I have been asked to help out at my old university with the Formula SAE team. This includes the use of an engine model (a competitor to Mickrick's WAVE simulation) from one of the leading engine consultancies, and arm of a company that runs a fairly successful F1 team. Not sure if I can name them as I am not really that up on the licence agreement (
). Suffice to say, this being their model, it will have been used for the design of inlet and exhaust and cam profile for their F1 entry.Anyway, I have been acclimatising myself with the model by plumbing in the data from my black S1, working from inlet to exhaust:
Forcewinder
HSR42
XB heads
SE 497 cams (std on the S1/X1/M2/S3)
Buell Pro Race exhaust (Albert's headers)
The model is fairly accurate in terms of engine dimensions and even includes the buldge in the headers which supports the C ring and flange
It is not perfect by any means, for instance the heat transfer models are not optimised and the cam profile intermediate points are estimated, but it gives me a baseline and it will not be a million miles out, with predicted power being high 90s / low 100s bhp.
I must admit, this is not all my own work, Gunter gave me a configuration for what I can only describe as a tuned up Blast (
) which I validated against the S1, I would say 95% of it has changed, however the data Gunter did give me enabled me to ensure the model was working correctly.As I said earlier, the model does output in bhp, however this is so subjective as it depends on so many assumptions related to the combustion process that I will stick to comparing volumetric efficiencies...
The simulation takes about 30 minutes to run, so firstly a comparison of different cams:
SE497 (stock)
SE536
SE551 (same as XB)
SE575
The model looks like this (which will probably mean nothing to the majority):
Attachment:
Grenade Sim.jpg [ 29.69 KiB | Viewed 916 times ]
Results from the study (torque is proportional to volumetric efficiency):
Attachment:
cam.jpg [ 55.42 KiB | Viewed 916 times ]
It should be noted that the volumetric efficiency shows the potential of the engine, i.e. you still need to optimise fuel and ignition timing. This is important for later discussion.
The 551s, fitted to XBs tend to give (with a my intake and exhaust configuration) the worst peak performance, but give better performance at lower speeds.
These results tie up well with other studies in terms of trending.
Although not the same configuration on both bikes, my S1 (497s) which shares the same top end as my Firebolt (551s but different exhaust and intake) is about 6% better on power on the same dyno. Probably not the best of comparisons as there are so many differences.
The 536s are a couple of % down on performance below 5000RPM compared to the 497s, which kinda ties up with 03's observations.
NRHS testing of the 551s v's the 536s found an approximate 5% difference, the 536s giving more power. Note that the data is a little noisy, from the dyno. The model gave me a difference of 2.1% which is about half the NRHS findings, however dyno results are very dependent on engine temperature, dyno condition and ambient conditions (despite the corrections used). The test engine was quite close in spec to my S1. Coming back to my comment earlier, volumetric efficiency is a measure of potential. Where the 551s have given better results at lower speeds, the optimum performance would be gained by less ignition advance than the 536s which, in the dyno test, would have probably given more power at the lower end. For that, you could either use a programmable system like my Dyna2000 or put an ECM in place.
NRHS testing of the 575s v's the 536s found an approximate 4% difference, the 575s giving more power. Again, the data is a little noisy. The model suggested a 0.3% difference, but the same caveats above apply in addition to the fact that this was a 1250 engine.
Best thing is that I have a model which I can compare the effects of small changes to inlet, exhaust and cam profile before cutting metal (or spending money).
