Jump to content
Sign in to follow this  

Recommended Posts

Hi,

at this moment I keeping myself busy with trying to retrieve a polar from a virtual sailplane

So here's what I've done so far - this panel is added to the plane you want to test:

screenshot_01.jpg

At the bottom right is a ballast gauge to set the water ballast to anything from zero to full.

The gauge at the top middle is doing the flying - stabilizing, slowing for best LD and speeding up to max.speed.

Once the flight is finished the polar is calculated and saved.

It can be used for the L-NAV straight away or viewed and saved for posterity with this:

screenshot_02.jpg

So you can build a library of polars.

It ain't finished yet but I'm getting there.

:cheers_s:

Share this post


Link to post
Share on other sites

Hey, impressive stuff...

Could this software be used to help diagnose the polar of a glider in development? At the moment I'm doing a similar thing manually (for the ASH25), clearing the weather and flying the glider smoothly while I measure the sink rates to draw the polar and compare it with the required real one, but it's a laborious process. The thing is I'd prefer to see a range of readings for the polar, rather than the computed coefficients to better compare with the desired polar (ok.. I also draw little 'x's on a piece of paper I have the correct polar printed on...). Also I iterate the polar measurement for each of the flap settings. So far it seems FSX does what you'd expect with the effect of ballast - the polar is shifted down and to the right in the ratio SQRT(weight / dry weight) so I've been concentrating on the 'dry' polar. Flaps is a bit more complex as the 'composite' polar assumes the optimum selection, so at the moment I overlay the various polars to try and ensure they cross over at the right speeds. This whole area is difficult to get straight in FS.

I'm not sure if you're already doing it, but using TE sink instead of vertical climb automatically gives you a slightly more accurate snapshot sink value in case the speed isn't exactly constant....

Now I think of it, how about you auto-fly the glider and zap the .air/.cfg file until the polar is perfect ! :D

cheers - Ian

Share this post


Link to post
Share on other sites

Here are some answers ( and questions):

Could this software be used to help diagnose the polar of a glider in development?

Yes - it would be perfect for just that. It already showed me the default airfiles for the DG808S are a complete disaster. B)

The thing is I'd prefer to see a range of readings for the polar, rather than the computed coefficients to better compare with the desired polar.

Just give me the format you want those readings in and I'll see what I can do... :rolleyes:

Also I iterate the polar measurement for each of the flap settings.

That's one I haven't thought of - I'll have to look into it - but I'm almost sure it will show up in the polars.

So far it seems FSX does what you'd expect with the effect of ballast...

That's what you would expect but so far I get completely different readings flying with or without water ballast.

I'm not sure if you're already doing it, but using TE sink instead of vertical climb automatically gives you a slightly more accurate snapshot sink value in case the speed isn't exactly constant....

There's nothing constant in any readings inside FSX (nothing changed from FS5 there...). I'm using buffers all over the place to get some average readings I can use.

Now I think of it, how about you auto-fly the glider and zap the .air/.cfg file until the polar is perfect ! :D

What do you think I'm doing? :lol: I'm lazy - I install a minipanel in the plane, load it in the perfect setting - no weather - no thermals - and let it fly for a couple of minutes - and there's a polar. Depending on the plane it'll take 10 to 15 minutes to have the results you want...

Cheers.

Share this post


Link to post
Share on other sites

Flaps were on my mind 8 years ago when I wrote this piece of code for an early version of a polar gauge:


PANEL_SERVICE_PRE_INSTALL: // set all controls to zero

   trigger_key_event(KEY_MIXTURE_LEAN, 0);

   trigger_key_event(KEY_THROTTLE_CUT, 0);

   trigger_key_event(KEY_SPOILERS_SET, 0); 

   trigger_key_event(KEY_ELEVATOR_SET, 0);

   trigger_key_event(KEY_ELEVATOR_TRIM_SET, 0);

   trigger_key_event(KEY_GEAR_SET, 0); 

   trigger_key_event(KEY_FLAPS_SET, 0);  

   break;

Just to make sure the flaps are in the off position from the start...

(the mixture and throttle settings are for my Astir CS - it used an engine to simulate winch launches...)

Cheers.

Share this post


Link to post
Share on other sites

Slowly - but surely - getting were I want to be:

screenshot_03.jpg

LS8-18 polars with 0%, 50% and 100% water ballast.

Next: polars from a plane with flaps and ballast...

Cheers.

btw. can anybody tell me how to attach these screenshot as thumbnails ?

Share this post


Link to post
Share on other sites

btw. can anybody tell me how to attach these screenshot as thumbnails ?

I click on the "Browse...". then the "Attach This File" buttons at the bottom of the compose post.

What are you doing? The insert image button at the top?

Share this post


Link to post
Share on other sites

Hi Max - are the curves on your graph the Av^2+Bv+C quadratics? I appreciate these A/B/C are the parameters you want to find for your gauge. I'm working on the actual .cfg and .air settings that will give me the correct performance for the aircraft as per the manufacturer specification, with all the complexity being at the low-speed and high-speed end of the polar. So the graph that would be helpful to me would be a plot through the actual readings, or a table of the results I could put into Excel, e.g.

93, -0.41,+0.121

98, -0.45,+0.082

106, -0.53,+0.002

123, -0.73,-0.011

(I added (A:INCIDENCE ALPHA, radians) in case that's possible). I'd prefer the simple graph though as this really is an iterative process.

Fundamentally I am having a nightmare with the ASH25

( a ) simply trying to get the L/D ratio to REDUCE below what should be the max L/D speed (you see on the LS8-18 how the L/D maxes out at the slowest flying speed, it doesn't peak at say 102kph and be less at 90kph.)

( b ) then I need to configure the flaps so they IMPROVE the L/D ratio at lower speeds - I haven't got on to this yet.

I've adjusted [flight-tuning] parameters in the aircraft.cfg, and tweaked the zero-lift drag (CD0) and lift/AoA table in the .air file, and after thousands of iterations I can get the L/D ratio broadly right through the speed curve and have the stall speed roughly correct, but I'm really struggling to program the low-speed part of the polar for negative flap (in fact FSX doesn't seem to work properly with negative flap anyway, so the technique is to put the most negative flap in as zero and *all* the other flaps in as positive but tune the polar correctly for the negative flap setting (that FSX thinks is zero) and then program the flap behaviour for a little bit of flaps out (which the pilot thinks is zero, but FSX thinks is a bit positive). I would expect the 'induced_drag_scalar' in the cfg would help but as far as I can measure it's effect extends up to 125km/h + so it's not much use in reducing the L/D below say 110 km/h.

I fully appreciate this isn't a problem for the A/B/C coefficients, as they fit a quadratic onto the glider performance that really has to fit well for max l/d and faster, as that's what the instruments need.

cheers

Ian

Share this post


Link to post
Share on other sites

Hi Ian,

to my experience it's hopeless trying to get the low speed part (below best glide or minimum sink)of the polar right, without "tricks". You know that in the airfile there is no table for the cd coefficient, corresponding to the ca table.

To my understanding the FS flightmodel calculates the drag entirely from the lift and airspeed, considering some extensions as the Oswald efficiency factor, winglets, some fuselage effects, and so on. However, flow separation at the airfoil and stall effects seem to be hardly considered.

regards,

Peter

Share this post


Link to post
Share on other sites

Hi Max - are the curves on your graph the Av^2+Bv+C quadratics? I appreciate these A/B/C are the parameters you want to find for your gauge.

Yes, they are - and it works ok for what I want. I also use them for the LNAV.

So the graph that would be helpful to me would be a plot through the actual readings, or a table of the results I could put into Excel, e.g.

93, -0.41,+0.121

98, -0.45,+0.082

106, -0.53,+0.002

123, -0.73,-0.011

(I added (A:INCIDENCE ALPHA, radians) in case that's possible). I'd prefer the simple graph though as this really is an iterative process.

I can give you a readout of the complete range from almost stall speed up to 250 km/h. Incidence alpha is no problem too.

What interval do you need ? Are 5 km/h steps sufficient ?

Cheers,

Max.

Share this post


Link to post
Share on other sites

Hey Max, two decimal places would be helpful for the sink number at the low speed end, and that would be great. For interval, 5kph is great at the low end - from 120km onwards a 20km interval would be fine if that reduces the measurement time.

I'll end up with the numbers on a graph like this and compare it with what I'm aiming for:

thanks, Ian

post-17895-088783900 1295602718_thumb.pn

Share this post


Link to post
Share on other sites

Hey Max, two decimal places would be helpful for the sink number at the low speed end, and that would be great. For interval, 5kph is great at the low end - from 120km onwards a 20km interval would be fine if that reduces the measurement time.

I'll end up with the numbers on a graph like this and compare it with what I'm aiming for:

thanks, Ian

So it could be something like this... :rolleyes:

Changing the interval at the higher speed end of the scale won't reduce the time needed so I'll let stay as it is.

Can you explain the two (or rather three) little peaks at the end of the scale on your graph ? Where the lift makes a jump.

Cheers,

Max

post-24134-011947400 1295605917_thumb.jp

Share this post


Link to post
Share on other sites

that version would be great for me.

I produced my graph with the 'jumps' just by typing the numbers in from your screenshot into excel, to demo the graph I'd produce - when I do it for real I'd expect to overlay the required polar from the real glider specification sheet. Your data has the jumps in it -e.g. between 197 and 202 kph.

cheers,

Ian

Share this post


Link to post
Share on other sites

Thanks for showing me. I loaded the file in Excel too and it showed I was flying too fast. A little less elevator took care of that... :rolleyes:

Cheers.

Share this post


Link to post
Share on other sites

Is the Polar Gauge ready for release, Max?

Hi Dave - no, not yet. I haven't yet solved the Win7 problems I've got but I ran into something completely different.

To create polars this way planes need to have reasonably correct air files and most of the ones I have (12 in all) don't.

Some of the older planes still use the fuel tanks for ballast - that's a lot of re-writing.

I downloaded a copy of AirWrench to do something about it only to discover it didn't know anything about water ballast (amongst things)

Jerry is looking into that...

Cheers,

Max.

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.
Sign in to follow this  

×
×
  • Create New...