On Thu, 2008-12-18 at 19:47 -0700, John Denker wrote: > On 12/17/2008 12:18 AM, Ron Jensen wrote: > > > I've made the constant "C" a configurable item, but I don't have a good > > name for it. The patch calls it <gagg-c> but there must be a better > > name... I just can't think of one. > > Physically, this C represents internal friction. So, maybe > "gagg_friction" or some such??? > > > The Gagg-Farrar equation uses a constant to express the power drop off > > of a piston engine as air density decreases. > > > > sigma= air_density / standard_density > > > > phi = ( sigma - C ) / ( 1 - C ) > > > > power *= phi > > > > In early testing it seems to work well. I used rho_air_manifold for > > air_density and replaced suction_loss, which was based in part on > > throttle position, with phi. This may resolve one of Torsten's issues > > with the FGPiston model. > > There seems to be a consensus that this engine model would > benefit from some TLC. > > The Gagg-Farrar equation is delightfully simple and reasonable. > However, AFAICT it assumes the throttle is wide open. That > makes sense for the original applications intended by Gagg > and Farrar ... but I don't think it does what FGPiston needs > to get done. > > In particular, I call attention to the lines in front of the > patch Ron contributed: > > suction_loss = RPM > 0.0 ? ThrottlePos * MaxRPM / RPM : 1.0; > if (suction_loss > 1.0) suction_loss = 1.0; > MAP = (something) * suction_loss; > > First of all, regarding this quantity "suction_loss", it's a > misnomer to call it a "loss" ... because as this quantity > becomes greater the MAP becomes greater.
True, this is actually a normal value, with 1 equaling static pressure, and the value decreasing towards zero as map drops. > Terminology aside, I don't think we should be multiplying by > throttle/RPM ... instead we should be dividing by (1+RPM/throttle) > roughly speaking. > This is much better behaved, especially > at low RPM. Sorry, not true. Using your model for manifold pressure, http://www.av8n.com/fly/throttle.htm#eq-map-tk, I was unable to find a value for beta which satisfied a high manifold pressure (over 28 inHg ) at full throttle / 2700 rpm and a low manifold pressure (under 15 inHg) at 0.2 throttle / 750 rpm, for a 320 in^3 engine. The model behaved even worse for the 5800 RPM red-line Rotax912 model. Further, you never bothered to declare how we will define a value for Beta. I considered a similar route last year, however since details of specific aircraft engine induction systems are very hard to come by I dropped the idea in favor of specifying the peak horsepower RPM. > Furthermore, these lines seem to be missing a rather important > dependence on ambient pressure. Well, you decided to replace the inlet pressure variable with "something" so you lost p_inlet, which is assigned to p_amb(ient) at the top of the function. > This is related to the fact > that while dynamic viscosity is insensitive to pressure and > density, the kinematic viscosity is directly sensitive. > I wrote up a first draft of the algebra involved in making a > quasi-plausible model of this stuff. It can be found at > http://www.av8n.com/fly/throttle.htm > This includes a graph of the non-decreasing power versus > revs curve that we discussed yesterday. Again, a real engine will not increase power with RPM indefinitely. There is an RPM which represents a "peak" horsepower. See, for example, this table http://users.erols.com/srweiss/tablehdp.htm which estimates peak horsepower as a function of a single piston's displacement and flow rate into that piston. The instant engine, an O-320 has four cylinders each displacing 80 in^3. The table lists 200 CFM and 80 in^3 as having a peak horsepower at 2778 RPM. If you will remember I estimated the flow rate for our engine to be 200 CFM @ 2700 RPM, and the configuration file specifies 2700 RPM as the "maxrpm" which I used as a synonym for peak horsepower RPM. As you can see, this table is very close to the current model. If you want the model to produce power beyond 2700 rpm we need to simply raise the specified maxrpm in the config file. > The analysis is not 100% complete; in particular it does > not include the frictional terms that Gagg-Farrar features. > I reckon if we add that in, we might get something pretty > nice. In particular, the Gagg-Farrar result should emerge > as a corollary of the more general analysis, applicable to > the special case where the throttle is wide open. ------------------------------------------------------------------------------ _______________________________________________ Flightgear-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/flightgear-devel
