Das Beast: My E30 track / street build

[Elysian]

Here's the inside of the plenum. No baffles or bell mouths. Slight raised lip on each runner.



Dimensions to scale here. Would definitely be interesting to see some CFD.

I made a 3d model based on your dimensions and some assumptions on things that are harder to measure, but I've never used Autodesk CFD and can't quite figure it out to flow test it...

[dvallis]

Cool. Let's see your model. Post a STEP file and I should be able to import it to SolidWorks

[Elysian]

https://www.dropbox.com/s/xtmig2ixwt0134c/M20Intake.stp?dl=0

I didn't add the mount plate for the throttle body, and had to assume the diameter on the inlet tubes, as well as the thickness of the metal.

[dvallis]

That looks great. Actual runner ID is 1.38" so your design is spot on. Metal thickness looks fine.

Could you add a throttle body inlet plate with 3 ID " tube? That's all it needs.



Nice inside details by the way. Looks like the real thing.

[dvallis]

#nevermind. I just added one myself. Now to get Solidworks flow simulation working.

STEP file is here.

[Elysian]

Awesome. I have Solidworks (educational license for veterans), but having a hard time installing 2020 on my home PC, need to figure that out.

[dvallis]

Well, THAT'S pretty weird. Here's a velocity plot. Cylinder #1 would be getting starved, not #6 Aaaaaannndddd ... that's why you need engineers designing manifolds. Yikes that's ugly. Glad we're using the stock manifold.



I gave it an intake volume goal of 1000 cfm, which is in the ballpark for this head. I know it's not exact but close enough. Note that volume of air OUT equals volume of air air IN. Also shows there is virtually nothing flowing in cylinder #1.



If you have Solidworks this is a pretty good CFD tutorial.

[Elysian]

Did cylinder 1 get capped off? Looks like there's a sliver of material over it in the flow bench section. I opened my step file in Fusion and it's not covered, just want to make sure.

[dvallis]

#1 got capped or the model won't run. But I'll check the port for extra material.

FYI you can't run a flow bench on the STEP file. It imports as a surface by default. I had to play with options, convert it to solids then add the runner flange and throttle body. I'll post the SLDPRT file today.

[dvallis]

Nope. Nothing blocking #1.

The Solidworks project including SLDPRT file and CFD is here.

[dvallis]

Just for curiosity sake I cranked the mesh resolution up to "Ya gotta be f..ing kidding me" and let the sim run this morning. I used a 20 core workstation that loved to have something better to do than email.

Nothing new but the you can definitely see more detail. Here's the velocity plot.



Flow plot.



Combined



I'm sticking a fork in this one and declaring the eBay manifold done. Waiting for the stock intake to come back from powder coating.

[Elysian]

Yeah that's wild, what a terrible "design."

[dvallis]

Here's what proper flow looks like. This is an S54 manifiold.

[digger]

Well, THAT'S pretty weird. Here's a velocity plot. Cylinder #1 would be getting starved, not #6 Aaaaaannndddd ... that's why you need engineers designing manifolds. Yikes that's ugly. Glad we're using the stock manifold.



I gave it an intake volume goal of 1000 cfm, which is in the ballpark for this head. I know it's not exact but close enough. Note that volume of air OUT equals volume of air air IN. Also shows there is virtually nothing flowing in cylinder #1.



If you have Solidworks this is a pretty good CFD tutorial.

Number 1 is flowing exactly zero? Repeat this with the stock manifold and see how good it is ( or isn't)

i never have all 6 moving air at the same time as it never happens in reality. also did you work out the mass flow and pressurize the inlet for something like 2b? if you have too high velocity on inlet it biases the flow to where the inlet points.

from the look of the design number 1 will be lower but i dont think it will be anywhere near as bad as this sugests

[dvallis]

#1 wasn't exactly zero, just some fraction of a CFM. I rounded to integer.

I gave the inlet an initial condition of 1000 CFM. Also tried some velocity initial conditions around 20 -100 m/s. Got the same results. I'll try a pressure.

The sim is working. I agree the CFM in #1 won't be zero, but definitely low. It's enough to give you a feel for the design, even if absolute velocities aren't what the real head will see.