Exhaust manifold design test result..

[Dave Freeman]

Just to complete as per gixxer my car is set up....

S4 engine, brand new, 6,000 miles

De catted

Stainless straight through exhaust

Electric charge cooler

Re chipped #4

Still to do list

New T35, 360 hybrid turbo

Ram air

Dave

[Gasman GT3]

Bit late posting this but I actually picked mine up from Alunox as I was working in the area.  Leon is a really nice guy and very knowliedgable.  Any doubts i had about his credentials disappeared when I walked into the workshop and saw what else he was working on, lightweight E type, Cobra, and Surtees F1 car.

Can't wait to fit it now, just waiting for a free weekend to go over to my mates in Wales who has a big workshop with a lift.  I will take lots of piccies and post the install.  I am also decatting the car as well as getting the Turbo refurbed/upgraded as i figure that 17 years of having bits of cast fired at it wont have done it much good.

[Dave Freeman]

Keep us posted as to how things go, enjoy.

[RobinB5]

@CHANGES, Dave, your original post was great, extremely enlightening! Do you know if/what simulation tools were used to help design the Alunox manifold? I only ask as I work for a company who develops CFD (computational fluid dynamics) software and would love to apply it to this manifold geometry, comparing and contrasting the fluid dynamics behaviour with the original manifold. Such a study would require a CAD model of both, which I assume would not be forthcoming due to IP concerns. Unless you know otherwise?

In the field of the design of flow systems the use of a relatively new approach, topology optimisation (https://en.wikipedia.org/wiki/Topology_optimization), is becoming more widespread. See for example this abstract: https://www.sae.org/publications/technical-papers/content/2017-01-1339/ Huge opportunity to be applied to manifold design. Considering I've just bought an S4 with an Alunox manifold I'm more than a little eager to do some studies myself  

 

[CHANGES]

@RobinB5

Thank you for taking the time to read the post.  The tools I used were very much old school algorithms/ equations. However when I was consulted by Alunox and noticed their primaries were the same size as my calculations we formed a mutual interest.  How Alunox arrived at the sizes was not discussed.   I must stress the calculations I used gave only an initial approximation to work from , as constant variable pulsing flow has its challenges..  We also needed to consider the pressure waves travelling in front of each pulse at circa 750 mph and how they react/ invert or nullify within the design.  On top of all that you need to take into account the velocity as this will determine the low pressure zone on the back of each pulse and how it interacts with any pressure waves at certain RPM..  Because we have a constantly increasing and decreasing volume, pulse sizes /pressures and velocities , a full range of calculations are needed .    The size of the Primary tube are so important to get correct to cover the full range of flow needed without creating to much or to little back pressure.   To much can cause detonation, poor cylinder evacuation , excessive EGT and drop in performance .. To little can cause a lack of velocity again effecting cylinder evacuation and VE..   Its only when you do high volume flow practical testing you start to refine the calculations and get a feel for how its going to pan out..  ..  But as stated this does not account for pulse flow , that gets calculated in after..       It was only when I introduced the Volumes my engine was going to produce in comparison to std unit that we had a difference in sizing.. this was at the collectors/ secondaries..   As the pulses converged two into one the original Alunox set up was to restricted for my engine at high RPM .  This would cause the pulses to compress and their velocity would drastically increase .  This in turn would introduce back pressure, the siphoning effect on the collector would then be drastically reduced or cancelled .  This is where my design on the collectors came in .. they were sized to create a consistent flow dynamic without stepping up so much as to introduce unwanted pressure wave inversion.  I was also using single scroll turbo so the intersection needed better flow streamlining.    A couple of pics following will show how that area changed.. 

  

I must say with all the computer modeling tools/ programs available , it will make this sort of design quicker and easier to quantify.  Interestingly I part used Fluid Dynamics when calculating the induction side , the computations seemed to suit the application better.    However Just taking the manifolds and measuring flow comparisons will give you very little useful data for reasons previously stated..   You must also use the given engines exhaust volume  to produce the test range parameters, remembering the exhaust volume has a direct correlation with induction performance/ boost / VE etc..   In short its one huge balancing act...   

Good luck with you studies.