Airfoil selection for perpetual 5m solar eGlider

[WulfKeeper]

G'day all

I'm about to kick off a concerted design effort for a ~5m 'perpetual' electric glider. There's a few of these getting around but in a nutshell, charge the batteries from PVs all over the 'top' during daylight, then run off the accumulated charge overnight - do it all over again for a long as the weather allows. I'll go into more depth as the ideas condense into thoughts but for the moment, I aiming for a big, slow monster

So... I need a nice, high lift, thick (low RE), slow speed (cruise around 15km/h) airfoil. There is a staggering array of options! I'd very much appreciate any thoughts, suggestions and input on selection criteria. Currently thinking something along the lines of a Rhode St Genese 34 or maybe the remarkably similar GOE 398?

I accept that I might not achieve perfection first time around But before I actually start cutting I thought I'd solicit the sage council of the hive mind

I'll probably end up building a few test wing spans (~1m) to test construction ideas and evaluate rigidity. Torsion and weight will be the enemies!

Current idea is that I'll be getting a flight controller (Cube-esque) to do most of the automated flying. This should allow me to follow the current thinking of 'let the software worry about the stability etc'. Yes, I'm sure a lot of you are shuddering at the thought However, from that I can (potentially) build the monster wing without any dihedral.

Cheers
Wulf

 

[JasonK]

rctestflight has done a whole bunch of solar plane attempts and has not manage to get one that would charge enough through the day that it could fly through the night. adding all of the solar panels ads plenty of weight, and because you need to mostly fly level, much of the day will get poor angles on the panels.

Here is a playtest with all of his tests and various builds:
https://www.youtube.com/playlist?list=PLXvxJNOIXBsM83XxVG7tDUmqlumxDKkBs

judging from his videos, I don't think your going to pull the stated goal of being able to fly through the night. Lets just start with an 'ideal' situation, 16 hours of light, 8 hours of dark, 100% solar power during the light (this last one is not true). Your battery would need to have ~ 9-10 hours of flight capacity (so what ever your 'stay in the air' energy cost is, it would need to be able to do that for 8 hours + extra capacity to not under change it) Then your solar panel setup would need to generate >1.5x the 'stay in the air' energy cost to fly during the light time and recharge the battery (really more then 1.5 as there are inefficiencies in charging the batteries).

 

[Piotrsko]

Do what the round the world electric people did.

Airfoil doesn't matter as much as strength and surface area or ease of build since you are way below anything critical as far as Reynolds numbers are concerned and full sized stuff DOES NOT SCALE DOWN. Diehederal from wing flex is sufficient.

Be a @quorneng fanatical follower.

 

[WulfKeeper]

rctestflight has done a whole bunch of solar plane attempts and has not manage to get one that would charge enough through the day that it could fly through the night. adding all of the solar panels ads plenty of weight, and because you need to mostly fly level, much of the day will get poor angles on the panels.

Here is a playtest with all of his tests and various builds:
https://www.youtube.com/playlist?list=PLXvxJNOIXBsM83XxVG7tDUmqlumxDKkBs

judging from his videos, I don't think your going to pull the stated goal of being able to fly through the night. Lets just start with an 'ideal' situation, 16 hours of light, 8 hours of dark, 100% solar power during the light (this last one is not true). Your battery would need to have ~ 9-10 hours of flight capacity (so what ever your 'stay in the air' energy cost is, it would need to be able to do that for 8 hours + extra capacity to not under change it) Then your solar panel setup would need to generate >1.5x the 'stay in the air' energy cost to fly during the light time and recharge the battery (really more then 1.5 as there are inefficiencies in charging the batteries).

---

Thanks for the reply Jason. Yeah, Daniel at RCTestFlight does some very entertaining RC work.

Rest assured I'm going into this project with my eyes open! It *can* be done. I don't want to clog this thread with links but if you're curious, have a look at AtlantikSolar's YT channel. They did 81 hours (world endurance record for aircraft <50kg). And they're not alone. There's some great resources around from journal papers to YT content. These days solar input > power consumption is *reasonably* easy.

I'm quite comfortable with the joys of MPPT's and light-weight flexible PV cells. Aircraft design... not so much

Cheers
W

 

[WulfKeeper]

Do what the round the world electric people did.

Airfoil doesn't matter as much as strength and surface area or ease of build since you are way below anything critical as far as Reynolds numbers are concerned and full sized stuff DOES NOT SCALE DOWN. Diehederal from wing flex is sufficient.

Be a @quorneng fanatical follower.

---

Thanks Piotrsko. Yes, the flights of planes like Solar Impulse 2 are fertile ground for ideas (scaling not withstanding!) E.g. Triangular cross section tail boom with PV cells on top? Yep, doing that too

Thoroughly agree on flex. Twist though... I've heard some horror stories about wing torsion rendering the control surfaces mute and the plane cartwheeling into terra firma. Balancing flex with torsion will be a major build challenge!

Cheers
W

 

[Hai-Lee]

With solar panels providing approx 100Watt per square metre you will need to consider wing area over span. To maximise wing area and to minimise drag I would consider a HUGE plank type flying wing OR a tandem wing design for maximum panel area.

I would NOT run a separate tail and to make controls less touchy I would revert to the old school separate inboard elevators and outboard ailerons. if using a plank style flying wing.

You also may need a flight controller that more closely resembles an Autopilot in operation as most controllers do not monitor the actual flight performance but just try to keep the plane attitude stable.

You need to consider including the throttle under the control of the flight controller so that the model will maintain an altitude range with the motor being used to climb when the minimum operating altitude is breached and is turned off when the maximum height is exceeded.

Finally you will need to keep the speed very low and use all forms of lift available and so thermals should be sought and held for as long as possible and so a vertical speed input into the flight controller with the ability to initiate a circling flight path in a thermal to reduce the requirement for motor driven climbs.

Of course you would need to use a prop that folds to reduce drag, (energy loss), and I would use a brushed motor with a switch, (home made), instead of an ESC because all ESCs draw current even when no throttle is applied. This would mean a full or zero throttle selection BUT the current used to control the mechanism could be as little as the current required to operate a 2.5 Gram servo.

Material selection I leave to you but if using Foamboard you should consider adding balsa to improve the strength where needed without significant weight penalty.

Just my thoughts.

Have fun!

 

[WulfKeeper]

With solar panels providing approx 100Watt per square metre you will need to consider wing area over span. To maximise wing area and to minimise drag I would consider a HUGE plank type flying wing OR a tandem wing design for maximum panel area.

I would NOT run a separate tail and to make controls less touchy I would revert to the old school separate inboard elevators and outboard ailerons. if using a plank style flying wing.

You also may need a flight controller that more closely resembles an Autopilot in operation as most controllers do not monitor the actual flight performance but just try to keep the plane attitude stable.

You need to consider including the throttle under the control of the flight controller so that the model will maintain an altitude range with the motor being used to climb when the minimum operating altitude is breached and is turned off when the maximum height is exceeded.

Finally you will need to keep the speed very low and use all forms of lift available and so thermals should be sought and held for as long as possible and so a vertical speed input into the flight controller with the ability to initiate a circling flight path in a thermal to reduce the requirement for motor driven climbs.

Of course you would need to use a prop that folds to reduce drag, (energy loss), and I would use a brushed motor with a switch, (home made), instead of an ESC because all ESCs draw current even when no throttle is applied. This would mean a full or zero throttle selection BUT the current used to control the mechanism could be as little as the current required to operate a 2.5 Gram servo.

Material selection I leave to you but if using Foamboard you should consider adding balsa to improve the strength where needed without significant weight penalty.

Just my thoughts.

Have fun!

---
G'day Hai-Lee

Yeah, this will be a full autopilot rig (as mentioned in OP it's likely to be a Cube-like FC with tweaked ArduPilot). Lots of ideas in that area. Ditto for flight 'strategy'. Probably a single big folding prop with a 750-1000 rpm motor (lots of work on eCalc already in the bag). Still haven't ruled out a dualie though. Maybe next revision...

Interesting thought on the ESC switch. Thanks, I'll have a chew on that

Material is very likely to be balsa and v. light 3 ply. Foam has some advantages but also a few 'issues'. At the risk of kicking off another foam v's wood holy war (and this being a balsa forum!), I'll leave it at that

Appreciate your input.

 

[Hai-Lee]

---
G'day Hai-Lee

Yeah, this will be a full autopilot rig (as mentioned in OP it's likely to be a Cube-like FC with tweaked ArduPilot). Lots of ideas in that area. Ditto for flight 'strategy'. Probably a single big folding prop with a 750-1000 rpm motor (lots of work on eCalc already in the bag). Still haven't ruled out a dualie though. Maybe next revision...

Interesting thought on the ESC switch. Thanks, I'll have a chew on that

Material is very likely to be balsa and v. light 3 ply. Foam has some advantages but also a few 'issues'. At the risk of kicking off another foam v's wood holy war (and this being a balsa forum!), I'll leave it at that

Appreciate your input.

With the Balsa just remember that you can laminate balsa yourself and improve its overall strength out of sight without adding too much extra weight. Also if it comes to adding ply consider using CF sheet if you can obtain any locally as it is far stronger and far lighter for the same strength.

Have fun!

 

[WulfKeeper]

With the Balsa just remember that you can laminate balsa yourself and improve its overall strength out of sight without adding too much extra weight. Also if it comes to adding ply consider using CF sheet if you can obtain any locally as it is far stronger and far lighter for the same strength.

Have fun!

Great point on the DIY laminating! I could cold-mold it. Experiment with grain direction... That'd add some strength. Hmmm... Love it! =)

I'll sure use unidirectional CF tape where applicable but CF sheet is a @#*&^@ to work with IMHO. But yeah, I won't rule it out.

Cheers Hai-Lee

For anyone following the OP along at home, seems the S4310 airfoil might be the go. A bit thicker than the SD7037 and a bit thinner than the GOE398. Might be a good compromise.

I might knock up a few ribs in a short span this weekend and see how it comes up.

 

[Piotrsko]

Thick airfoils tend to be more torsion resistant.

 

[quorneng]

Just a point but a brushless is electrically more efficient and much lighter than a brushed motor. As you are relying on limited solar power source both electrical efficiency and weight are key factors. In the same way airframe weight is equally important.

In addition high lift full size air foil sections do not necessarily work as well at model sizes. They create the lift all right but at a disproportional level of drag which needs power to overcome. It all comes down to the Reynolds number.
Air viscosity is significant at slow speeds so the 'parasitic' drag, that generated by everything else other than the bare wing, needs to kept as small as possible.
Low power flight is quite a structural and aerodynamic design challenge.

 

[WulfKeeper]

Just a point but a brushless is electrically more efficient and much lighter than a brushed motor. As you are relying on limited solar power source both electrical efficiency and weight are key factors. In the same way airframe weight is equally important.

In addition high lift full size air foil sections do not necessarily work as well at model sizes. They create the lift all right but at a disproportional level of drag which needs power to overcome. It all comes down to the Reynolds number.
Air viscosity is significant at slow speeds so the 'parasitic' drag, that generated by everything else other than the bare wing, needs to kept as small as possible.
Low power flight is quite a structural and aerodynamic design challenge.

Thanks Quorneng. Yep, all good points. Quite a challenge, indeed Thanks.

 

[WulfKeeper]

Great point on the DIY laminating! I could cold-mold it. Experiment with grain direction... That'd add some strength. Hmmm... Love it! =)

I'll sure use unidirectional CF tape where applicable but CF sheet is a @#*&^@ to work with IMHO. But yeah, I won't rule it out.

Cheers Hai-Lee

For anyone following the OP along at home, seems the S4310 airfoil might be the go. A bit thicker than the SD7037 and a bit thinner than the GOE398. Might be a good compromise.

I might knock up a few ribs in a short span this weekend and see how it comes up.

Knock up a few ribs... like that was ever going to happen. Been reading 'The Gougeon Brothers on Boat Construction' to get up to speed on laminated veneer construction. Still an amazing book (first published 1979!) Seems a fair few ultralight plane wings are built using a similar technique. Should be interesting how it comes up at this scale.

Made the frame for the mold though! Stay tuned

 

[WulfKeeper]

For those of you still playing along at home...

Well it took a bit longer than I thought to get this far... and still haven't actually started laying up the wing. But it's getting closer! I'll go and grab the epoxy and associated paraphinalia once we get out of lockdown.

This test section will be about 800mm long. Just testing the construction technique and resulting performance before I committ to the 5m version

Only pin tacked the wet the 1st layer balsa into the mold 45deg LtoR. Looks Ok so far. Perhaps unsurprisingly the 0.8mm balsa got wet through real quick. Took to the mold nicely. I'll wait until they dry then so same with 2nd layer (also 45deg but RtoL). The balsa will be under less stress during lay up if it's already leant the shape it needs to be.

LtoR: Ribs, Ribs in frame, Dry laying the 0.8mm balsa 1st layer (stringers underneath glued to ribs).

View attachment 203858

 

[Piotrsko]

Fyi: Hughes invented /refined the process building the HK1 Hercules ( WW2 era) since plywood isn't a 3 hundred foot long sheet.

Keep it up, it's really interesting as a project & I like your structure, but you may be heavier than you think

 

[WulfKeeper]

Thanks for the comment Piotrsko. The inspiration for the cross lamination was actually Australian surf boats I'll go and see what I can find on the HK1. Update: Ah yeah, the 'Spruce Goose' (made of birch). Pity it didn't get through its trials.

Yeah, it'll be interesting how the weight turns out. All the strength in the 'skin'. Hopefully a few ribs as opposed to a *lot* of ribs and cross bracing etc. It sure won't be as light as my typical 2m builds but at ~5m? Lots of failed twisty wings at that size on the Net...

Stay tuned!

 

[Jackson T]

Sounds like an awesome project! I'm interested to see how you go about mounting the solar panels to not ruin the wing's smooth surface.

 

[WulfKeeper]

Sounds like an awesome project! I'm interested to see how you go about mounting the solar panels to not ruin the wing's smooth surface.

These days it's not a particualr challenge, JacksonT. The Sunpower C60 cells are quite flexible and pretty much the go-to solution these days. Used on almost every scale, SolarAtlantik to Solar Impulse 2!

Seems the prefered mounting method is to connect the cells and use an office laminating film (through the laminator!) to make a skin. That sure sounds like 'fun'...

 

[Jackson T]

These days it's not a particualr challenge, JacksonT. The Sunpower C60 cells are quite flexible and pretty much the go-to solution these days. Used on almost every scale, SolarAtlantik to Solar Impulse 2!

Seems the prefered mounting method is to connect the cells and use an office laminating film (through the laminator!) to make a skin. That sure sounds like 'fun'...

Ahh, laminating! Wouldn't that still have a bit of a ridge at the panel edges? It would still be an improvement though. I wonder how much weight it would add. Thanks for the reply, good luck with your build

 

[Piotrsko]

Thanks for the comment Piotrsko. The inspiration for the cross lamination was actually Australian surf boats I'll go and see what I can find on the HK1. Update: Ah yeah, the 'Spruce Goose' (made of birch). Pity it didn't get through its trials.

Yeah, it'll be interesting how the weight turns out. All the strength in the 'skin'. Hopefully a few ribs as opposed to a *lot* of ribs and cross bracing etc. It sure won't be as light as my typical 2m builds but at ~5m? Lots of failed twisty wings at that size on the Net...

Stay tuned!

According to Clinginpeel the plane photographer, the HK1 made more than 6 sucessful flights, but was not good at generating "feel" on the control yoke, thusly irritating Howard tremendously.

Twisty wings means an assembly failure. I have built 4 meter thermal stuff that was more rigid than 1 meter aerobatic. As long as it isnt heavier per meter, you should be good. Ideally it should be lighter per meter.