Re: [Flightgear-devel] JSBSim Piston Engine Idle

[Ron J. <wi...@je...>]

On Tuesday 11 December 2012 13:49:51 Eric van den Berg wrote:
> So it is kind of modelled like an air pump. Interesting method.

Piston engines are basically air pumps.

We currently calculate power by dividing the mass fuel flow by the 
user-entered bsfc multiply by correction factors for mixture and spark and 
subtract a little bit for good measure. The little bit ensures the engine 
stops spinning if it isn't producing power. The propeller code will keep it 
spinning under certain conditions, and needs to be fixed up to let it start 
the engine spinning again...

 If I were to start over with this model the biggest thing I would change 
would be replacing the power calculation with the Otto cycle[note 2] pressure 
calculations. We already (as mentioned) calculate the area of [1:2]+[6:1] as 
pumping losses.[note 3] The trick to this method is calculating the [3:4] 
pressure rise to get the area of [3:4:5:6] This should let us roll the egt 
and cylinder temp calculations into the power loop and make them more 
meaningful. Right now they are largely just indications, altough egt does 
indicate power somewhat correctly.

That may be done in the future as we can then more easily add a diesel cycle.

Ron


[2] http://hyperphysics.phy-astr.gsu.edu/hbase/thermo/otto.html#c5
[3] http://mae.wvu.edu/~smirnov/mae320/figs/F9-2.jpg

> BTW
> p0 =101325 Pa
> R = 287.05
> Cp_air = 1004.68
> gamma = 1.4
>
> Handbook of Aviation fuel properties, third edition:
> net heat of combustion of AVGAS, all grades : min. 43.5 MJ/kg, 44 typical
> density of AVGAS: 710 at 15degC
> C_p_AVGAS = 2.065 kJ/kg K @20degC, approx linear to 2.710 @140degC
>
> hope this might improve accuracy (a bit),
>
> Cheers,
>
> Eric
>
> On 12/11/2012 07:35 PM, Ron Jensen wrote:
> > On Tuesday 11 December 2012 09:46:10 Eric van den Berg wrote:
> >> I see.
> >> Looking at the code (I think) I can see you are trying calculate the
> >> pressure losses in the injector/throttle valve, "airbox" and inlet
> >> tubes. Using throttle position and engine speed (was expecting cylinder
> >> displacement here also).
> >
> > The displacement is somewhat irrelevant in that it is a constant and can
> > be ignored. The modeler provides two data points; the pressure at full
> > throttle and maximum RPM, and the pressure at 0 throttle and idle RPM.
> > These are used to determine the impedance of the airbox and throttle
> > respectively. In this scheme, the engine is also treated as an impedance
> > which varies with ( 1 / engine speed ) giving infinite impedance at 0
> > RPM[1] and falling towards, but never reaching, 0 impedance as engine
> > speed increases.
> >
> >
> > We experimented with many other and more complicated intake models early
> > on, and this is the best behaved of the lot.
> >
> >> Basically your MAP at idle is to low, thus the
> >> pressure loss too high. As 99% of the pressure loss comes from the
> >> injector/throttle position, I would say for idle power setting the
> >> injector air valve should be a bit more open?
> >>
> >>
> >>
> >>
> >> I assume it is only calculated for  indication and not for engine power
> >> calcs?
> >
> > Actually, the manifold pressure is used in three ways in the power
> > calculations. First, it affects the mass flow rate. We assume an
> > adiabatic process so the loss in pressure is accompanied by a
> > corresponding loss of density. Second, the volumetric efficiency is
> > reduced by the intake pressure being less than the exhaust pressure
> > further reducing the mass flow rate. Finally, the pressure difference
> > between intake and exhaust creates a direct power loss as work is
> > performed to pull and maintain the manifold pressure drop.
> >
> > Ron
> >
> > [1] Note: Engine speed actually used is mean pistons speed not RPM.