So, we established in part 1 of this development story that there appears to be room for improvement with the OEM plenum/inlet. In this blog post we will take you up to the prototype testing phase via some design and development!
How bad is the OEM inlet?
We noted in the last blog that the air distribution to the cylinders was not even across all 6, the CFD results were only showing flow velocity which is only a small part of the puzzle. The CFD software can output highly detailed pressure and flow results to each individual cylinder, I’m not going to post up all of the numbers here but I can tell you that the difference in flow to the cylinders is approximately 2% between the best and worst inlet runner which does n't sound so bad but our target is <0.5% so how do we go about that? To improve the flow characteristics it was simply a case of design an alternative in CAD then run another CFD study and check the results…… then do it again and again and again as many times as required until the design is optimised or in other words no matter how much you change it you don’t get any more improvement. That sounds simple enough but in reality you have to have to consider several aspects such as can the part designed actually be manufactured? What price point are you striving to achieve? Is the design going to be robust enough? etc.
The development journey - Test#1
We will be marketing the inlet as a “race” part so decided to go straight for the 82mm GT3 throttle body as the parts are relatively cost effective and for what you lose in small throttle openings you more than make up for when at WOT (wide open throttle). This won’t make an improvement on the distribution between cylinders on its own but should help reduce pressure drop at WOT. All CFD test from this point are with the 82mm GT3 throttle.
First of all we tried just replacing the rearmost part of the plenum with a guided/sectioned type design which incorporated the 82mm throttle. The results showed a slightly higher pressure at the manifold outlet, which is good, likely down to the 82mm throttle but the flow distribution was significantly worse than the OEM parts with a 5.5% difference between the best and worst inlet runner. One point of note is that the OEM parts and this rear only part retain the front tube that links the two main plenums, there is almost no flow at all in this rear tube and it appears to only be present as a pressure equalisation link which is fine but it’s not very effective (at WOT) and means the flow from the throttle body to the front most cylinders is a much more complex and longer distance than the rear most cylinders. Could this be the reason we are seeing uneven flow to each cylinder? The below pictures are of the CAD part tested
The development journey - Test#2
Next up we went back to basics and literally created an “ideal” flow path from the throttle to each of the 4 plenum inlets in CAD, laid them all on top of each other and hollowed the part out, this is visualised with the different colours on the picture below then the hollowed out part below. This created a nice looking “swoopy” part as can be seen. The CFD results were not quite so pretty though with the results only very slightly better than the previous test with a 5% difference between the best and worst inlet runner. The resulting pressure for each runner was quite different where there was now more of a bias towards the forward most cylinders. This was pointing us towards a solution to keep the rear and front transverse tubes but link them with some restriction between them to encourage the flow to be evenly spread between the two.
The development journey - Conclusion
Based on the above results we tried several variations to the model you can see below, shown in the top picture in its most simple form then below that looking almost ready for manufacture, mainly changing the size of the link pipe between the two main tubes to get the flow in both front and rear tubes to match and also changing its position from left to right to ensure the flow going to the left and right cylinder banks match. After several iterations we had the CFD results yielding a 0.3% variation between the best and worst cylinders and a minimal pressure drop. Go back and compare the CFD result video below to the one on part one of the blog and you'll see what we mean! That result is one we are happy with but would it work in the real world….. ?
Time for a prototype! Subscribe to keep updated on the next stage of this blog either on the MRN website or our FaceBook page.