1.05 Intakes - Lets figure them out

[Beetlecat]

[quote name='Geschosskopf']I disagree with this. I liked the old display because I could note it at a glance just by having resources open. Right-click menus, especially for jet engines, are just too cluttered up with stuff I don't care about to be useful at all, plus get in the way of enjoying wathcih my plane fly.

The need for less air has been around since 1.0, although not as marked as it is now. So what the IntakeAir display showed was whether or not you'd flame out before your thrust fell off the table. It was a nice check on everything else, simple, quick, intuitive, and out of the way. I wish they'd bring it back.[/QUOTE]

It still exists in the Alternate Resource Panel, but it behaves very stragely now. Not at all like we're used to, and not what you're looking for.

It's better to develop a new display instead -- Nanogagues now has an indicator that shows the % from the engine right-click menu. Possibly assuming the average between engines if it could be different among a group of them. I'm sure TriggerAu will add it to Alternate Resources gauges. Edited November 20, 2015 by Beetlecat

[Red Iron Crown]

The old intake air meter was terrible, the order of placement of parts affected it immensely. Place one small intake and then place an engine and the meter would read near empty most of the time, place the same engine followed by the intake and it would read near full most of the time, in the exact same flight conditions. Good riddance.

[Wanderfound]

[quote name='GoSlash27']The goal now is to have just enough intakes to keep the engines lit[/QUOTE]

Which raises the question: how much is that? We've got numbers for air supply, but nothing yet for air consumption.

[GoSlash27]

Wanderfound,
About .001 m^2 effective intake area per engine is adequate in 1.05. This is roughly 2 of the variable ramp intake (radial mount) per engine at high altitude and hypersonic speed.

Best,
-Slashy

[Wanderfound]

[quote name='GoSlash27']Wanderfound,
About .001 m^2 effective intake area per engine is adequate in 1.05. This is roughly 2 of the variable ramp intake (radial mount) per engine at high altitude and hypersonic speed.[/QUOTE]

Thanks.

It gets complicated, though; I tried running two RAPIERs off a single shock cone. While I suspect it would be fine at speed, it was giving me asymmetic thrust on the runway.

[GoSlash27]

[quote name='Wanderfound']Thanks.

It gets complicated, though; I tried running two RAPIERs off a single shock cone. While I suspect it would be fine at speed, it was giving me asymmetic thrust on the runway.[/QUOTE]

Wanderfound,
Yeah. The shock cone is weak at low speed. It's intake speed is only 5.

Best,
-Slashy

[Morse]

On 17.11.2015, 5:46:51, NathanKell said:

jThe reason the change was made in 1.0.5 was because I quite totally rewrote the intake module, so the info displays for it *do* make sense now, including flow rate.

BTW, speaking of flow rates, is there a way to put the flowrate for intake air into the resource panel? Just to have a general idea of how well the intake performs.

[Psycopyro]

I did some maths for the intakes, don't know if correct but my last tests seems ok.

I used formulas:

R = 8,3144621 J·K^-1·mol^-1

M = 0,028965338 kg·mol^-1

P and T as per graphic

I have the density of air in kg/m3

Then, depending of how many kg/s I need (depending on engines) I can calculate how many intakes I should install.

Intake area in m²

Velocity in m/s

Available air flow (kg/s) = Air density * Intake area * Velocity (in intake), I guess this value is top with the maximum of part (10kg/s for shock cone)

Do not have the numbers there, but the result is for 12x Whiplash, I need 2x pre-coolers, and 3x Shock cone at 25km - 800 m/s.

Now the tests with pictures:

 

I tried to minimise the drag as possible, without tail fin but aerobrakes, but these ones burned during re entry.

300T on runway can deliver in LKO 14400 unit of liquid fuel + oxydiser

No clipping - Stock parts 1.0.5.

 

[Bit Fiddler]

On ‎2015‎-‎11‎-‎11 at 7:13 AM, NathanKell said:

If you have 'enough' intakes you won't get asymmetric flameouts, because the engines will flame out together due to the flow multiplier being too low, a good bit before the air intakes cease being able to shove enough units at you. If you don't have enough intakes you'll hit that problem above.

However, my noting it publicly ought be taken as indication that I'm aware it's a problem.

Ok so this is a very rough flow chart of my understanding of how KSP deals with intake air, I know the details are not going to be correct but just wondering if this is indeed the logic flow KSP uses.

 

 

First we assume we have 4 parts that want "Intake Air"

1 small "Air Sipper 30" jet engine with a max intake need of 3

2 large "Air Hog 100) jet engines with a max intake need of 10 each

1 atmosphere separator with a max intake need of 1

 

so the flow looks like this:

1) by looking at all the air intake parts...  we find we have a total of 6 units of intake air being collected

2) the game then looks at all the parts wanting intake air and divides this air between them evenly, so each part is fed 1.5 units

3) it first looks for buffer overflows and finds the atmosphere separator only needs 1 so it gives the remaining .5 units back to the pool

4) therefore each remaining part will divide 5 units between them and receive 1.25 unit of intake air.

5) next the system looks for "flameouts" and it sees these massive "AirHog 100" engines wanting 10 unit of air but only receiving 1.25 units.  unfortunately this means it will be running at 12.5% efficiency, however it is an air hog after all, so it flames out at 15%. 

6) Therefore BOTH the "AirHog 100" engines flame out and return their air to the pool.

7) now the one remaining "Air Sipper 30" is being fed a total of 5 units of air. so it happily takes it's desired 3 units and runs at 100% efficiency. 

8) the remaining 2 units of intake air are wasted and lost.

 

 

Is this the basic idea or have I missed something?

 

 

 

 

 

[NathanKell]

Nope, it's not at all like that, sorry.

1. The vessel has its set of parts. These are all organized in a list, in the order they appear in the .craft or .sfs file.

2. Each frame, each of those parts is updated, and any modules on those parts run their updates.

That's it.

Now, it happens that during the intake module's update, it checks its speed and angle of attack and the density of the air it's in and computes how much air it intakes. That air is added to the part's resource counter (and if that's full, other parts' resource counters).

And it happens that during the update of each module that uses air that it requests air from the vessel. That air is then consumed. If there's enough air available to match the request, the full output occurs; if there's less, less output occurs, and if the result is under the threshold it flames out / deactivates.

Now, remember that engines' thrust is a function of air density and speed as well. So the higher you are, the less air they will request. And the faster you are, the more air they'll request (broadly; turbofans actually choke around Mach 1). So usually (a) the intakes and engines keep in sync, and (b) the engine will flame out saying 'air combustion failed' (i.e. air density too low to sustain combustion) long before the intakes stop supplying sufficient air.

However, if you don't have enough intakes, engines will flame out from "intake air deprived", and _that_ will happen asymmetrically. Remember the list above: everyone goes in order. So if you're generating half the air your engines need, and you have two engines, the first engine will consume it all and the second engine will have none left to consume. So one engine will run fine and one will be flamed out.

[Bit Fiddler]

ok... not as expected but I think I follow...   so I saw once long ago a mod that would assign an air intake to a specific engine...  all it was doing was re-ordering the parts in the ship file?  so it would place a scoop, then an engine?  and then repeat for the rest of the parts?  is there a fairly simple formula you can plug in numbers and look at in the design phase to see how much air you will need?  or does that require something like FAR to do those calculations for you?

 

I was trying to run as close to stock as possible for memory reasons, and do not have FAR loaded.  so any help other than "crash and burn and redesign and repeat" would be great.  or I guess if I fly up to where my engines flame out and then look at the % of air used on the tool tip if it is not 100% I have too many scoops?

 

 

 

 

 

[NathanKell]

There isn't much stock UI info available to help you do that, you'd need some info mods to help. Not even sure MJ can predict it. It's not a simple equation anyway--it's the interaction of two curves (one based off static density and one based off mach) to know what flow level your engine's running at, and another curve (on the intake) based off mach to know how much air it's generating, let alone all the linear stuff (velocity and density based for intakes, and base flow * throttle based for the engine).

[Val]

My rules of thumb for intakes are:

• Shock Cone: 3-4 engines per intake.
• Adjustable Ramp (Radial): 2 intakes per engine.
• Everything else: 1 intake per engine.

Note: I don't have any good rules of thumb for mixing Juno engine and Small Circular Intake with other sizes.

[kot_behemot]

On 23.12.2015 at 6:56 PM, Psycopyro said:

Do not have the numbers there, but the result is for 12x Whiplash, I need 2x pre-coolers, and 3x Shock cone at 25km - 800 m/s.

Now the tests with pictures:

I tried to minimise the drag as possible, without tail fin but aerobrakes, but these ones burned during re entry.

300T on runway can deliver in LKO 14400 unit of liquid fuel + oxydiser

No clipping - Stock parts 1.0.5.

 

Impressive, but how did you manage to push that behemoth into LKO? With what engines?

[bewing]

On 11/12/2015 at 2:40 PM, GoSlash27 said:

Right,

After fiddling with the numbers and empirical testing, I have come up with a recommended list of engine/ intake combos for minimum drag.

Tech level 6: Panther, 1 Nacelle per, and Tailcone A as required.

Tech level 7: Panther, 2 variable ramp inlet (radial mount) per

Tech level 8: Whiplash, 1 Precooler, and Tailcone A as required.

Tech level 9: RAPIER, 1 Precooler, and Tailcone A (better heat dissipation and lower drag) as required *or* 1 Shock cone inlet (better heat tolerance and lighter weight).

Best,

-Slashy

Not that I doubt all of this academic theorizing, but I doubt all of this academic theorizing. Maybe things have changed a little in 1.1; but if I make my spaceplane with the Tailcone A and Precooler (and any engine!) and measure my fuel in LKO -- then replace the Tailcone and Precooler with a shock cone -- I get much better fuel numbers (and orbital deltaV) with the shock cone.

[GoSlash27]

1 hour ago, bewing said:

Not that I doubt all of this academic theorizing, but I doubt all of this academic theorizing. Maybe things have changed a little in 1.1; but if I make my spaceplane with the Tailcone A and Precooler (and any engine!) and measure my fuel in LKO -- then replace the Tailcone and Precooler with a shock cone -- I get much better fuel numbers (and orbital deltaV) with the shock cone.

bewing,

 It might be that 1.1 has changed things, or it could be that your design relies on thrust instead of minimal drag. I really couldn't say without pics.

Best,
-Slashy

[Psycopyro]

On 03/05/2016 at 8:34 PM, kot_behemot said:

Impressive, but how did you manage to push that behemoth into LKO? With what engines?

12x Whiplash and 2x Vector

That's works but it's a really pain to fly and to land such plane, I prefer my reusable engine for cost reduction

http://forum.kerbalspaceprogram.com/index.php?/topic/105145-stock-payload-fraction-challenge-105-edition/&page=26

 

 

[AeroGav]

On 04/05/2016 at 10:08 AM, bewing said:

Not that I doubt all of this academic theorizing, but I doubt all of this academic theorizing. Maybe things have changed a little in 1.1; but if I make my spaceplane with the Tailcone A and Precooler (and any engine!) and measure my fuel in LKO -- then replace the Tailcone and Precooler with a shock cone -- I get much better fuel numbers (and orbital deltaV) with the shock cone.

Tailcone A is no longer the lowest drag part you can start and end a stack with.  Did some tests at 250 m/s and 500m,  the http://wiki.kerbalspaceprogram.com/wiki/Advanced_Nose_Cone_-_Type_A    

shows lower drag numbers in the right click menu with debug turned on.  It is also lighter and more compact than the Tail cone, justifying its higher position in the tech tree.   BTW the slanted version of the above part performs just as well, it's a purely cosmetic thing.

I could not detect any difference in drag between the Shock Cone and the Pre Cooler however, so unless his engine is surging from lack of air i don't see this affecting him.

One thing Bewing, you mentioned you like to lock Stability Assist on after takeoff and let that control your launch profile - the tail connector part generates significant amounts of lift and it's right at the front, so it's possible it might alter your flight profile to something less efficient?

I find other factors determining my shock cone/pre-cooler choice. The pre cooler is quite a long but not very heavy part and i find it useful to put just ahead of CG in the stack to shift the heavy cockpit etc. further forward, where it's got more lever arm to balance the mass of the engine.

[Pecan]

4 hours ago, Psycopyro said:

12x Whiplash and 2x Vector

That's works but it's a really pain to fly and to land such plane, I prefer my reusable engine for cost reduction

http://forum.kerbalspaceprogram.com/index.php?/topic/105145-stock-payload-fraction-challenge-105-edition/&page=26

 

 

Is this your considered, version 1.0.5, analysis or are you necroposting?

[Psycopyro]

Actually don't know, I play since 1 year and physics change 3 times... I will try it.

Edit:

True on 1.0.5. I cannot even take off on 1.1.2. have to re-design that...

Edited May 27, 2016 by Psycopyro

[Northstar1989]

On 11/16/2015 at 10:15 PM, GoSlash27 said:

royying said:

 

Is intake spamming still useful now?

No, it's not. Intake spamming is counterproductive from 1.0 on. All it does is add drag in exchange for air that you can't use. If engines are fed beyond their useful speed and altitude, they will produce nothing but noise, heat, and *negative* thrust at the expense of fuel.

The goal now is to have just enough intakes to keep the engines lit while they are still useful while adding as little drag as possible. Precoolers are the new hotness in 1.05.

Best,
-Slashy

I've said this before, but this system is NOT reflective of how real jet engines work, and I think that's a problem.

Real jets are completely unaware of outside speed and altitude.  All they know are the characteristics of the intake air stream received (its temperature, speed, and oxygen content) and the ambient pressure outside the exhaust nozzle (just like with closed cycle rockets, ambient pressure can compress the exhaust stream, reducing thrust.  Just like with a rocket, it is also possible to expand it to a lower pressure and higher speed in the nozzle, increasing thrust when the exhaust pressure exceeds the ambient pressure.  Generally, you get the least thrust for a given exhaust stream at sea-level...)

This means that, *in theory* intake spamming *should* work with open cycle jet engines (note that technically, rockets are a type of jet engine, hence the open/closed cycle terminology).  The *problem* is that you would need to reduce the airstream to the same internal speed without exceeding temperature tolerances.  When you compress the airstream to increase its pressure to a usable level, you increase its temperature.  When you slow it down, you compress it.  This is a problem as intake compressors will overheat and melt if you compress the airstream too much...

This is where precoolers could, theoretically, help in real life.  Although no precooler has ever been attached to a working jet engine on an aircraft that has flown in real life (this is why the speed and altitude curves are a useful approximation of the performance of a jet engine in KSP as it currently stands- because they reflect the performance limits of the jet's compressors and engines themselves when you don't pre-cool the airstream...), the British *ARE* developing a highly potent precooler system for the SABRE engine (off which the KSP RAPIER is likely inspired), so it's not so far-fetched that KSP has one.

In real life, a precooler would work by reducing the temperature of the airstream before it reaches the jet engines, allowing you to slow it down and compress it further.  With a powerful enough precooler system, you could theoretically slow down and compress an airstream from *any* speed and pressure to one usable for even the most low-speed of jet engines without loss of thrust or efficiency.  Of course, the cost is the mass of the precooler and the need for a heat-sink for all the heat you remove from the airstream...

This is *NOT* how precoolers work in stock.  In stock they're basically just fancy inline air intakes with no effect on thrust or fuel-efficiency.  KSP-Interstellar Extended, however, *USED TO* include precooler code that allowed the stock precoolers (as well as precooler parts unique to the mod) to act much like they would in real life.  That is, when attached to an air intake and an engine, they would fool the engine into thinking it was operating at a lower speed/altitude than it actually was.  This allowed the mod's Thermal Turbojets to operate up to much higher speeds and altitudes while still producing usable thrust.  This also helped prevent jet engines from exploding due to compression-based overheating (KSP-I included code entirely separate from the stock heat system that added heat to engines purely based on the speed and altitude their intakes were operating in).  This was especially relevant when the mod included powerful next-gen fission reactors, fusion reactors, and even microwave beamed power networks that could all easily propel a spaceplane to Mach 5 or 6 with a sufficiently precooled intake airstream and well-designed plane, at least in FAR.

You'd lose most of your Thrust as you reached Mach 6 in KSP-I, because one precooler was as good as 12 (in real life, you could theoretically just alternate coolers and compressors to reach any desired pressure and airspeed with enough precoolers and compressors).

This technology was invaluable for allowing things like my construction of the first-ever working HTHL spaceplane in RSS 64K without cheats (OK, to my knowledge- somebody else might have done it earlier and never told anyone- although another player posted one a month or two later and thought they were first, as I didn't really document it much on the forums...  I did have dated Imgur albums however...), the last in which my seies of designs could have even worked in full scale RSS (I built three successive working models in a six-plane series with countless intermediate prototypes between these reference designs- the 3rd was the first suborbital, the 4th was orbital but crashed often, the 5th stable, and the 6th had a higher payload capacity and could have even worked in RSS full-scale with a smaller payload).  Note that I limited myself strictly to the lower tiers of fission reactors for this, as I considered fusion and antimatter reactors a tad unrealistic...

Anyways, to conclude this very long tangent and get back on-topic, the stock precoolers don't work like this at all.  Someone at SQUAD really should consider rewriting them to function more like this- as in real life precoolers aren't air intakes at all, and function mainly to (in combination with compressors- which KSP seems to abstract as part of the engine parts, based on their mass vs. the intakes...) allow jet engines to produce usable thrust at higher speeds and altitudes than they would otherwise work at.

Regards,

Northstar

Edited September 9, 2016 by Northstar1989

[kspacc]

This is the same order as you would expect based on the respective intake areas.

But, it's only at sea level.

Wouldn't it be more helpful to have a set of similar charts showing these tests at various altitudes?

Or am I missing something?