• 0
KnightRider

Designing high-performance FAR aircraft

Question

I've gotten the basics of aircraft design with FAR installed down (CoM ahead of CoL, landing gear close behind CoM, check stability derivatives in the FAR analysis screen, etc) as well as the basics of designing low-drag aircraft (area ruling etc), but I'm interested in any resources or information about optimizing aircraft as much as possible to perform well in terms of maximum speed and altitude.  What kinds of features should a plane have to perform better in those regimes?  I'm currently working towards sort of an X-plane program in my career mode install (i.e. building faster and faster planes leading up to a spaceplane), and I'd like to optimize the aircraft as much as possible.  I'm working with FAR and RSS in terms of mods, so the difficulty is greater than stock.

Also, how can I make my high-speed aircraft perform better at low speeds and altitudes?  I understand that it's difficult to do that, since getting an aircraft to perform well over a wide range of speeds and altitudes is very difficult, but I'm currently having issues with my latest X-plane stalling at speeds below 120 m/s, which makes landing a pain.  I typically end up trying to come in as slowly as possible (in the vertical direction) at 120 m/s and then flaring/cutting power, but this just leads to me bouncing down the runway several times, flipping over, and destroying the plane.

I'm not posting from my computer with my KSP install on it, so I can't post the craft file at the moment, but I will as soon as I get back to my KSP computer.  Basically, it's a single-engined delta wing aircraft with no horizontal tail or canards, powered by a single J-33 jet engine.  I essentially attempted to get it as close as possible to look like the F-106 (pic included).  It performs pretty well, achieving a maximum speed of almost mach 2 and a maximum altitude of ~20 km on a 20 kN engine (the engine only produces a fraction of its rated power due to airflow restrictions).  However, its performance while landing leaves a lot to be desired.convair_dart.gif

1 person likes this

Share this post


Link to post
Share on other sites

17 answers to this question

  • 0

Posted (edited)

6 hours ago, KnightRider said:

I've gotten the basics of aircraft design with FAR installed down (CoM ahead of CoL, landing gear close behind CoM, check stability derivatives in the FAR analysis screen, etc) as well as the basics of designing low-drag aircraft (area ruling etc), but I'm interested in any resources or information about optimizing aircraft as much as possible to perform well in terms of maximum speed and altitude.  What kinds of features should a plane have to perform better in those regimes?  I'm currently working towards sort of an X-plane program in my career mode install (i.e. building faster and faster planes leading up to a spaceplane), and I'd like to optimize the aircraft as much as possible.  I'm working with FAR and RSS in terms of mods, so the difficulty is greater than stock.

Also, how can I make my high-speed aircraft perform better at low speeds and altitudes?  I understand that it's difficult to do that, since getting an aircraft to perform well over a wide range of speeds and altitudes is very difficult, but I'm currently having issues with my latest X-plane stalling at speeds below 120 m/s, which makes landing a pain.  I typically end up trying to come in as slowly as possible (in the vertical direction) at 120 m/s and then flaring/cutting power, but this just leads to me bouncing down the runway several times, flipping over, and destroying the plane.

I'm not posting from my computer with my KSP install on it, so I can't post the craft file at the moment, but I will as soon as I get back to my KSP computer.  Basically, it's a single-engined delta wing aircraft with no horizontal tail or canards, powered by a single J-33 jet engine.  I essentially attempted to get it as close as possible to look like the F-106 (pic included).  It performs pretty well, achieving a maximum speed of almost mach 2 and a maximum altitude of ~20 km on a 20 kN engine (the engine only produces a fraction of its rated power due to airflow restrictions).  However, its performance while landing leaves a lot to be desired.

I played a little bit with FAR, the trouble is it makes my head hurt and kills a lot of Kerbals.

2016-01-03_00002_zps1jozolcq.jpg

I managed to get the stalling speed of this thing down to 70 m/s with full span high lift devices.   The problem i could not solve, was that the ailerons would stall with high lift devices deployed at landing pitch angles .   If I reduced control surface deflection, i'd have insufficient roll authority at higher speeds.

Ignoring the requirement to look like an F-106, you can use a canard style design since you're pulling the nose up with lift rather than pushing the tail down with downforce.  Therefore the pitching surface is contributing to total lift rather than subtracting from it.  (That's why the SAAB Viggen is a canard delta, for STOL).     You can also make the wings arbitrarily long if the outboard sections are highly swept, and have them extend well beyond the back of the fuselage.  

As I understand it, for supersonic flight you need to do two things.   One, the wings need to fit within the "bow wave" shock created by the nose cone.   The longer the span, the more they need to sweep, you might want to use a compound sweep angle where sweep angle increases as you move outboard and aft.  

Second is "area ruling",  so rather than making a delta with a flat trailing edge, which results in an abrupt reduction in cross sectional area, it's more of a notched delta double delta.

collage_zpsi1scyltv.png

The frustrating thing about trying to design for low transonic drag whilst simultaneously having a lot of wing area is that tiny changes produce huge changes in drag characteristics.  This made adding control surfaces very time consuming. The other issue is that FAR is more realistic than stock aero, in that lift increases with the square of airspeed, so you have to add a huge amount of lift to reduce the landing speed by not very much.   I wonder if you combined FAR with infernal robotics and procedural wings to make a swing wing design you could have your cake and eat it?

As for your F-106,    is the CoM quite close to the elevons by any chance?  I suspect they may be having to push "down" rather hard to get the nose up on landing, which may add to the high speeds.  How about adding control surfaces to the leading edges of the wing, and using the leading edge devices ahead of CoM to control pitch, with the elevons only controlling roll.  That way you are lifting up not pushing "down".  Alternatively, lock the elevons' pitch function and use some monstrous RCS system (multiple Vernor thrusters?) to control pitch on landing approach.... 

Edited by AeroGav
1 person likes this

Share this post


Link to post
Share on other sites
  • 0

Oh incidentally, 100 m/s is the speed the space shuttle landed at .    120 m/s is just insane.

The real  F-106 BTW

https://en.wikipedia.org/wiki/Cornfield_Bomber

173 mph landing speed, which is still damn terrifying and fast even for a supersonic fighter.  But that's only 77 m/s.

Eurofighter Typhoon 140knots,  72 m/s

SAAB Viggen  (A delta canard with much smaller wings than the craft i made) has an approach speed of 119knots, 61 m/s.

Share this post


Link to post
Share on other sites
  • 0

Posted (edited)

On 15 March 2017 at 11:23 AM, KnightRider said:

Also, how can I make my high-speed aircraft perform better at low speeds and altitudes?  I understand that it's difficult to do that, since getting an aircraft to perform well over a wide range of speeds and altitudes is very difficult, but I'm currently having issues with my latest X-plane stalling at speeds below 120 m/s, which makes landing a pain.  I typically end up trying to come in as slowly as possible (in the vertical direction) at 120 m/s and then flaring/cutting power, but this just leads to me bouncing down the runway several times, flipping over, and destroying the plane.

As with the real ones: slats, flaps and spoilers. And for the ultra-high wing loading early supersonic ships like the X-planes, drag chutes for post-touchdown braking assistance.

Flaps to lower stall speed...unless you're flying a delta, in which case they'll be too far back and generate too much pitching effect. So slats instead; they'll raise the maximum AoA before stall, allowing you to reduce the minimum airspeed at which you can maintain level flight.

Basically, when landing, you want to convert your wing into an undercambered airfoil, as those generate the most lift and drag. Like this:

attachment.php?attachmentid=40326&d=1423

With a mid-wing plane, you generate that airfoil shape by lowering flaps at the back. With a delta, you instead raise the AoA of the main wing and lower slats at the front.

You want to cruise in at minimum possible speed and altitude, dropping the nose at the last moment as you touch down. No flare at all if you can avoid it. Get slow and low early (i.e. a couple of km out), lock the nose into your desired pitch for the approach, and control altitude with the throttle.

Spoilers are handy for increasing your ability to control descent rate while keeping the nose stable, but as with flaps you need to be cautious of excess pitching moment if they're distant from CoM.

Edited by Wanderfound

Share this post


Link to post
Share on other sites
  • 0
On 3/16/2017 at 8:20 PM, Wanderfound said:

As with the real ones: slats, flaps and spoilers. And for the ultra-high wing loading early supersonic ships like the X-planes, drag chutes for post-touchdown braking assistance.

Flaps to lower stall speed...unless you're flying a delta, in which case they'll be too far back and generate too much pitching effect. So slats instead; they'll raise the maximum AoA before stall, allowing you to reduce the minimum airspeed at which you can maintain level flight.

Basically, when landing, you want to convert your wing into an undercambered airfoil, as those generate the most lift and drag. Like this:

attachment.php?attachmentid=40326&d=1423

With a mid-wing plane, you generate that airfoil shape by lowering flaps at the back. With a delta, you instead raise the AoA of the main wing and lower slats at the front.

You want to cruise in at minimum possible speed and altitude, dropping the nose at the last moment as you touch down. No flare at all if you can avoid it. Get slow and low early (i.e. a couple of km out), lock the nose into your desired pitch for the approach, and control altitude with the throttle.

Spoilers are handy for increasing your ability to control descent rate while keeping the nose stable, but as with flaps you need to be cautious of excess pitching moment if they're distant from CoM.

I'm trying spoilers ATM.

As far as reducing drag goes, is there any way to implement a supercritical airfoil in KSP with FAR or any other mods?  Also, are there any other ways of increasing the critical mach number of an aircraft that's been designed with FAR in mind?  What about reducing wave drag at supersonic speeds?  At the moment my only knowledge is that you're supposed to use a thin, highly swept wing, use area ruling, and keep everything external to the aicraft within the shock cone.  Are there any other tips?

Regarding wings, is a delta wing better than a trapezoidal/diamond shaped wing for supersonic drag, or is it the other way around?  I'd assume the delta wing is best, since it gets the most wing area inside the most limited span, which helps both with area ruling and keeping the wing inside the shock cone.  But I've heard that the diamond wing is also very good, and the F-104 used it instead of a delta and performed very well.  Is a trapezoidal wing better than a delta?

Share this post


Link to post
Share on other sites
  • 0

I found that a well area-ruled FAR craft slips through the transonic region easily,  but once you get past mach 3 no matter how your craft is designed L/D ratio is going to be awful.    Real life aerodynamics mirrors this to an extent, but there is a way around, using compression lift / building "wave rider" designs.    I doubt that FAR models this however, and have no idea how you'd create such shapes with the parts available to us.

Fortunately the orbital velocity in a non-RSS scaled install is only about mach 7,  and the % of craft weight that gets supported by orbital freefall appears to be determined by what % of orbital velocity your current airspeed is (linear relationship, eg. mach 3.5 = 50%).  This quickly works in your favour.

If all you want to do is go fast in a straight line, it's really easy to make a low transonic drag design with tiny, swept wings.   Max L/D in supersonic flight is at about 15 degree angle of attack,  which means you can get away with tiny stubby wings.  But fighters are expected to turn,  spaceplanes and bombers have to carry fuel or weapons ,   and i prefer my craft to be re-usable and not destroy themselves on landing, but that's just me.

Share this post


Link to post
Share on other sites
  • 0
On ‎3‎/‎14‎/‎2017 at 7:23 PM, KnightRider said:

I'm currently having issues with my latest X-plane stalling at speeds below 120 m/s, which makes landing a pain.

I use drogue 'chutes to assist with landing. This is also what the STS used.

I usually end up with bullets approaching the runway, well above 120 m/s on approach. For instance, this recent Rick Kerman Report demonstrates an approach exceeding 200 m/s and still landing safely. Skip ahead to 8:30 for the approach:

Spoiler

 

A couple of episodes back I showed how to adapt this space plane from stock to FAR, mostly adding tail planes.

Or what about air brakes? Do you have any kind of mechanism to quickly increase drag when you're just above the runway?

An early LTS Sparrow FAR craft used one pair of elevons extended down and the other pair extended up / inverted. At first the air ripped them off the wings, so I reduced the extension amount until it was just enough. It wasn't as effective as the drogues, but it got the job done. And since KSP doesn't check for colliders for parts on the same craft, you might even be able to have overlapping elevons and then have them extend in opposite directions when braking.

Edited by Gordon Fecyk

Share this post


Link to post
Share on other sites
  • 0

Just one remark in regards to area rule: what you should try to minimize is not only the rate at which the cross sectional area changes but the total change as well. This is why real life designs have their fuselages "waisted" to maintain roughly constant cross sectional area (such as the F-106 shown in this thread - notice how the fuselage gets narrower as the wings get wider - this way they could make a regular delta wing without additional wave drag). This can be achieved to an extent in KSP as well by careful placement of air intakes and other radially attached parts.

Edited by m4ti140

Share this post


Link to post
Share on other sites
  • 0
7 hours ago, AeroGav said:

I found that a well area-ruled FAR craft slips through the transonic region easily,  but once you get past mach 3 no matter how your craft is designed L/D ratio is going to be awful.    Real life aerodynamics mirrors this to an extent, but there is a way around, using compression lift / building "wave rider" designs.    I doubt that FAR models this however, and have no idea how you'd create such shapes with the parts available to us.

Fortunately the orbital velocity in a non-RSS scaled install is only about mach 7,  and the % of craft weight that gets supported by orbital freefall appears to be determined by what % of orbital velocity your current airspeed is (linear relationship, eg. mach 3.5 = 50%).  This quickly works in your favour.

If all you want to do is go fast in a straight line, it's really easy to make a low transonic drag design with tiny, swept wings.   Max L/D in supersonic flight is at about 15 degree angle of attack,  which means you can get away with tiny stubby wings.  But fighters are expected to turn,  spaceplanes and bombers have to carry fuel or weapons ,   and i prefer my craft to be re-usable and not destroy themselves on landing, but that's just me.

But I've seen videos of people making mach 5+ aircraft with FAR.  And there are real life designs like the SR-72 and Skylon that don't appear to use wave rider designs and that should be able to achieve speeds around that range (even if they can't reach their design speeds, they'll surely be able to go faster than Mach 3.5).  What design features do they have to allow them to achieve such performance?

4 hours ago, m4ti140 said:

Just one remark in regards to area rule: what you should try to minimize is not only the rate at which the cross sectional area changes but the total change as well. This is why real life designs have their fuselages "waisted" to maintain roughly constant cross sectional area (such as the F-106 shown in this thread - notice how the fuselage gets narrower as the wings get wider - this way they could make a regular delta wing without additional wave drag). This can be achieved to an extent in KSP as well by careful placement of air intakes and other radially attached parts.

With regards to area ruling -- how do I achieve it in KSP?  All the fuselage parts, both stock and in mods like B9, appear to be flat-walled, with no taper that could be used for area ruling.  I suppose I could implement it using procedural parts, but is there any other way?

 

Also, what do you mean by "regular delta wing?" Are other variations of the delta wing better at reducing wave drag?

And do you have any specific tips regarding the placement of intakes to reduce wave drag?  I'd assume I should place them at the narrowest possible point to improve area ruling, but are there any other guidelines?

Share this post


Link to post
Share on other sites
  • 0
57 minutes ago, KnightRider said:

But I've seen videos of people making mach 5+ aircraft with FAR.  And there are real life designs like the SR-72 and Skylon that don't appear to use wave rider designs and that should be able to achieve speeds around that range (even if they can't reach their design speeds, they'll surely be able to go faster than Mach 3.5).  What design features do they have to allow them to achieve such performance?

5 hours ago, m4ti140 said:

I never said you couldn't, just not very efficiently.  But, if your L/D ratio is only 2.5 to 1 at mach 5, but orbital freefall is supporting two thirds of the plane's weight,  drag will only be equal to 28%  of the plane's weight.  

Chasing the Nth degree of wave drag reduction isn't that important, the main thing is to have a reasonably clean design combined with an engine/intake combo whose thrust won't fall off a cliff at the speeds you're designing for.  Most supersonic military aircraft are limited by engine temperature and inlet drag i think ? Variable geometry inlets are expensive.

Share this post


Link to post
Share on other sites
  • 0
36 minutes ago, AeroGav said:

I never said you couldn't, just not very efficiently.  But, if your L/D ratio is only 2.5 to 1 at mach 5, but orbital freefall is supporting two thirds of the plane's weight,  drag will only be equal to 28%  of the plane's weight.  

Chasing the Nth degree of wave drag reduction isn't that important, the main thing is to have a reasonably clean design combined with an engine/intake combo whose thrust won't fall off a cliff at the speeds you're designing for.  Most supersonic military aircraft are limited by engine temperature and inlet drag i think ? Variable geometry inlets are expensive.

Does ksp with FAR accurately model variable geometry intakes in terms of inlet drag?

Also, is there any info out there on whether or not FAR models compression lift?  From what I've seen compression lift can be generated by a caret wing, which is essentially a delta wing at an anhedral angle.

Share this post


Link to post
Share on other sites
  • 0
18 hours ago, KnightRider said:

Does ksp with FAR accurately model variable geometry intakes in terms of inlet drag?

Also, is there any info out there on whether or not FAR models compression lift?  From what I've seen compression lift can be generated by a caret wing, which is essentially a delta wing at an anhedral angle.

I just realised you might be able to make a wave rider shape by making a 2d procedural wing then rotating 90 degrees?

When I last played FAR, it was bundled with Realengines and Realfuels .   I made a plane with two ATAR jets and a ramjet.  At altitude, got up to mach 3 easily enough and was still accelerating strongly, but I had to throttle back because the ATARs started overheating.   The RAMJET still wasn't producing much thrust, never had much luck with them.  The point is though, it  was fairly similar to this one

2016-01-03_00002_zps1jozolcq.jpg

and was quite clean but could have had even better wave drag and been much easier to design if I'd put a smaller wing on it.  But top speed would still have been limited by the turbojet's ability to withstand compression heat.

The above plane (the only FAR craft I can still find pics for) BTW was built with Interstellar installed.  I had a pebble bed reactor powering a dual mode nuclear turbojet.  Gave about 120kn on a pretty flat thrust curve up to mach 3 or so, then switch to close cycle where it operates as a nuclear thermal rocket consuming the hydrazine stored in the wing tanks.  About 85kn and 600ISP I seem to remember. SSTO'd nice and easy on the stock scale system. 

RSS is going to make this very hard.    Even if you get up to mach 5 point something you're only 20% of the way to orbit.  Perhaps a two stage to orbit vessel is the way to go?   Install a stage recovery or flight manager mod?

Share this post


Link to post
Share on other sites
  • 0

Some tips based on my FAR experience:

- Weight is everything, if your landing is at the end of the mission, dump everything (fuel, supplies, monoprop, unwanted kerbals, etc) when you know you'll hit the runway.

- Same weight concern applies during craft build: Set the strength/mass slider for wings and ailerons at around 0.4 (depending on how many g you want to pull, 0.4 is good for 7-8g). Don't overdesign the plane, add lightness.

I was pleasantly surprised at the first landing with this plane, it just wanted to keep flying!

bCwMwfP.png

 

1 person likes this

Share this post


Link to post
Share on other sites
  • 0

20170320181022_1_zps1u09kavy.jpg

Made this today.     It has a really tiny wave drag area despite the huge wing.   It unsticks around 85 m/s at 5 degrees AoA.   The stalling angle is about 40 degrees according to static analysis, but that's irrelevant because i can't pitch up to more than 5 degrees AoA because the front canards stall out.    When finished, I hope to  make this into a Panther/NERV SSTO.

Hopefully I can get more pitch out of the thing, but there's no point going too far with that or you'll just tailstrike anyway.    You get waay more lift in supersonic flight than you do with stock Aero.  At mach 1.7, it flies at 14km at only 1.4 AoA.       According to static analysis, best l/d ratio is at 5 AoA subsonic and 15 AoA supersonic.   Once it's running on nuke engines and i'm no longer trying to get enough air for the jets, i expect it's going to fly really high up !   Heating won't be a problem..

Share this post


Link to post
Share on other sites
  • 0

20170320223850_1_zpsgx8de0nm.jpg

Getting there ! I haven't applied any of Hesp's weight reduction methodologies yet.   But as I'm planning on hanging a pair of NERVAs under the wingtip hardpoints,  I wonder if I should keep them strong?

20170320225115_1_zpslwgqmryu.jpg

1 person likes this

Share this post


Link to post
Share on other sites
  • 0
42 minutes ago, AeroGav said:

Getting there ! I haven't applied any of Hesp's weight reduction methodologies yet.   But as I'm planning on hanging a pair of NERVAs under the wingtip hardpoints,  I wonder if I should keep them strong?

If it helps, I was able to reduce the mass of the LTS Kestrel FC's wings and control surfaces by 50% for the Kerbin 2 challenge. They flexed a bit, but didn't break. Of course I didn't have three tonne LV-Ns attached to those wings. Could you strut whatever the LV-Ns are attached to against the fuselage? Strut connectors can't add that much drag.

I usually fly with DPCR enabled too. This scales the input authority against the strength of the air frame.

And one nice thing about FAR: You don't have to resort to nose cone tricks to work around the stock drag model.

That is one gorgeous looking craft too.

Share this post


Link to post
Share on other sites
  • 0

20170320235057_1_zpsp6sgvl2w.jpg

This thing must have good transonic drag, at 0.97 Mach we were getting L/D of nearly 10 to 1.   Up to MAch 3.5 or so we were getting 4 ish, but this is where you need a wave rider and compression lift - at mach 4 we're down to 2:1.  Fortunately we are most of the way to space.  Adding the 1000 m/s to bring us to orbital velocity used about 1400 dV,  so gravity/drag losses are about 33% from this point.

20170321000057_1_zpsjivg6lee.jpg

I wonder if I could fire those things at a target...

20170321003641_1_zps5ebra0ul.jpg

The NERV engines can be decoupled before landing.  Firing the last stage triggers small parachutes in the nose cones of those NERV pods at the same time as the separators.   The 3 ton NERVs descend at 11 m/s on the little parachutes, but surprisingly they come down intact if they land in the water.     The reduced mass has an effect on our stall speed...

20170321003732_1_zps3wspiqzh.jpg

After floating down the runway and most of the flat plain around the KSC we finally manage to touch down before the hills whack us.  That landing speed is 121 knots in real money.

Share this post


Link to post
Share on other sites
  • 0

Well now,  I tried to make a wave rider to exploit compression lift.

waverider-image2.jpg

That involved a large number of modular wing panels and a considerable amount of time.  Stuck an aerospike on the back, set infinite fuel cheat, and off we went.    

It was rubbish.   At subsonic speeds it is desperately unstable in the roll axis, and has the L/D ratio of a brick.  At mach 4, in the upper atmosphere, it was more stable, and at one point I saw L/D of 1.9 to 1 .  Ahah, this must be where the hypersonic compression lift takes over, i thought.   Mind you, my scimitar-winged ssto was seeing similar numbers at that point, and was a heck of a lot better at slower speeds.     Unfortunately, as speeds increased, L/D fell just as it did with the swept wing aircraft, so I can't say it was better at any point in the flight envelope, no sign of compression lift actually working.    

Whether FAR models this or whether I'm just not as good as NASA at building Hypersonic waveriders (hmm, how many of those have actually flown again?) is the question.

After having a lie down,   I spent the afternoon building an unswept version of my airplane.     Literally just rearranged the pieces to make a straight wing of low aspect ratio (a shade over 2 to 1) of the same area.

Wave drag area doubled to 0.55m,  which still isn't terrible - i think the fuselage layout, with the long nose and engine nozzle appearing just before the wing starts,  with the canard layout keeping the wing mounted far aft,  has inherently good area ruling.

But,  stall speed was about 5-7 m/s lower only.

So , is the reason that supersonic fighter planes land so much faster than say, a Cessna 172, not because of their swept wing planform, but more because 

  • The Cessna is an empty tin can, with a pair of deck chairs to sit on and a ribbon on a stick for instrumentation
  • The fighter jet is loaded with boxes of top secret, obsolete and staggeringly expensive electronics except for the bits that
  • are filled entirely with Kerosene.    Every cubic centimetre of space inside that is not a grey box with wires hanging out the side, is filled entirely with fuel
  • In contrast, the Cessna's fuselage is mostly empty space, except for cobwebs,  dry leaves, and food wrappers
  • The Cessna's fuel tank is the size of a lawn mower's, as is it's engine

So, to make this into a Cessna 172 trainer,  I'd need to swap the cockpit for a service bay and command chair, replace the Panther with a Juno,  reduce the strength of the parts to 5g,  and carry only 50 units of LF.

1 person likes this

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now