PDA

View Full Version : Piston engine mechanical efficiency



dmaher
January 2nd, 2012, 23:21
Table 508 Piston engine mechanical efficiency...

I can tune this table to match either...
a) power vs. altitude
b) power vs. manifold pressure

But, carefully matching either one leaves the other distorted.
It seems like the best solutions is a compromise…neither is ‘right’ but the sum is somehow least ‘wrong’.

So…
I can find a value that will match exactly how power drops with throttle at sea level.
But the power will then drop off too fast with altitude.

Or …
I can find a value that will match exactly how power drops with altitude.
But the power will then not drop enough with throttle.

There is no other method to tune efficiency (?).
Aye...what’s the solution? Any suggestions?

Danny

This chart just for illustration, but it is the one I'm trying to match.
http://www.baytower.ca/photo/power2.jpg

power2

Brett_Henderson
January 3rd, 2012, 05:43
This is an interesting endeavor :salute:


HP is pretty much a function of MP and RPM. Are you probing for the difference between MP limited by a throttle setting, and MP limited by available MP (altitude) ? .. For example.. 25" at sea-level requires a partially closed throttle, and 25" at 6000msl would require not only a wide-open throttle, but high atmospheric pressure to begin with .. and that same 25" (at a given RPM) will produce different, actual HP, at different altitudes .. ?

The first problem, is how MSFS deals with altitude. It does a reasonably good job of modeling the drop in power via the drop of available MP with increased altitude.. but the flight-model's weakness in its method for calculating power (your ultimate target) shows its ugly head, as you try to optimize the mixture. It "penalizes" an other than optimal mixture, by restricting fuel-flow on both, rich-of-optimal, AND lean-of-optimal, mixture settings. You can see this by taking an aircraft with a fuel-flow gauge up to 6000msl without ever leaning.. then while level at 6000msl begin leaning. As you lean, power will (as it should) increase, but fuel-flow will also increase, as you're leaning. The optimal mixture (best power), and fuel-flow peaks, are simultaneous.. showing clearly that MSFS makes use of fuel-flow as an unrealistic variable, throwing the whole power algorithm into question, when calculating power at altitude... which is what I believe you're hoping to achieve.. ?

So, from my estimations (I could be in error), the compromise you seek, is itself compromised by MSFS's flight model's attempt to find an overall compromise.. :mixedsmi:

Also, there's the MSFS's unrealistic modeling of lapse-rates (temperature decrease with altitude), an important variable in your quest (as mentioned in your chart).

Now, the big question.. Are you hoping minimize these compromises just for the sake of doing it.. ie... "seeing" the MSFS HP variable behave more realistically (I can relate) ? Or are you trying to achieve more realistic performance as seen in cockpit, by a combo of: MP, RPM, IAS across the altitude/throttle-setting sectrum ?

Aside from a few, inter-acting XML "gauges" using variables yet discussed, to constanty re-calculate HP AND flight-dynamics, and then "force" the model "use" those results, the latter is a tall order.. considering the MSFS compromise isn't that far off.

Still though, and interesting endeavor.. will make for an interesting/eductional discussion :applause:

fliger747
January 3rd, 2012, 11:28
It would be nice if there were more points available to specify values for. I use these tables a lot to attempt to replicate the available power tables as closely as is possible. As mentioned above the best result may be a least worse situation, but it works out fairly well on average.

Several items: For supercharged piston engines with multiple stages of supercharging available, FS does not allow for direct modeling of the shifting of the supercharger ratios, and the power loss involved in the use of the supercharger. An example the R2800 can loose as much as 400 hp to the supercharger in high blower! To approximate the correct speed performance at altitude it is necessary to use other means since the delivered HP will be wrong, such as mach drag etc. Many factors affect internal motion resistance of a piston engine. To use the R2800 again as an example, about 200 HP was gained at high RPM by a redesign of the oil scavengening system to reduce the power loss just from slinging all of that oil around in the case! Enough to allow a 100 RPM increase in max RPM.

Non turbocharged piston engines actually do a bit better as altitude increases than one might think, as waste extraction has less back pressure o fight. FS models turbocharged engines reasonable well, but in actual RW operation they can ba PITA (not the animal protection outfit).

If you wanted to see how much of a least worse solution you have arrived at you could do a linear regression....

Cheers: T

dmaher
January 3rd, 2012, 11:53
Thanks, that’s a lot of helpful insight.

Brett, It didn’t occur to me there was an issue with fuel flow…that could account for a lot.
If the lapse rate is slightly off too that would skew things a bit further.
I use auto mixture for all the tests to be confident there’s best power.
The hope is that getting the power sort of empirically right sets up a good foundation for the rest of the model.
Modeling power in an intuitively style would give a good result too, but it would take a lot of experience. The gauges could certainly be scripted to deliver the ‘corrected’ values.

T, the option I’m looking at most is matching the lower chart (MP vs Power), getting right the low throttle and corresponding low power. And then using turbo (this engine is not supposed to have turbo) to manage power with altitude. Unfortunately there’s no way to script a turbo's wastegate in XML. If I could I might have a chance to match the ALT vs Power too. Without, the effect isn't that flexible.
Using aircraft.cfg boost and critical altitude I think I can make a reasonable Alt vs Power chart.
But it’s going to have way too much power in some regions.

The linear regression idea is also a great idea, I have two viable tbl.508's and the best compromise is the average. It's a way to empirically approach the intuitive style - which gives me a bit more confidence. I'm very curious to see how that would chart in some test flights. It would also allow the engine to stay normally aspirated. A big plus for authenticity.

Yeah, in a program that allows 4096 pixel texture maps, having 4 points each for tbl.508 and tbl.509 seems truly unfair. Even just 2 more points would open up new possibilities.

Brett_Henderson
January 3rd, 2012, 14:08
In a way, it's like the MSFS guys faced a similar dilema. They needed to not only have power altered realistically by available MP, they needed power to ALSO be altered by other than optimal mixture, per altitude (and then throw into the mix, the power potential difference between 25" with a partial throttle, vs. 25" at wide-open throttle.. via induction-system fluid dynamics)(possibly the mountain you're trying to climb ?). Did they create the direct tie twixt fuel-flow and power as a fix.. or was it an existing weakness in the model ?

Regardless.. the innacurate fuel-flow irks me to the point where I re-do the gauge bitmap, and use it as the gauge it more closely resembles, an EGT gauge (climbs, peaks and then falls off during leaning).. though that still leaves me grumbling about realistic fuel consumption, as you choose to run LOP, or ROP for specific reasons.


The hope is that getting the power sort of empirically right sets up a good foundation for the rest of the model.


Yes and no.. The actual power yield you end up using might (likely) require far more tweaking of things like; wing-efficiency, parasitic-drag, and induced-drag... when tryng to get a model that will demonstrate realistic runway use during takeoff, AND a realistic climb-profile, AND realistic airspeed at cruise, AT different altitudes.. than a less than realistic power yield.

IOW.. if "realistic" power nets accurate cruise speed at 8000msl, it will likey leap of the runway, in a fraction of realistic runway use... and so on.. **sigh**

I've found that (for singles)(I always shoot for cruise altitude realism, and then climbs, and then takeoff performance, repeating that sequence until I'm satisfied) a higher than realistic power yield, as a base, makes all the other tweaking yield a less compromised overall flight model.


The gauges could certainly be scripted to deliver the ‘corrected’ values

Indeed .. ;)

fliger747
January 7th, 2012, 09:54
The method I like to try to use, at least as a baseline, is to get the power correct (as possible) then adjust the drag to meet published cruise values, which are probably the most accurate performace figures we have available. In service aircraft seldom meet test figures.... Some items such as rate of climb are the hardest to replicate and are quite weight and air speed dependent

Then there is thrust, in which one must go into the prop tables, a world of it's own.

It is possible to achieve correct cruise figures with either low thrust/low drag or high thrust high drag ratios. Getting this correct yields the closest takeoff and climb values. MS does constant speed props OK, but falls down a bit with fixed pitch installations. Very small adjustments of a fixed pitch value affects absolutly everthing. For slow aircraft you might consider using a low speed theory value other than 80 knots, which just about covers the whole flight regiem of say my SuperCub....

Fuel consumption, we can adjust this to get a closest value. WWII large piston engines averaged about 0.5 gph/hp, but obviously this would vary a lot depending on mixture, altitude, RPM etc. I reccomend the use of the Herve Sors utility AFSD to monitor these parameters in flight.

Not perfect, but some pretty good results can be achieved.

Have fun! T