratio will then increase due to the higher pressure in the boost line and low pressure at the compressor inlet. Hence, a compressor map will show you that the turbo is in the surge area with higher compression ratios.

7. Sep 20, 2007

### Kenny.s

Thats what's confusing me though, if a 90 degree bend has the effect of reducing flow why does it appear to increase it in this instance?
Also, I would have thought the increase in velocity would help reduce the pressure difference across the compressor blades, if there was a large pressure difference the compressor would have to apply more effort to pressurise the air? or am i looking at this the wrong way?

8. Sep 20, 2007

### stewartcs

The flow is not increased; it stays the same if the compressor is turning at the same RPM. The volumetric flow rate is dependant on the velocity of the fluid (air in this case) and the cross-sectional area of the pipe. Assuming the compressor is turning at a constant rate, the volumetric flow rate will be constant (however, the mass flow rate to the engine will increase as the density of the air is increased due to the compression of it). If the cross-sectional area is reduced, then the velocity must be increased to conserve energy (neglecting any thermal losses). Hence, the flow rate into the compressor is constant. The mass flow rate into the engine is increased due to the compression of the air (makes the air more dense).

The increase in velocity does not help reduce the pressure differential across the compressor blades; it actually does just the opposite - increases it. If P1 is the inlet pressure and P2 is the outlet pressure of the compressor, then the pressure differential is P2-P1. If the velocity at the inlet is increased, P1 will be decreased. Since P2 is still at the same value and P1 was decreased, the differential is increased.

Yes, if the differential is increased, the compressor will have to apply more energy to compress the air. Since the compressor is driven by the turbine, and the turbine is driven by the engine exhaust gas, and the engine exhaust gas is a function of the fuel/air mixture, and the amount of air in the cylinder is dependant on the density of the air which is compressed by the compressor, the extra effort required by the compressor is normally compensated for by the extra exhaust gases expanding across the turbine blades. Once the energy from the turbine is lost, the compressor losses it energy as well and cannot continue the compression cycle.

Remember what happens once the compressor stalls, the air in the charge pipe surges back through the compressor and out the inlet (assuming the BOV didn't relieve it first). That is the surge you are trying to prevent by keeping the compressor compressing. The compressor keeps compressing as long as the differential pressure is high enough (of course it must be driven by the turbine as well). If it drops too low, the compressor will stall. If the compressor stalls, and the previously compressed air in the charge pipe is high enough, the compressor will surge. Surge generally happens when the throttle plate is shut suddenly. This greatly reduces the exhaust gases at the turbine, which is driving the compressor. The compressor had enough rotational energy to continue rotating temporarily until the energy it had to put out to continue compressing was too great to overcome the trapped air in the charge pipe. Hence, it flows back (surges) through the compressor.

I think the confusing part is the compression ratio. If it is too high, the turbo will surge. If it is too low, it will cause the compressor to stall, which creates the situation that causes a surge also. The pressure differential should be as stable as possible to prevent stalls.

That's about all I know about turbos! Hope it helps some and doesn't confuse you more!

9. Sep 20, 2007

### Kenny.s

That does make some sense, think i need to read it a few more times though lol

So if the velocity is increased the pressure at the inlet P1 drops?

The surge i was experiencing was as the turbo came on boost from low revs, apply full throttle at say 2.5-3k rpm in 4th gear so the engine is loaded up, as the boost came in it would get to around 1.2 bar then rise and fall rapidly, could be heard in the car as a fluttering type sound, gradually settling down as the engine speed climbed to 4k rpm and the target boost of 1.5 bar was acheived.

I'll have another think about it tonight and see if it makes anymore sense in the morning!

Cheers

Kenny

10. Sep 20, 2007

### stewartcs

Yes, if the velocity is increased at the inlet, the pressure at the inlet (P1) is decreased - assuming the volumetric flow rate is the same.

11. Dec 28, 2007

### killerbee

First post. Hello Everyone. Thought I would bring this back to life, as I noticed what I consider misconceptions. I only mention it because I recently came across a severe induction pressure drop in inadequate turbo plumbing.

As air negotiates a bend, the inner stream must slow dramatically, compared to the OD stream, which must travel further in the same amount of time. With air being compressible, this is not followed exactly, and a lot of turbulence can result. With a sharp 90 located right near the compressor mouth, a large pressure gradient can exist.

The objective, typically, is to reduce the intake losses, such that PR is as low as possible. Pressure ratio, PR=COP/CIP. Whenever CIP decreases (due to restriction presumably), so does the thermodynamic heat created in the compression process, a bad thing for air density.

12. Dec 29, 2007

### FredGarvin

I hate to say this, but this is a false statement. There is nothing that mandates that the outer flow has to traverse the longer distance in the same amount of time. What does cause the problems you are talking about is due to the large amounts of rotational flow and captured eddies that form in the elbow. The inner flow separates from the wall which can cause a reversal in the flow direction near the wall. That is why you will see a lot of piping codes/rules of thumb that mention a minimum number of pipe diameters downstream of any disturbance before installing anything that requires a smooth flow, i.e. flow measurements, etc...Putting a 90 in front of the inlet of any turbomachinery is not a wise thing to do.

13. Dec 29, 2007

### killerbee

Fred,

if it was water, would your statement be the same?

14. Dec 30, 2007

### stewartcs

The inner stream will not slow down compared to the outer stream in the bend. Immediately after the bend, the velocity will be predominately faster on one side of the profile until it recovers and fully develops again. Take a look at the picture in post #2, it shows this graphically pretty clearly; or if you prefer another, look here.

The velocity profile is given for the air flow at the exit of the 90 degree bend in the link above.

The picture in post #2 shows the profile throughout. Like Fred pointed out, there is no theoretical reason for what you are saying will happen (or at least I don't know of one).

15. Dec 31, 2007

### killerbee

The pic shown above appears to be a pressure gradient profile. Maybe not.

The pressure profile results from the shear derived from the compressible nature of adiabatic gas flow. The turbulence is the source of the prerssure drop, energy lost by virtue of the gas absorbing work. No different than shaking a glass of cold water, which is then warmer after the event.

We may be arguing over semantics. A gas does exhibit some characteristics of a liquid, in this conduit example.

on edit, after looking at some old material, I realize that there is an error in my statement. The stream velocity does accelerate quite a bit on the outer part of the fitting after the corner is approached. This velocity increases the closer you approach the OD of the conduit (not laminar flow obviously)

Last edited: Dec 31, 2007
16. Dec 31, 2007

### stewartcs

The picture is a velocity profile.

17. Jan 1, 2008

### killerbee

18. Jan 1, 2008

### killerbee

What I am really interested in, is the qantitative impact of, say, 3-4 psi of inlet restriction. I have a vehicle this applies to, a poorly designed compressor mouthpiece.