Greg,

Are you also considering that the true effective wing area?
With a double taper you can optimize visual to some extent with some tradeoff to optimize performance. Can you sweep a variation of planforms and dihedral?
Just wondering how much we loose at each joint if any?

Greg,

Nicely done! The next level of complexity is to optimize the airfoil for each design variant and see if the same trends/sensitivities still hold true. Yes, there is some reinvention of the wheel in all this number crunching, but for me at least, doing the numbers really gives one a "feel" for the design space instead of just having the results handed to you by someone else. If you can find a copy (maybe check with JT), have a look at the proceedings from the MARCS sailplane symposium from way back in the late 80s (between '86 and '88 if memory serves) where I gave a talk on XC design optimization. Yes, this optimization was a bit on the crude side (single airfoil, 2 ch design). What also may be of interest is the stuff regarding speed to fly. One wee little thing to remember is that way back then, I did the analysis using a hand calculator. Spreadsheets were a gleam in someones eye way back then...

Jack,

You bring up an interesting aspect regarding visibility. I suspect that something between mean and average chord is the appropriate choice for visibility optimisation. One would need to do some real-life experimentation to further refine what the independent variable should be for the optimization process.

Dihedral is a pretty minor player in the performance side of the equation. I suspect that the dihedral constraint gets defined by spiral stability as one really wants the vehicle to have reasonable spiral stability for flying at high altitudes.

As for the losses at the dihedral joints, these are pretty small in the grand scheme of things. The next item in the evolution of XC performance will be in the reduction of wetted area of the fuselage as just about all of the current aircraft flying have more than a wee bit of performance degradation at inter-thermal speeds due to the drag of the fuselage.

Joe, good to see you monitoring/adding to this forum. Sorry Greg, I got rid of my MRCSS stuff a few years ago in a purge of stuff I had been lugging around for 20 years. Still have all the Soartech stuff, but no MRCSS.

JT

OK I added in the 14 x 150 and it comes in about 30 seconds behind the 14 x 160. Aspect ratio starts to take a piece out.

The next level of complexity is to optimize the airfoil for each design variant and see if the same trends/sensitivities still hold true.

OK we are back on the learning curve here. For a first cut I will take TrekBiker's thinned and decambered foil set and plug it into the existing wing planforms. See if we can gain back some speed that we lost with the bigger chords.

wetted area of the fuselage

I'm loathe to part with the GJ design, especially since I have one sitting in the corner. However there are options and I only want to build one more of these for a while.

JT,

I purged my copies just a few months ago... Lug something around for a quarter century without using it and then throw it out, and one will quickly find a need for it!

Really wish I could have been at Deans shindig this year! Sounds like it was a classic event, with some really great conditions.

Greg,

Have you looked at how the design optimum changes with course length yet? Where the bigger chord plane really shines is in the shorter course events due to the clear advantage of a higher starting altitude.

Wile E. was sized for the 20 to 26 mile events that we were flying back in the '80s. Ended up making some longer tips later to have a bit more soaring performance for longer courses (e.g. Great Race). The short tip Wile E had a 140" span... Not a great thermal plane, but it could run quite nicely!

Joe Greg has my Soartech #8 ~ Is there a specific areahe can reference if you remember?
Yes I still have my Wile killer - straight line speed and L/D is it's best but flaps were addedto assist in the thermal dept.

Getting faster at this now that I have a method.

A 15 mile race does not change the ranking. I think you would have to go to a really short race to take advantage of the extra height of the 14" chord, something like the 10K we used to run in central FL.

Now the interesting part. I took a set of airfoils thinned and decambered from the AG23 series used in #8. Put those on the same airplanes plus one that TrekBiker designed. They all beat their respective wing in the AG23 series group, but the only one that stands out is the 12" chord. Now you have both aspect ratio and wing loading, and those win out over altitude. XFLR5 gave this wing a flatter glide slope below 1200m and that is what does it. Either there is a peculiarity in the program or this is the truth. In any case it highlights the importance of planform and airfoil progression. Now I understand the difference between TLAR and running the analysis. There must be more to be gained, especially with camber change.

#8 V2 Airfoils 97.11
150 Span V2 97.37
170 Span V2 98.6
12" Chord V2 89.2
14" Chord V2 97.54
TrekBiker 12" 93.13

Greg,

I would be cautious in interpreting your data until you understand why there is such a large delta between the 12" and 14". Something is amiss there...

Wait til you play with camber change! There are minutes to be gained there.

The next caution is in going towards a theoretical optimum without regard to real-life constraints. Would you spend 30 seconds of potential time for a 40 mile course to get something that works a low level thermal much easier? What about having the capability for less drag at very high speed to get through some of those nasty sink patches?

You are getting close to the time where you may want to run the designs through a variety of conditions (thermal strength, course length, inversion height, etc.) and do a weighting on the various course results to arrive at an "optimum" design. As Dr. Drela mentioned to me a long time ago, the art is in defining the optimization function, everything else is just engineering.

Joe,

Thanks for the tips. Attached is the XFLR5 polar data for the thinned and decambered airfoils. It definitely favors the 12 inch root. I will run it again and check the dimensions to be sure there is no mistake. It does pay a bit in min sink.

I have a few other models for different days, inverted good day, inverted weak day, etc that need to be refined. I wouldn't want a plane that goes fast but never gets off the field on a weak day, or lands out on every low save. We have been spoiled on the last two outings and that is what I started looking at. I counted back this year and here is the weighting of conditions:

4 big day no inversion
1 blown out no flying
1 windy, low lift
1 inversion 500m cap low lift

I know from watching it the MXC has a little better run L/D than what I am flying now and I can only beat it by flying higher. What I want is something that will run and track better than the MXC and thermal out as well as #8, ideally without having to slim down the fuse.

No error in the data entry part.

I'll just release this as I generate it. I redid the curve fit for TrekBiker's plane polar, not sure what happened the first time but it was not right. Here is a whole pile of planes run in a 26 mile course on an inverted day with the following assumptions:

Lift 0.4 m/s, adjusted down from the last model because XFLR5 is optimistic for wing only
Sink -.13 m/s
Max altitude 500m
Course length 26 miles
Distance between thermals 1 mile
Speed to fly using sink rate and no expected thermal
Final glide in sink at speed to fly
Result is in minutes

#8 218.5
150 span 261.4
170 span 201.4
12 chord 235
14 chord 221.7
14 chord 150 span 260.5

#8 V2 airfoils 219.1
150 span V2 262.4
170 span V2 200.2
12 chord V2 231.9
14 chord V2 227.1
TrekBiker wing V2 215.6

The clear winners here are the stretched spans, with the losers being the clipped wings. The 13.25" chords are more or less maxima in the design space with 12 and 14 dropping off. Not much difference in airfoils, trading places depending on the planform. The planes get spread out pretty far here, with the more efficient climbers getting ahead in the weak lift.

I've flown in this kind of day. Once you recognize it there isn't much you can do except know when the thermal is topped out and bail in time. The lift is regular, and the good teams are all going to fly the same race. Assuming the top 4-5 teams are equivalent pilots, the plane will win the race. This is in direct contrast to the boomer day when the team with the fat chord that dares fly the highest will most likely win. Recent memory shows about 4 boomer days for every inverted day, and sometimes if you are patient the inverted day turns into the boomer day. Spreadsheets are getting easier to manipulate.

No guarantees the spreadsheets are without error but they are giving me reasonable answers. Same course, 26 miles, inverted, but stronger lift and sink. Here are the assumptions:

Lift 1 m/s
Sink .33 m/s
Course length 26 miles
Distance between thermals 1 mile
Speed to fly using sink rate and no expected thermal
Final glide in sink at speed to fly

Result is in minutes

#8 83
150 span 83.9
170 span 86.2
12 chord 87.3
14 chord 85
150 span 85.8

#8 V2 airfoils 81
150 span V2 82
170 span V2 83
12 chord V2 79.8
14 chord V2 83.4
TrekBiker wing V2 81.3

With stronger lift now the 12" chord thin airfoil wins based on higher speed to fly because of its aspect ratio and wing loading. #8 with thin airfoils and TrekBiker wing not far behind. The big losers are the thicker airfoils, with #8 as built the maximum of that subgroup. Looks like the thinnest wing that will hang together on course is a good place to start, with camber change to modify the polars.

The lesson here is if you are going to fly on inverted days against good pilots, the plane has to be optimized to win these particular races. The pilot can't do much unless with a bit of luck the inversion layer breaks through and with a less competitive plane restarts higher or takes advantage of upper level lift on course. I distinctly remember flying an unballasted #8 on this type of day and watching two MXC's pull away from us, and there was nothing we could do about it. Now I understand why.

Next step is to get the polars off the recent Supra build and see how that does on various courses with camber change.

Greg,
Why is sink assumed to be 1/3 of lift?

Regards Dean

Dean,

I looked at the GPS data from the first lap at the last Sacramento contest. I added or subtracted the sink rates at minimum sink and 35 mph using polar data from Dan Edwards paper. This gave me actual air rise and sink rates. I came up with about 1/3. I know there is bigger stuff out there but particularly on inverted days I think it is reasonable.

Regards,

Greg

Added in the new Super Supra that we flew this weekend, in summary format. Here is what shows up.

The Supra is a much bigger plane than the 12 inch chord with V2 airfoils but ties it on the 50 mile good day. Up high running fast the speed to fly is fast enough to push it down the polar far enough to move into reflex mode. The Supra wins big on the inverted, light day because it can outclimb everything else. Better in cruise mode, not reflex. The Supra loses by 2 minutes to the 12 inch chord V2 that XFLR5 likes so much. Slightly better in reflex. However, the far superior min sink helps you out in case of an ultra low save. This is worth more than the 2 minutes saved over an hour and a half race.

I'm sold on camber change. I suspect the Supra would win even bigger if we optimized the area and chord for visibility and upped the wing loading a little.

Not sure how to write a function to weight the different conditions, but I would say the final design has to approximately tie the 12 inch chord V2 on the big day and beat the Supra on the two inverted days by a little. Any more that can be squeezed out would go to the big day, as there are more of these and everybody remembers them more.

If anybody wants the spreadsheet to play around with PM me.