Andy Ross wrote:

> Vivian Meazza wrote:
> > The attached diff models the output of a gear-driven
> > supercharger. I've used a 3rd order polynomial which matches the few
> > examples I have been able to find, giving near-linear output up to the
> > max-power rpm of the engine, with a little tail-off below this point,
> > and a larger one above. More importantly, it produces realistic
> > results.
> Cool.  Can you resend the patch as a "cvs diff -uw" output for
> readability, and remove the unrelated changes?  Are there any sources
> for the curve you are fitting to?  I'm not completely unwilling to
> stick magic numbers into the source code where they produce acceptable
> behavior; I just want to put a comment in explaining where the blame
> lies. :)

Well, first I thought of using some real maths based on this lot:

Then I found this:

The supercharger described is a smaller (much) version of that fitted to the
Merlin, but it has the same arrangement of throttle ahead of the compressor.
I'm very familiar with the engine - it is the same as the one in my track
car (un-supercharged). The supercharger output described is pretty much
linear up to the maximum rpm at 7000 rpm. It tails off a bit before that. I
reasoned that the Rolls Royce engineers would have gone for pretty much the
same output shape. I also wanted to take into account the surge boundary for
later use when we model the so-called 'boost control cut-out' so I added a
more rapid fall-off in output beyond the maximum power rpm. I did some curve
matching, and came up with a suitable 3rd order polynomial. It's a very good
fit up to max power; thereafter ... I'll look into that issue later in the
context of the boost control cut-out. It probably doesn't matter too much.
Not too much magic in the numbers. 

As to the throttle equation - it's a hack but it works - the P51d and the
Hurricane are easier to land. This is pure voodoo except that 750-850 rpm
seems to be the chosen idle speed for most piston engines. Big - slower,
small - faster. I'd like to do better. It might be possible to put a lower
limit on omega in PropEngine.cxx. 

Meanwhile, I attach a revised diff using -u - w. 

Hope this all makes sense



Attachment: YASim.diff
Description: Binary data

Flightgear-devel mailing list

Reply via email to