Ion Plane Research Challenge

[Archgeek]

1 minute ago, Deddly said:

@Archgeek So does it actually help to have the engines inside the structural fuselage? Sorry I didn't quite get whether it's worth it or not from your post.

As for wings, remember that those little fins you're using can be placed on to each other as well.

Stupidly, it seems the gains from hiding the engines (which oddly aren't complete, the lead one still gets like .03kn drag somehow) are almost exactly countered by the wider cross-section.

True, but that somehow feels cheesier than the lack of exhaust occlusion.

[Deddly]

Now that I finally have my new computer, I can have another serious go at this.

I reasoned that I should take one step at a time, so I enabled infinite electricity to see if it's even possible to get off the ground without worrying about the extra weight and drag of power production. Here's my result so far:

As you can see, I just about managed to get off the ground after gaining speed down the slope near the sea shore. Unfortunately I wasn't able to maintain altitude, but I remain hopeful that "moar wing" should help. I think if we can get takeoff speed down to 15 m/s, this might actually work.

[Ultimate Steve]

Okay. This may be a bit of a Nerco, but @Deddly has done such a good job setting up this challenge and all of the information is in this thread. And technically it has only been 2.5 months...

Anyway, the point: I have refined my ion plane (It is now Ion Plane 7B) and have found a slightly better launch location, and have actual landing gear installed. I have set a new altitude record:

Unfortunately, Even with gear, it is really hard to land:

#ReviveThisThread!

[Deddly]

Excellent work, @Ultimate Steve! I wonder if you could remove the tail fin from that to reduce drag, since you have upward-curving wings that should help against side-slip as well?

Oh, and rule 2.5 specifically states that necroposting is not necessarily forbidden

[Ultimate Steve]

Yay! Also, I got up to 7903 or so, but I can't upload pics right now. The wheel physics are driving me crazy!

[W. Kerman]

Doing Munian challenge, but saw this, AAAA! Too many challenges!

[Cunjo Carl]

While working on chores, my hind brain just splatted me through a quantitative analysis of low-altitude ion planes. Would anyone like the model? It's a little complex, and probably not worth it for calculations (trial and error is easier) but it can be nice to see how the pieces relate sometimes...

[Deddly]

@Cunjo Carl, anything that relates to the challenge is absolutely welcome

[zolotiyeruki]

On 8/7/2016 at 3:46 AM, Deddly said:

Excellent work, @Ultimate Steve! I wonder if you could remove the tail fin from that to reduce drag, since you have upward-curving wings that should help against side-slip as well?

Oh, and rule 2.5 specifically states that necroposting is not necessarily forbidden

IIRC, aero surfaces don't actually create parasitic drag in the stock aero model in KSP.  They add weight, which contributes to induced drag, but no parasitic drag.

[Ultimate Steve]

Okay. So, a while ago, I began a quest to see if I could carry one of the huge xenon tanks as fuel. Eventually, after trying to store it on its side, I ended up trying to make an ion plane in a MK2 cross section, so I could properly store the tank. It kept getting bigger, because I used the mk2-1.25m adapter to store fuel, so I added another xenon tank to balance it. I realized I would need two of those to maintain aerodynamic efficiency. I left the other one mostly empty, so I wouldn't need 4 of the huge xenon tanks.

A month or so later, after coming back to this, I decided to do some numbers off of my successful ion plane (see above post). I came up with these:

1. For an Ion Plane above 6km, every one lift will lift 0.446 tons of craft. You can go lower, but these are my approximations.

2. For an Ion Plane above 6km, every 2 ion engines and one fuel cell (0.74 tons, I call them "Propulsion Units") will provide enough thrust for 1.2 tons of plane. Again, you can use a few less Propulsion Units, but this is my approximation.

2a. Subtracting the mass of the Propulsion unit, each propulsion unit will give you enough leeway for 0.35-0.45 tons of other stuff, like wings, fuel, struts, reaction wheels, etc.

3. For every 1 liquid fuel you burn in the fuel cells (plus oxidizer) you burn about 50 xenon.

4. For Ion planes above 6km, your craft will most likely (in terms of mass) be 65% ion engines and fuel cells, 20% wings (10:1 lift/mass ratio wings) and 15% fuel, structure, and other important stuff.

Okay, the math rage is over. So, I applied these "rules" to my mk2 ion plane. That mk2 bay is heavy, and that falls into the 15% category. That means I need... 72 ion engines and 36 fuel cells. So, how far did it fly?

Yay! New altitude record! (I launched from 6600 meters, but lets ignore that for now) However, I still can't land it. It is really uncontrollable at lower altitudes, the wings shear off, and I don't have any landing gear anyway.

But we may see a Kerballed flight soon!

Takeoff:


A closer picture:

Yay! Science!

EDIT: It may be worth mentioning that my ultimate goal is to launch up to 10km with a kerbal and return him safely to Kerbin.

 

EDIT EDIT: The plane is now the Ion Plane IXA (9A).

Edited August 11, 2016 by Ultimate Steve

[zolotiyeruki]

A couple of thoughts:

1) instead of lots of modular wing segments, I'd suggest you use a bunch of FAT-455 main wings.  7.8 lift rating (about 4 of the wing connectors), same lift/mass ratio, and fewer parts means less flexing and crashing.  I count 23 wing connectors per side, so you'd need about 6x FAT-455 wings per side.  Also, the big xenon tank is a 1.25m part, so why exactly do you need the cargo bay?

[Ultimate Steve]

1 hour ago, zolotiyeruki said:

so why exactly do you need the cargo bay?

3 reasons. Nobody has done it before, it looks nice, and it completely eliminates the drag from the ion engines and fuel cells.

1 hour ago, zolotiyeruki said:

I'd suggest you use a bunch of FAT-455 main wings.

Hmm, they do have the same lift-mass ratio. I did not know that. I'll be trying this later. Thanks! (Although some wing flex helps the plane remain stable, but I agree my wings were flexing a lot)

Thanks!

[Ultimate Steve]

Okay. Thank you very much, @zolotiyeruki! I added the FAT wings, and some more xenon/lfo, and a few landing gear iterations later, The Ion Plane IXd (9d) reached...

13722m! Yeah! YEAH!

It also reached 129m/s in level flight (actualy upward flight) but I don't have a picture. I should do a speed run sometime. I could also carry a bit more fuel without adding any more engines or wings. A few issues: It doesn't have much authority over sideslip, and the front canards are really wobbly (I should move those reaction wheels). So, my goal:

1. Reach 10,000m. DONE! By a lot. I might increase it to 15 or 20km.

2. Do it with a Kerbal: Not yet, but it should be soon.

3: Recover both the plane and the Kerbal safely on Kerbin. Again, no Kerbal, and no landing gear, and the outer wings sheared off again. I also ran out of EC.

BUT, for science purposes, I ended up HyperEditing some electricity in (I promise I didn't use it in the actual flight. Only for getting the plane to the mountain) so I could control the plane to try my hand at landing. And what do you know -

A no damage water landing! (Apart from the outer wings shearing off, but that was in the air)

As a side note, at 13km the ion engines have 1.8kn of thrust and 3700 isp.

[zolotiyeruki]

It looks like you're currently pitching up by 10 degrees to maintain altitude.  So, next optimization:  in the SPH, Give all your wings and horizontal winglets a 5 degree AoA.  That way, your fuselage will have less AoA and therefore less drag.  Ideally, you want your craft to point exactly prograde to minimize drag.

[Ultimate Steve]

Four things, all in order:

1) I got to 17288 meters.

2) My screenshots folder now has over 1000 screenshots in it.

3) @zolotiyeruki I tried the AoA adjustment and I could not control the plane at all.

4) I have begun work on the Ion Plane X, the one that will hopefully accomplish my goal. Unfortunately, I am having quite a bit of problems, mainly the fact that it is performing worse than the Ion Plane IX, and it keeps flipping out halfway through the flight.

 

[zolotiyeruki]

Make sure *all* your horizontal lifting surfaces have the same AoA, and that your CoM is still ahead of your CoL.

[Ultimate Steve]

Okay, I tried building a new plane (Ion Plane X) but I couldn't get it to perform well. At all. So, I decided to strap a seat to the Ion Plane 9f. What could go wrong?

*5 hours later*

Because I decided to add a bit more fuel, some recovery devices, and adjust the canards so they don't wobble, I ended up at the Ion Plane 9p. Plus, the ladders kept malfunctioning.


Jeb launching off of the mountain. This is actually the second Kerballed attempt. The first one exploded (although Jeb survived).

Jeb is now airborne, powered by (nearly 100) ion engines, all stuffed inside the cargo bay. The climb rate is high, and the plane is accelerating.

Jeb and the plane pass ten kilometers with nearly five minutes of fuel remaining.

Jeb is now at 12km and still going strong.

15km has been reached, nearing 16 with 1.5 minutes of fuel remaining.

This is my top speed (that I have proof of, I actually got to around 222m/s) in above level flight. Approaching engine cutoff.

Aaaand...

18048 meters! New record for both crewed and unscrewed! Go Jeb! I could have flown a bit higher, but I wouldn't have any electricity to control it with. Also, you can see that at this point the CoM had shifted so far forward that even pitching all the way up meant I was in a dive. At this point, I tried to land the beast.

The drogue chute is out and the airbrakes are deployed now. You can also see that I have flown 93Km. At this point, the outer wings began to oscilate, so I deployed the main chute. They kept oscillating and fell off, though. It's not 100% re-usable, but I can live with losing a couple thousand funds of wings when the plane costs nearly a million.

Fortunately, the plane is actually more controllable without those wings.

We have flown 99.3km from the starting point. Later, the marker disappears, meaning that Jeb has flown over 100 Km.

At this point, I cut the chutes so I can land the thing.

Because the airframe was designed for ion power, it is REALLY good at gliding, and therefore really hard to slow down for a landing.

And, splashdown!

Nothing exploded, except for the outer wings when they fell off.

(Also, it says I only went 20Km. That screen is really inaccurate.)

Now that is a lot of distance.

Jeb Just missed the sunset.

Some personal bests:

1) Altitude of 18048m! We started at 6500m and pulled up by 6300m, though, so it went up by around 12km.

2) Speed of >218m/s in level flight!

3) There was a Kerbal on board!

4) The Ion Plane 9p flew over 100km! (Granted, it started from 6500m up)

Okay, now I'm going to try to shift my focus to low altitude ion planes. Maybe.

Also, video of this coming Soon^tm.

 

[Cunjo Carl]

 

Ok. I think I've succeeded to make it quite approachable, but unfortunately it's probably a half-hour read at minimum. Er, I hope it helps in some way!

Beware, Math within!

Spoiler

I was considering making a definition of terms, but for such a short document, I think ctrl+F should suffice. Let me know if anything is unclear!

To start, we'll be attacking the problem with the style of a chemical engineer, seeking to use unitless variables and lumps of coefficients wherever possible.
We'll assume steady state flight conditions (constant height, velocity) and attempt to find the wing pre-tilt angle and ratio of engines-to-wings that allows us to fly at the minimum height.
Along the way, we'll develop equations that tell us something interesting about how flying works.

 

A. Level flight, the balance

Since we're flying in a straight line, many of our forces will need to cancel. Thrust and drag, for instance, also lift and weight. These are the basic concepts from which everything else will be built.
a_x = Thrust - Drag + CoriolisForce = 0    ...  Since we're not diving hard, Coriolis force is negligible so this becomes      Thrust  = Drag   Eq.1
a_y = Lift - Weight + CentrifugalForce = 0 ...  Since we're going slow, centrifugal force is negligible so this becomes        Weight  = Lift     Eq.2

 

B. Defining the Goodness Factor, G

Two useful equations separately, but they'll be more useful if we can combine them for now. Hey, thrust and weight... aren't those used together somewhere? Let's divide equations 1 and 2. We can see the critical TWR and L/D develop directly

Thrust/Weight = Drag/Lift     Eq.3
TWR = Drag/Lift      Eq.4    We can also phrase this equation as  ...
TWR * (^L/_D)= 1      Eq.5     for level fight. Hey! If it were >1 you'd go up, and <1 you'd go down.
TWR * (^L/_D)= G      Eq.6      The above formula (5) describes a specific situation, but we actually want to know how well our plane can handle that situation (flying without crashing). We can make the (somewhat) adhoc change to this, where G stands for Goodness. G is the thing we can optimize to make our plane more efficient across a broader range of altitudes. Its real world meaning is the factor you can reduce the throttle and still fly.

So we can fly either by having a good TWR (lots of engines) or a good ^L/_D (Lots of wings). I wonder if there's an optimum?

C. Breaking up terms into real-world mechanisms

Hey, there's a reason I didn't become an artist!
For now, let's look at what we mean by lift and drag. Each is more nuanced than it seems. When we check the picture we see that the actual 'up lifting' (WLU) of the wings is only one component of its "Lift" (WL) and the other part is actually acting as drag (WLB)! Let's define some terms.
WingLiftUp (WLU)  WingLiftBack (WLB)  WingDrag (WD)  And wing drag is always pointing back. Also, often the largest form of drag is
BodyDrag (BD)  caused by other non-wingy things.

So, in terms of our earlier equations...

There's WingTilt (wt)  , which is very much like AoA, and

                MassWings (M_w)   MassEngines (M_e)          MasseXtras (M_x)                     TotalMass (M)      Or, much more conveniently...
massFractionWings (m_w)   massFractionEngines (m_e)  massFractioneXtras (m_x)     and the total mass fraction is 1. Think of mass fractions as the percent of your craft that is that thing, except ranging from 0 to 1 rather than 0% to 100%. To calculate them, we can simply write m_W = M_W/M and so on for each. Incidently, from here on, ''engines' refer to everything used to make the plane go, including xenon, fuel and fuel cells.

We'll introduce one final concept, the TWR of an engine itself. This is the value listed on the KSP wiki on the parts page. It relates to the TWR of your entire plane especially simply when using mass fractions.
TWR = m_e*TWR_e    Eq.7   The TWR of your whole plane is the TWR of your engines times the percentage of your plane that is engines.

Plugging all of these concepts into Eq4, we get...

m_e*TWR_e*WLU / (BD + WD + WLB)   =   G     Eq.8   Standing for, once again, Wing Lift Up, Body Drag, Wing Drag, and Wing Lift Back.   The greater we can make the top half of the equation, the better our plane will fly.

 

D. Fleshing out terms into calculable pieces (determined by numbers in the game's physics file)

Each of the Lift and Drags can be defined in terms of physical constants/charts within the game, as follows.

WLU =  liftAngleCoeff * (lift)machMultiplier * .5* P*MW/RT  * v^2 * WingLiftCoefficientPerTon *  M_w * globalLiftMultiplier * cos(wt)

Yeowch, that's a lot of stuff. What all is in there? Let's sort it into useful chunks. 

DynamicPressure      DP = .5* P*MW/RT  * v^2  This describes the amount of air pushing against the wing. It's the reason you stall if you go too slow or lose control flying a space plane above 40km.
Lumped Coefficient  C_WL = liftAngleCoeff * (lift)machMultiplier *  globalLiftMultiplier * WingLiftCoefficientPerTon   This describes a host of multipliers the game applies to describe the wing's efficiency. The first two describe how the ing changes in effectiveness with tilt and speed (at higher machs). I have a handy calculator for these if desired- it can also be made into a short equation for our purposes. The global lift is a big convenient knob for people to turn when they want to lift better and the last one is something I made up based on the fact that every wing has a wingLiftCoefficient (liftiness) of 10*mass. It's more convenient to speak in terms of mass, so we separate out the 10 and leave the mass there. As a note, the basic fin is the only wing that's slightly better, scoring a 12 instead!
Finally, the cos(wt) describes the fact that if your wing is tilted too far back, its lift is no longer pointing straight up, so not all of it is useful.

WLU =        cos(wt) * Mw * C_WL  * DP    Wing Lift Up. Now that's a manageable number of concepts.
WLB =    sin(wt) * Mw * C_WL  * DP     Wing Lift Back. Note that WLU and WLB have the same lumped coefficient, so the U/B has been dropped in its name.
  WD =              Mw * C_WD  * DP   Wing Drag. Here, rather than using wing _lift_ coefficients, we'll be using the wing _drag_ analogues.
 BD = (FF_C + FF_e*Me + FF_X*Mx) * DP   Body Drag. The body uses Drag cubes, and we'll be describing it ROUGHLY in terms of fudge factors. Because the body never tilts and our speed stays low, it will actually work ok. FF_C describes the drag of cones leading and trailing the fuselage(s), and FF_e describes the relative drag of the engine pieces in the middle of the fuselage(s), and . The factors will change heavily depending on what parts you use. Again, this aspect is just an estimate, but it'll work ok. Finally, drag cubes have their own lump of constants that effect them, but they're all actually constant for our conditions and speeds, so we can lump that C_BD term in with the fudge factors.

 

E. Bringing it together

Before we go plugging everything in to Equation 8, we'll notice that all of the lift and drag terms have DP in them, which will cancel out. We can also convert the inconvenient absolute masses (Big Ms) into handy relative mass fractions (little ms) by multiplying by Eq. 8 by (1/M)/(1/M).

m_w*m_e*TWR_e*C_WL*cos(wt) / (FF_C + FF_e*m_e + FF_X*m_x + m_w*C_WD + sin(wt)*m_w*C_WL )   =  G  and gathering like terms
m_w*m_e*TWR_e*C_WL*cos(wt) / (FF_C + m_x*FF_x + m_e*FF_e + m_w*(C_WD + sin(wt)*C_WL))     =  G 

And this is our stopping place for the first round (I'm bushed!) Before we leave it, it's critical we understand what it means for our planes. We came in hoping to optimize out mass of engines relative to the mass of the wings, and optimize the wing tilt (AoA) of our wings. We can do one of the engine/wing mass right now, the other will take more work.

The equation has a top half and a bottom, and the top half can be thought of as the things that make your plane fly (thrust+lift) while the bottom is the things that drag it back (drag). As far as making it fly is concerned, the relative mass of the wings and engines are equally important, so the optimal amount is an equal weight of wing and engines REGARDLESS of how efficient those wings or engines are. That's surprising! In fact, almost all of the optimization here will be involved in that complicated bottom half regarding drag. The only reason we might favor wings or engines is if one is less draggy per ton. In the end, that's not too surprising! Drag minimization typically winds up being the key for making a good KSP plane.

The drag is made by nose cones, extras, engines and wings. Each creates drag a bit differently, and has its own tricks for minimizing. Wing drag (C_wd) tends to be relatively low, and is made even lower by traveling ~50ish m/s. The backwards component of lift ( sin(wt)*C_WL ) is difficult to avoid, but is lowest for low wing tilts, raising the speed. Meanwhile, the only way to make headway on the body drag terms ( FF_C + m_x*FF_x + m_e*FF_e ) is to raise the wing tilt so the plane goes slower and the good lift term is larger by comparison. It's a bit of a mess.

As interesting as what the ability to fly does depend on it's fascinating what it DOESN'T depend on. Except for the TWR boost at lower pressures, it doesn't outwardly depend in any way on pressure. Pressure unilaterally increases lift and drag, so it cancels itself out! At higher altitudes the plane will need to go faster to maintain lift, but it can go faster because there's equally less drag. The place where this doesn't even out is that lift coefficients dive sharply going beyond 100m/s, and we'll loose a ton from that. So as long as our conditions can keep us under that... hm, that's rather tricky isn't it.

 

To move forward from here.... I could try my hand at making a low-drag ion plane? Or maybe I could make a calculator for people to plug and chug into the equation above? I could develop the equations for determining optimum wing tilt. Or I can go watch telly for once. It's all good- if something sounds interesting/useful let me know!

By necessity I'm leaving this a little unfinished. Please let me know if any points are unclear, of if you'd like things further developed.

 

 

Summary: For the sake of the effects that keep the plane aloft (thrust+lift), engines and wings are equivalently important. So much so, that an equal mass of wing and engine is the ideal, regardless of their relative efficiency. The only factor which would make us choose more of one over the other is their drag. Drag is hugely important, and highly nuanced. A low drag is often the deciding factor between a decent plane and a superb one. It's a thing I've struggled with a lot for other challenges, and would be happy to share tricks and heuristics.

 

[zolotiyeruki]

@Ultimate Steve --Another thing to keep in mind is that the stock aero doesn't care where the wings are (other than for CoL calculations), and doesn't care about their shape, either.  So instead of putting fragile wing extensions on your plane, you can stack up FAT-455's straight onto the fuselage.  It doesn't matter if they overlap.  The game only cares about how much wing area is pointed in which direction.  Aspect ratio isn't considered in stock aero, either.

Edited August 13, 2016 by zolotiyeruki