Intake Air Box

[VWRacer]

After reading Jake's thread about building his own intake system, I though I'd share this report I came across.

http://www.me.up.ac.za/research/meng.htm

Scroll down to the thesis by Mr KW Makgata, titled "Computational analysis and optimisation of the inlet system of a high-performance rally engine". (Note: The download is 3.7 MB.)

"The dissertation deals with the unsteady CFD analysis using Fluent of an inlet system of a high-performance rally engine comprising an airbox, trumpets and 4 inlet manifolds, with the rest of the engine represented by a transient pressure boundary condition determined from a 1-D gas dynamics code, EngMod4T. The airbox geometry is parameterised and use is made of mathematical optimisation techniques in the form of the optimiser LS-OPT to improve the base design. The improved airbox design is also evaluated for off-design conditions. Limited experimental validation is performed using steady flowbench tests."

There is a lot of dense engineering language that can be difficult to follow for non-engineer types, but the bottom line is that they were able to improve the performance of a basic airbox by more than 6% by twiddling with its shape. Here is a synopsis of what they did that's clear enough for anyone to understand.



It may not be directly applicable to our engines, but it's still pretty cool!

[fspGTD]

Thanks for the link, Stan!

That is cool...  My random thoughts on this:

* I think the realistically used powerband for a rallye racer must be broader than 6000-7000RPM.  That is only a 16% speed increase.

* Despite it being sharply focused at 6500RPM, 6% power gain is impressive for just a plenum change.

* Interesting to see the optimized plenum shape the computer came up with had generally a tapered shape (and somewhat of a larger volume.)

* Pretty cool to see what the computer thinks airflow is doing in those 3D pictures.

* Too bad they didn't confirm the simulated optimized airbox by actually constructing it and dyno testing it compared to the original.

* I was not impressed with the accuracy of the model in the steady state test (off by several percentage points.)

[greggearhead]

Interesting article, and for its purpose - to prove that computer modeling can be a less expensive testing method than trial and error, it is fairly good.  However, it is also not perfect.  

Like Jake mentioned, they did no hard testing of the new design to verify any improvement.  I have been around engines and such long enough to know that what works in theory (on paper or a computer screen) isn't always best at all.  

It is interesting that the author equates a 6% increase in arflow into a 6% increase in horsepower.  I have never seen that equal of an increase before (doesn't mean it isn't accurate, just that I haven't run across it).  

I don't think they are saying that the power-band is only 6000rpm to 7000rpm, but in inertia tuning, it can only improve a narrow rpm band, as I am sure most know.  

There were some statements about Inertia tuning and resonance tuning used on inline vs. Vee engines I disagree with, but that is small potatoes.  

Thanks for bringing up the link - I love reading stuff like this.  Wish I still got Racecar Engineering magazine.

[VWRacer]

While I agree that the thesis has some short comings, bear in mind the purpose of a thesis: to prove that the author has the ability to do original research - not to find something that 'works'. In this case, the thesis would have been equally successful if the author had shown that the size and shape of an airbox had NO effect on its performance. That's the nature of a thesis.

In the real world, though, we need look no further than VW's own Mk1 gasser manifold to see that the author's solution is as equally valid in the real world as it was in the virtual world of computational fluid dynamics.



In the photo above the TB would mount at bottom left, with the air moving from left to right before entering the individual cylinder intakes. Notice that the intake tapers down from the entrance to the end, just as Mr Makgata found was best. The width of the VW manifold is not as wide as in the thesis, but my guess is that it was optimized for a lower RPM band.

I still think that the Mk1 GTI intake manifold holds lots of promise for high performance diesels, and I'm looking forward to finding out.  

[fspGTD]

I'm excited about trying the GTI manifold out as well...  

This computer modeling stuff reminds me of some finite element analysis modeling I used to model the stress and strain of my racing wheel centers before and after being lightened.

model with stress and strain anaysis, before lightening:


We confirmed the accuracy of the model by weighing the wheel centers and comparing their mass with the computer model's calculated mass.

It was interesting to see which lightened designs the computer thought were "safe"...


...and which would have stressed the material too much:


The computer modeling was also useful in predicting the amount of weight savings for the different lightened designs.

After a few iterations, I ended up with a design that according to the computer model, barely increased the peak material stress (and a modest increase in strain or "bending"), while it saved about a pound per wheel off the unsprung, rotating mass per wheel.

The University of Washington Formula SAE team who helped me do the computer modeling work later helped me do the machining as well in their machine shop:


(PS - check out that big chunk of aluminum we used as a support fixture! :shock:)

The "real test" of actually racing on the modified wheel centers has yet to be done, but I'm looking forward to it!


...ready for testing.