Fuel System Challenge

[Hotel26]

Good day, all.

I am facing a challenge with the fuel feed system and am hoping to receive some expert advice!

This link is to an unmodded .craft file for KSP 1.3.1.  The craft is the B payload for Operation Beach Head loaded atop a standard Aquila launcher/transit vehicle.

 

Problem Statement:

The lowest 3 engines ignite during the lift-off stage and are intended to burn out together.  To make this work in KSP versions 1.0.5 (which it did), fuel lines had to be run from the lower tanks to the next-level higher *engines*, rather than the tanks above.  (This was due to one of the peculiarities/flaws of the original KSP fuel feed system.)

Since 1.2.1, 1 tank on the lower deck runs out before the other two.  I investigated this in 1.3.1 with the fuel feed overlay system and verified that the fuel lines are all correctly attached and everything seems fine.  In addition, I can see no way with the new fuel priority system to prioritize this any differently.

Any advice, insight, expertise will be greatly appreciated and acknowledged!

[And yes, I did promise that Aquila is the ugliest ship in the Kerbal universe.  If anyone has created an uglier Gorgon, I'm sure we'd all enjoy seeing it!!]

Edited February 4, 2018 by Hotel26

[Foxster]

You just need to set the Fuel Delivery Priority (available by setting Advanced Tweakables on) to the same value for all the tanks in the first stage. 

You can also then remove all the fuel ducts around that stage and up to the next. 

If you want to save a shed-load of dV with this craft then consider streamlining the tops of the tanks. You may also want to consider adding mirrored parts with symmetrical placement and also not using those quad-couplers which are stupidly draggy. 

[Hotel26]

20 minutes ago, Foxster said:

You just need to set the Fuel Delivery Priority (available by setting Advanced Tweakables on) to the same value for all the tanks in the first stage. 

I''ll give it a try right now.  I'm sure I didn't understand the fuel priority system very well!

Can I really lose the fuel lines?  At lift-off, 6 engines are firing and drawing fuel off the lower 3 tanks ONLY so that when the first stage is jettisoned, all tanks are still full...  i.e. asparagus staging.

More about the craft.  It's primarily a booster to take payload into low orbit.  But it may then it may get refueled and ganged together for interplanetary transit.  Starting in space with full fuel, drag does not matter, only mass?  Also, what are "quad-couplers"?

Edited February 4, 2018 by Hotel26

[Foxster]

To be brutally honest - it's a mess. 

There is  broken symmetry, nuke engines apparently fed from the old LF/Ox requirements mix, pipes all over the place, unneeded thermal systems, loads of drag, loads of RCS...

Personally I'd start again. 

If you let me know just what you are trying to put into space (the two rovers?) then I'll design a craft for you. 

[Hotel26]

12 minutes ago, Foxster said:

what you are trying to put into space

That's the rub.  Anything.  Load it onto the payload platform and Go.

The honesty is appreciated!  It really needs refactoring.  Mainly because the second stage works as a piston and it's always been tricky to get it to slide out past the couplers on the bottom of the outer tanks in the uppermost stage.  I fixed that recently because I know how to use the gizmos now(!).  It'll get rebuilt shortly but I wanted to fix this second major problem first.

About to test your priority suggestion...

P.S. the thermal; systems, for example, are for long burns of the nukes for interplanetary transit insertions...

Edited February 4, 2018 by Hotel26

[Foxster]

Ah, right. I think I see what you are trying to do. 

What you have got yourself into is that the craft has been rebuilt umpteen times and it's kinda lost its way. Sometimes you just have start again from scratch. Simplify things as much as possible. 

Edited February 4, 2018 by Foxster

[Hotel26]

Nope.

I set the fuel priority on the 3 tanks in the middle layer from 40 down to 50 like the bottom tanks.

Asymmetric draw.

OK, so then I took the fuel lines out...

Still asymmetric draw on the bottom 3 but with one emptying faster rather than remaining fuller.  In addition, now the tanks above are draining, which means it's not fully operational as asparagus.

Six engines firing at all times; fuel for all only coming from the next tanks to stage away.

And tho I said there was asymmetric draw at the bottom, all six stopped together which means that there was a counter-balancing asymmetry in the upper level.

[Foxster]

As I said, a chunk of your problem is that the tank/engine stacks are not fitted with symmetry turned on, meaning that you have to mess with each stack separately. The craft has become unnecessarily complicated. 

[Hotel26]

7 minutes ago, Foxster said:

he craft has been rebuilt umpteen times

This is Aquila VII (seventh generation), so you are right.

But never rebuilt from scratch because it was too complicated internally and I didn't trust myself.  I'm now a bit better at mapping craft pulling a test instance apart carefully and I wrote a Parts program that prints the part tree.

But it has taken myself until now to steel myself for the effort.

To some extent, I think you make a map, put a sweatband on, and then simply start building it over and over like a Marine stripping down and rebuilding his weapon against a stopwatch until he can do it in three minutes; he can do it in the dark; he can do it while asleep; and he can do it while dead drunk and being shot at...  but that's just my imagination. 

 

Edited February 4, 2018 by Hotel26

[Foxster]

[Hotel26]

Also, one thing that always hampered me in the VAB is when you lay down six-way symmetry and then want to put 2-way or 3-way on top of that.  The VAB corrects you to N-way as it already is unless you say 1-way.

This explains part of what you may be seeing.  especially with this crazy design.

[Foxster]

What you have to do for that is create one of what you want first, putting the 2/3 way on that and then take the whole thing off and place it again with 6 way. 

[Hotel26]

What burns me is that this used to work!  And I had read the paper on the internals of the early fuel system (I believe it was on the bug-tracker system actually!) and had steered around it.

So with Aquila, due to the "nexus tank problem" (A and B draw from C but C draws from D and E, but will do so asymmetrically [due to a lazy short-cut Squad deliberately took that actually killed some interesting possibilities for me more than once]!) and because I needed to be able to feed the top six tanks for the nukes so they didn't cut out at a bad time, I've always had to pump fuel from the last orange central tank into the outer 3 manually and I don't mind that.  Because connecting the upper stage with asparagus would create that 'nexus'.  (I actually pump it back out on days with a light load when periapsis is reaching 0km alt and the upper three still have fuel, which means its time to jettison the final stage so that it doesn't litter orbit.)

Fundamentally, what I am doing is simple asparagus staging.  It should be feasible.  And it used to work... smoothly.

17 minutes ago, Foxster said:

What you have to do for that is create one of what you want first, putting the 2/3 way on that and then take the whole thing off and place it again with 6 way. 

OK.  Think about the scenario where the payload platform has 6 peripheral tanks and I wish to place the load ABABAB to keep the mass distribution symmetrical.  It's going to end up as 6 individual placements.  That's OK.  Only discrepancy would be with rotations not being exactly the same, for example.

[Hotel26]

This is F-Troop Gizmo.

I'll be putting nukes and lots of solar into it so that it can take some ion-engine packs in the stern.  Plenty of packing space on top with lots of margin for wide loads.

Mammoth followed by Mainsail followed by NERVs followed by thousands of Dawns.

I'm probably going to double the height on the orange tanks so that I can get enough solar panels down those columns to drive the xenon.

It's 9 launches in one, which cuts down on Mission Control costs.

It's dedicated to the folks at Squad for bringing us KSP. 

[Good bye, Aquila.  Requiescat in pace.]

Edited February 4, 2018 by Hotel26

[Hotel26]

Loaded with *all* of the Operation Beach Head payload, this doohickey arrives in 90km orbit with LF:18073, OX:20505 in hand.

• 2 Gimlet drill rigs
• 2 Hawk tug/landers
• 1 Hawk M700 surveyor
• 2 fuel pods
• 4 Habitat (2 in orbit; 2 on the surface)

That means I don't need my Minotaur super tanker any more, maybe?  And that brings my fleet down from two workhorses to just one.

Not yet loaded with nukes for interplanetary transit; nor the Omega ion-drives.

And hey, it triples as an Orbital Fuel Dump.

Edited February 4, 2018 by Hotel26

[Rocket In My Pocket]

1 hour ago, Hotel26 said:

It's 9 launches in one, which cuts down on Mission Control costs.

Me being initially surprised/doubtful and then realizing it's actually a good idea:

Edited February 4, 2018 by Rocket In My Pocket

[Hotel26]

On 2018/02/04 at 10:49 PM, Rocket In My Pocket said:

Me being initially surprised/doubtful and then realizing it's actually a good idea:

I'm excited!

It's a:

• heavy lifter [LKO]
• it's a super tanker
• it's an orbital fuel dump/habitat
• it's an interplanetary transit vehicle
• and it's an expeditionary force

and my mother (83) could fly it, it's so simple & sweet.

I have to put a lot more stuff on it.  Mainly though nukes and I won't get enough solar sails on it even 2 orange tanks tall, so I will use my motorized solar sails (16 x Gigantor) that stow for take-off/aero-braking but deploy for ion flight.

The details are what will take the most effort.  So far this has been 45 minutes.

Think about this: the computer hardware business went componental decades ago and use "blocks".  The software industry use packages and libraries: reusable software.  Aerospace??

Aquila was built during the 0.29 era.  Now I *know* about gizmos...    (The monkey bars in the structure are purely ornamental!!  I got taught this in the SPH by my masters.  )

Edited February 6, 2018 by Hotel26

[Hotel26]

One last pic -- Gizmo on the pad:

At the moment, it's just 2 stages.  BANG!  Bang.  Stop.

[monstah]

2 hours ago, Hotel26 said:

It's 9 launches in one, which cuts down on Mission Control costs.

It took me until the very last picture to understand what you were doing, and man this is genius!

[Hotel26]

23 minutes ago, monstah said:

It took me until the very last picture to understand what you were doing, and man this is genius!

*blush*  thank you, Sir.

It'll be going up in KerbalX when I'm done!

Edited February 4, 2018 by Hotel26

[foamyesque]

I have a question though: Why no nosecones? You can place them directly on those docking ports and rig an action group to decouple them all at once, probably during a coast-to-apsis phase to reduce debris & keep them from being jammed against the rocket from thrust. You could probably also improve your aerodynamics by putting adapters on top of those big Kerbodyne tanks. There's nothing to be done about the girders except reducing the count, but you might want to investigate doing that as well, since it looks like you only need one girder's worth of separation.

[Flavio hc16]

i'm for streamlining that monstrosity too! ( the 9in1 launch). Put some nosecones and adapters, their weight will be greatly paid back with reduced drag ( and also a bit more fuel, that is always good)

[Hotel26]

6 hours ago, foamyesque said:

Why no nosecones?

Thank you for the suggestions!!

This was literally a 45-minute build from conception to completion of first trial.  (And I confess, the exercise started as a joke borne out of some frustration.)  And I also know there will be weeks of details to refine this.

Under the fairings are the Gimlet drill rigs (rhymes with "pigs" if you saw my video...) and they have 4x medium thermal radiator panels.  That get damaged when the fairings blow off.  So the fairings are going to have to go.  Drag, drag, drag.

When I built Aquila, I didn't know about gizmos (around about 0.29) and I've never had enough spacing on the platform deck.  So here, I struck out and went bold!  Who knows how big I thing I might want to include in the future.  It could be a Pterodactyl on vertical launch, for example.  However, you are probably right that I could pull it in.

As for the girders, here's a little secret: they are purely ornamental.    It's the gizmos that are holding this craft together plus a judicious tincture of auto-strutting.

2 hours ago, Flavio hc16 said:

also a bit more fuel, that is always good

Very good thinking.  Whatever the payload doesn't consume for ascent goes into the orbital bank as fuel credits for topping off the next flight!

[Hotel26]

When I was building Aquila, I found that if you build a large payload platform and then load it up, you are going to have to provide a commensurate amount of thrust from below to power it.

The result is that, upon ignition, the central column crumples like an origami waterbird.  So Aquila utilizes 4 thrust points acting on the platform.  Under nukes alone, it uses 3, but the thrust (and drag!) is greatly reduced, too, once in space.  Actually, that's not fully true: on launch, I stage the nukes in at about 30km and that means that thrust is applied to all 7 points on the platform.

This detail is easily to overlook while first admiring the breath-taking ugliness of the machine.  Unfortunately.

This is the idea that I carried over into Gizmo.

Secondly, it's not easy in KSP to run a fuel line in other than a straight line.  You cannot connect it from a tank to one above bypassing an engine in between.  This is why I had the staggered tank design, like a church organ.  It allowed me to connect A diagonally to B and then diagonally back to C above.

Aquila suffered in execution because I lacked skill / expertise in the VAB and because I patched the design through 7 generations without a rebuild from scratch.

I didn't know for example that you can access internals hidden in the core by pulling things out with the gizmo, temporarily, to get a better view.  Therefore, for example, I could never fix the 6-way placement of the RCS when I only needed 4-way.

Recently, some SPH masters tutored me, which has advanced my understanding and probably my capabilities for the future.  I'll name @swjr-swis specifically, but there was a round also with my very first space plane in August 2016; thanks to all then as well.

Nevertheless, for all that, I poured as much ingenuity and love as I could into Aquila and I will mourn it.  I'm not done on the fuel problem.  My next step will be to run it up in 1.0.5 and verify it worked properly.  I will also need to check that I don't have any stray fuel lines buried inside the core that I've forgotten about.

Then I will make a launchpad test bed.  It will have 3x small engines running from 3 tanks, A, B, and C, and with A feeding B and B feeding C.  This is basic asparagus staging and it should work.  At start, all engines should run off fuel in A and consume it all before touching a drop out of B and C.  A's engine will quit, fuel starvation, as expected.

If I can get past the fuel problem, I will rebuild this ancient bird...  it has what Gizmo will never have: personality.

The only way to begin appreciation of it is to fly it.

Edited February 5, 2018 by Hotel26

[Hotel26]

I made this fuel feed test bed and experimented with it.

From this, I gained some understanding of how to use the fuel priority system  From left-to-right, the fuel priorities are 0, -1, -2.

The test bed works.

I applied that to Aquila...  Specifically, I set all three lowest-level tanks to the highest priority.  Note that there are 3 independent fuel flows being asparagized in Aquiila.  The fuel priorities now decrease as you ascend the stages.

And...

 

Launch of overhauled Aquila failed.  Asymmetric draw.

Very glum.

Edited February 6, 2018 by Hotel26