[Cadence]

TIC gets full points for such an in-depth review

What is the final word six+ months after the first post?

We need an up pipe with a flex/expansion section and no bad restrictions.

Is Grimspeed the #1 choice for an STI with the stock turbo?

[RFFG]

Originally Posted by Cadence

TIC gets full points for such an in-depth review

What is the final word six+ months after the first post?

We need an up pipe with a flex/expansion section and no bad restrictions.

Is Grimspeed the #1 choice for an STI with the stock turbo?

x2...



In for results...

[DrPhyzx]

Originally Posted by Turninconcepts.com

Second lets take a look at the position of the flex joint. Through this whole pipe the flex section with its neck down is the worst place in the entire pipe for flow. What makes it even worst is the fact that it's RIGHT BEFORE it enters the turbo. Nothing like stalling the gases before you put them to work. Yuck. Don't get me wrong. I still feel that a flex section is needed, but a better place for it would be down low so the air has a chance to smooth out before entering the turbo.

Agreed on many points, some doubts on others, but I am sure you are wrong on this last one. There is no way the gasses are going to "smooth out". The turbulence, once generated, will only grow along the pipe. The lower you place that section, the more drag the added turbulence is going to create when integrated over the length of the pipe.

[P&L Motorsports]

I say put it lower, here's why.

Watch water flow down a clear PVC drain, it sort of follows the contours of the piping and rolls around like a child going down a water slide.

Now, watch a fluid going through a pressurized environment, even as simple as a spray bottle.

All the air gaps are taken up, the fluid has filled every void, and there really is no chance for any sort of visible cavitation of flow.

Now, I'm no fluid dynamics major, but I know enough to have a serious conversation with any physics geek out there. The fact remains, I think it would be easier for the flow to settle back into a more laminar environment if the flex is earlier, then right before the volute entry.

Just my $.02.

-Jorge

[DrPhyzx]

Originally Posted by P&L Motorsports

I say put it lower, here's why.

Watch water flow down a clear PVC drain, it sort of follows the contours of the piping and rolls around like a child going down a water slide.

Now, watch a fluid going through a pressurized environment, even as simple as a spray bottle.

All the air gaps are taken up, the fluid has filled every void, and there really is no chance for any sort of visible cavitation of flow.

Now, I'm no fluid dynamics major, but I know enough to have a serious conversation with any physics geek out there. The fact remains, I think it would be easier for the flow to settle back into a more laminar environment if the flex is earlier, then right before the volute entry.

Just my $.02.

-Jorge

I think you are misunderstanding key points. First, cavitation and turbulence are two different things. There is no cavitation in the uppipe. This is not a two-phase system (liquid and gas): there is only gas.

Instead, think of wind tunnel movies where a smoke trail passes over some irregularity. The irregularity seeds a wisp of turbulent flow, which grows in lateral scale as the trail moves onward. Turbulence originating at the walls of the downpipe will grow from the edges inward until it fills the pipe. The more you can limit the turbulence to near the walls (in the boundary layer) the less you will impede the flow. Once turbulence has been created, laminar flow will only reestablish on a very long timescale if at all: seconds at least... enough time for the gas to travel a few football fields let alone to the turbo.

Now... all of that said, I would not be surprised to check the Reynolds number and find that uppipe flow is turbulent anyway. Do you know the flow rate? With that, it is easy enough to calculate.

[P&L Motorsports]

Originally Posted by DrPhyzx

I think you are misunderstanding key points. First, cavitation and turbulence are two different things. There is no cavitation in the uppipe. This is not a two-phase system (liquid and gas): there is only gas.

Instead, think of wind tunnel movies where a smoke trail passes over some irregularity. The irregularity seeds a wisp of turbulent flow, which grows in lateral scale as the trail moves onward. Turbulence originating at the walls of the downpipe will grow from the edges inward until it fills the pipe. The more you can limit the turbulence to near the walls (in the boundary layer) the less you will impede the flow. Once turbulence has been created, laminar flow will only reestablish on a very long timescale if at all: seconds at least... enough time for the gas to travel a few football fields let alone to the turbo.

Now... all of that said, I would not be surprised to check the Reynolds number and find that uppipe flow is turbulent anyway. Do you know the flow rate? With that, it is easy enough to calculate.

I may be able to get that, actually.

Are you looking for maximum airflow through the cross section of the up-pipe diameter, or at a given point in time? Again, I'm no expert, but for the latter flow rate would be subjective to things like RPM, and engine airflow, which is subject to a ton more stuff (obviously).

EDIT: While I'm at it, in your wind tunnel analogy, the only question I have is how that dynamic changes in a pressurized environment.

Case in point. If you spray a fuel injector in an atmospheric environment, you get a nice cone pattern, but under pressure and velocity I've seen those spray patterns get reduced to streams of fuel. Albeit to a different level, but fluid dynamics do apply to gases in similar ways to fluids. One is just infinity more viscous then the other.

I like this! I'm learning something!

-Jorge

[DrPhyzx]

Originally Posted by P&L Motorsports

I may be able to get that, actually.

Are you looking for maximum airflow through the cross section of the up-pipe diameter, or at a given point in time? Again, I'm no expert, but for the latter flow rate would be subjective to things like RPM, and engine airflow, which is subject to a ton more stuff (obviously).

EDIT: While I'm at it, in your wind tunnel analogy, the only question I have is how that dynamic changes in a pressurized environment.

Case in point. If you spray a fuel injector in an atmospheric environment, you get a nice cone pattern, but under pressure and velocity I've seen those spray patterns get reduced to streams of fuel. Albeit to a different level, but fluid dynamics do apply to gases in similar ways to fluids. One is just infinity more viscous then the other.

I like this! I'm learning something!

-Jorge

Yeah, I have a lot of fun thinking about this stuff too. Fluid dynamics are not my specialty (I'm a particle physicist), but I do have to pay attention to this stuff when designing things like cooling systems for instrumentation and do teach it to undergraduates. Let's assume what we are interested in is the flow at max boost when the turbo is fully spooled up. That's the case in which we want to make sure the uppipe is doing its best.

I'd probably just do a ballpark calculation to start with, to see if it's obviously turbulent or laminar. For that, the volume flow rate of the turbo, the temperature and pressure of the hot side (pressure across the turbo would be fine) and the cross sectional area of the uppipe should be enough. Unless I can easily find data on the viscosity of air as a function of temperature and pressure I'd probably use viscosity at STP. Again... if it's turbulent by orders of magnitude, this isn't going to make a difference.

That said... for such a short tube at these velocities, it may just not matter whether the flow is theoretically turbulent or not since the turbulent flow may not fill the pipe until you're already at the turbo. However, creating additional turbulence (e.g. flex joint) is something you want to do as far along the pipe as possible on basic principle for the reasons already noted.

In all seriousness, doing this right means supercomputer time. I wouldn't be surprised if that research has been done, and the results may even show in the FHI design (limited by the constraints pointed out in the original post).

[BunaSTi]

suscribed

[boxtwo]

Very interesting reading, thanks to the contributors! Come to a STi forum and get a lesson in physics, I love it! Supercomputers??? Cool, that's my dad's bread and butter. Sicortex is their new company.