WANTED: Mothership Propulsion Design

[Wcmille]

Looking for a propulsion design that fits these specifications:

• Mothership outbound payload is 80 tons (includes crew, docking ports, RCS, solar panels, batteries, etc.)
• Mothership return payload is 16.7 tons. The other weight is left at the destination.
• Total outbound burn time must be 15 minutes or less, 4000 dV. TWR does not matter if the burn time for 4000 dV is under 15 min. Please note that estimating outbound burn time as half the total burn time often gives incorrect results.
• Total return burn time must be 15 minutes or less, 4000 dV (beyond the outbound dV).
• Assume no refueling occurs at the destination. Assume fully fueled on initial departure.
• Propulsion must be nuclear or chemical (no ion drive); assume a complete stock tech tree.
• Cost doesn't matter.
• Once in orbit, design must be 100% reusable. Boosters/Stages are OK only if they can be recovered and reattached. Boosters that can fly home are a huge plus.
• How it gets to orbit doesn't matter; this doesn't need to be part of the design.

The best design fits the above parameters and optimizes for the least total fuel mass.

A design using 40 NERVs and 383 tons of Mk III Liquid Fuel tanks will solve this problem, but I believe better designs are possible by using boosters and mixed engine types (e.g. using NERVs, Rhinos, and Poodles together)

[Rune]

1 hour ago, Wcmille said:

Looking for a propulsion design that fits these specifications:

• Mothership outbound payload is 80 tons (includes crew, docking ports, RCS, solar panels, batteries, etc.)
• Mothership return payload is 16.7 tons. The other weight is left at the destination.
• Total outbound burn time must be 15 minutes or less, 4000 dV. TWR does not matter if the burn time for 4000 dV is under 15 min. Please note that estimating outbound burn time as half the total burn time often gives incorrect results.
• Total return burn time must be 15 minutes or less, 4000 dV (beyond the outbound dV).
• Assume no refueling occurs at the destination. Assume fully fueled on initial departure.
• Propulsion must be nuclear or chemical (no ion drive); assume a complete stock tech tree.
• Cost doesn't matter.
• Once in orbit, design must be 100% reusable. Boosters/Stages are OK only if they can be recovered and reattached. Boosters that can fly home are a huge plus.
• How it gets to orbit doesn't matter; this doesn't need to be part of the design.

The best design fits the above parameters and optimizes for the least total fuel mass.

A design using 40 NERVs and 383 tons of Mk III Liquid Fuel tanks will solve this problem, but I believe better designs are possible by using boosters and mixed engine types (e.g. using NERVs, Rhinos, and Poodles together)

*whistles*

...You don't ask for much, do you? 8km/s on a single stage, that's mass ratio ~3 on NERVA's Isp. So yeah, you are definitely asking for a very high TWR, because you are going to have to burn through 160mT of liquid fuel in 15minutes... assuming your engines are massless.

 

Rune. My suggestion? Drop the burn time requirement.

[sgt_flyer]

yeh, i tried to test out configurations for the burn time requirements (with only 16.7 tons of return payload after a 81 tons outbound burn), not accounting for additionnal superstructure needed to attach the engine pods - and i need around 640 tons of fuel on 46 nervas for single stage with the given parameters. it would burn 2/3rd of the fuel for the outbound burn, and the 1/3rd remaining for the return burn - to stay on a single stage.

besides, given the weight of the thing, going on anything else than nuclear is going to need an awful lot more of fuel.

Edited December 30, 2015 by sgt_flyer

[Temstar]

4000m/s delta-V is a lot, I'm guessing this is headed for Moho?

5 hours ago, Wcmille said:

• Once in orbit, design must be 100% reusable. Boosters/Stages are OK only if they can be recovered and reattached. Boosters that can fly home are a huge plus.

What does this mean? Does it mean the launch vehicle for the mothership also has to be fully reusable? Or are you saying that just the mothership itself has to be fully reusable (staging or otherwise)?

One suggestion I have, you won't happen to have any D or E class asteroid on your radar near by do you? Go see if you can find one with high ore concentration. People have found things like this:


That's basically a better fuel tank than actual fuel tanks.

Edited December 30, 2015 by Temstar

[Rune]

There is a way to do this, tough, now that I think about it further. Even in a reusable fashion: break the back of the rocket equation, as usual. If you only had to go 4,000m/s, that could be done with beefy chemical propulsion... so use ISRU. Then it's 4km/s, refuel, 4km/s. IMLKO is around the same as a single-stage nuclear bird that does the round trip, but the TWR is about twice.

 

Rune. For a 80mT payload, that would still be a ~500mT ship at LKO. VERY roughly (50% of dry weight as pure payload, the rest engines+tankage, 340s Isp).

[Wcmille]

Temstar, I really like this idea. I think it essentially requires you to only pack 4000dV, then you refuel during the trip. I'd have to think about narrow the asteroid weight range and resource concentration would need to be.

Rune, this is quite possible; I have an abstract solution already. I'm trying to find the optimal answer.

34 Nukes and 395 tons of fuel MkIII tanks (8:1 wet:dry ratio). Using the first 265 tons of fuel tanks (about 232 tons of fuel) gets you 14.88 minutes of outbound burn and 4033 dV. Using the remaining 130 gets you 4056 dV with 7.30 minutes of burn on the flight home. Could be I've done the math wrong; very easy to get a sign or parenthesis wrong in a spreadsheet.

I think the design could be improved by using a Rhino that fires briefly (for the first couple minutes) with the nukes on the outbound journey. This reduces the number of nukes and turns them into fuel.

3 hours ago, Rune said:

Then it's 4km/s, refuel, 4km/s. IMLKO is around the same as a single-stage nuclear bird that does the round trip, but the TWR is about twice.

What is IMLKO?

[sgt_flyer]

one problem though, is you can't 'cut' fuel tanks  

 

closest possible to your numbers is 7x long MKIII fuel tanks, + 5 nervas / engine. (and you'll still need a superstructure to connect to your payload, bipropellant RCS  fuel to steer that monster - unless you account for those within your payload's weight)

[Rune]

1 hour ago, Wcmille said:

Temstar, I really like this idea. I think it essentially requires you to only pack 4000dV, then you refuel during the trip. I'd have to think about narrow the asteroid weight range and resource concentration would need to be.

Rune, this is quite possible; I have an abstract solution already. I'm trying to find the optimal answer.

34 Nukes and 395 tons of fuel MkIII tanks (8:1 wet:dry ratio). Using the first 265 tons of fuel tanks (about 232 tons of fuel) gets you 14.88 minutes of outbound burn and 4033 dV. Using the remaining 130 gets you 4056 dV with 7.30 minutes of burn on the flight home. Could be I've done the math wrong; very easy to get a sign or parenthesis wrong in a spreadsheet.

I think the design could be improved by using a Rhino that fires briefly (for the first couple minutes) with the nukes on the outbound journey. This reduces the number of nukes and turns them into fuel.

What is IMLKO?

Initial Mass at Low Kerbin Orbit. The kerbal equivalent of IMLEO, basically. It's a useful figure of merit for a mission architecture, very comparable. And yeah, it might be possible (I haven't run the numbers myself, but it works for a 1mT payload/nuke in the editor)... but lag due to the part count is going to be horrendous. A low-thrust system will likely be much smaller, allowing you to run it at high simulation speeds. For example, meet the Medusa:

80mT dry, 250mT with all the extras lander included, and makes the same 8km/s... but with a mere 0.06G's of acceleration. Currently targeting 200 parts. But of course I am a known minimalistic that doesn't care about setting the autopilot and an alarm, and going to get a drink and grab a book.

 

Rune. YMMV!

[Wcmille]

11 minutes ago, Rune said:

Initial Mass at Low Kerbin Orbit. The kerbal equivalent of IMLEO, basically. It's a useful figure of merit for a mission architecture, very comparable. And yeah, it might be possible (I haven't run the numbers myself, but it works for a 1mT payload/nuke in the editor)... but lag due to the part count is going to be horrendous. A low-thrust system will likely be much smaller, allowing you to run it at high simulation speeds.

You're very right about the part count; I will have to consider that. My concern with longer and longer burns is that the error increases until you are pushed into a higher orbit to stay accurate. For high dV burns, this may be well above the gate orbit height, which is the same as losing dV. For the design you have above, how long is the outbound burn?

[Wcmille]

39 minutes ago, sgt_flyer said:

one problem though, is you can't 'cut' fuel tanks  

 

closest possible to your numbers is 7x long MKIII fuel tanks, + 5 nervas / engine. (and you'll still need a superstructure to connect to your payload, bipropellant RCS  fuel to steer that monster - unless you account for those within your payload's weight)

Yes, the 16.7 is all the reaction wheels, RCS, crew quarters, docking ports, etc. I don't think there's much need to use RCS, provided it has reaction wheels. It does not dock with other things. It just needs to hold still to let other things dock with it. It also doesn't need to turn particularly fast (provided it will hold still during a burn), since it has several days between burns to face the next burn.

[Rune]

14 minutes ago, Wcmille said:

You're very right about the part count; I will have to consider that. My concern with longer and longer burns is that the error increases until you are pushed into a higher orbit to stay accurate. For high dV burns, this may be well above the gate orbit height, which is the same as losing dV. For the design you have above, how long is the outbound burn?

BurnS. That's the issue. What you need, my friend, is a guide on the art of the long burn. That link right there is "how the pros do it", which is that way because it really makes everything easier. And it's not even that hard to learn! You will feel all NASA-like, planning burns and ejection angles in no time... Once in space, all you need is the push of a feather!!

Of course the other option is perfectly valid, too. With the ISRU parts, chemical ships can grand-tour the system, 4km/s at a time, with stupendous TWR.

 

Rune. Ok, there is a limit under which the game can fail to notice your acceleration, and I've found it. It's something like 0.05m/s², on the high part of a high-altitude elliptic orbit.

Edited December 30, 2015 by Rune

[Wcmille]

This method seems very good for low-velocity hyperbolic orbits, but how do you leave Kerbin if your ejection dV is 2000+? Once you have about 1000 dV put into an orbit, it will start to go hyperbolic.

How do ships with extremely long burn times arrive in a system without wasting DV? If you don't have the TWR for a capture on the first pass, you are not stopping. If you have to start the burn far from the Pe, you are wasting dV. This probably doesn't matter in a Hohmann interplanetary approach, but a high-velocity one-tangent burn requires a lot to stop.

[Jeanjvs]

I don't know if this helps, but i designed the KFV-Odissey, a mothership with ~10.000 m/s delta-v made take a 30 ton payload (science, mostly) to Eeloo and back. It needed two 6-7 minutes burns to make the transfer and it also had a small lander.

I had been designing it since 0.90, but it relied on aerobraking to return from Eeloo to Kerbin. Since i never found a way to protect the NERVAs from the heating, i just gave up and started to design smaller ships with ISRUs. But, if you scale this thing up and solve that problem, i think it could meet your requirements.

[Jakalth]

You could also go with a design like this:

 

The tug has a TWR of 0.55 in Kerbin's gravity meaning it can land on places like Vall, Moho, Dres, and even Duna.  It's got 4,600m/s delta-v by it's self, and can have it's range extended simply be trailing more fuel tanks behind it.  It's a design based on suggestions and tips from Rune.  This little guy may not be able to effectively pull the 64 tons behind it, but if it's scaled up to 6 engines, with it's stance widened a bit, it should be able to handle that much cargo weight easily.

 

Although, this type of craft works best if refueled on site.  But with it's TWR, it can undock from the cargo train, drop down to the surface of the world of choice, preferably a moon, refuels using it's ISRU, fly back up with it's tanks full and refuel the train, only needing to do a few trips to top it all off and go to it's next destination.

[sgt_flyer]

Here's some pictures of the WIP on a nerva based engine pod (note, i still need to tweak the cooling... can manage around 7/8 minutes of burn at full thrust before some of the nerva explodes due to overheating.) - 562 tons currently, propelled by 35 nervas. the pod had approximately enough D/V for the payload requirements without the slanted fuel tanks, but as i needed to add radiators and such, so i needed a bit of additionnal fuel. it might be slightly over 15mn for a 4000m/s of delta-V burn, but as it has it's own RCS fuel, you should be able to limit the RCS on the mothership itself. (needed those RCS pods to test it, as they have a specific hidden staging to prevent the vernors from using the nerva's fuel). the 5 nervas on the core fuel tank would run out of fuel before the outer ones. (though it should possible to add a MK3 to 3.75m adapter)t

he rocket on the side is an upgraded version of my chaos star 500 rocket (and the 500 is the smallest of the series ^^) it can lift the whole pod to LKO in one piece (for only 449 parts at launch, including the 138 parts nuclear engine pod! - and yes - the nuclear pod fully fits within the rocket's fairing)

 

still need to build a docking subassembly for your mothership. (if you have a preference for the diameter for the docking subassembly (so you can strut the subassembly to your mothership) - with the mass of the thing, when doing RCS manoeuvers, it would impart a huge amount of torque on a single 2.5m docking port)

nuclear pod RCS system : 1 large RW, 1 large probe core, 1RTG to power the two, and two rocket fuel pods to feed the vernors. (pitch / yaw and roll possible with the vernors).

 

Edited December 30, 2015 by sgt_flyer

[Wcmille]

2 hours ago, sgt_flyer said:

Here's some pictures of the WIP on a nerva based engine pod (note, i still need to tweak the cooling... can manage around 7/8 minutes of burn at full thrust before some of the nerva explodes due to overheating.) - 562 tons currently, propelled by 35 nervas. the pod had approximately enough D/V for the payload requirements without the slanted fuel tanks, but as i needed to add radiators and such, so i needed a bit of additional fuel.

You raise a really good point here -- how does one calculate precisely the number and type of radiators?

[sgt_flyer]

it's even more complicated than that - you also need the correct 'intermediate' part that can transfer the heat - cubic struts are really bad at transerring heat back to the fuel tank (as the fuel tank can absorb enormous amounts of heat - if i could transfer this damn heat through the intermediate part. (cubic struts are bad for this). it'll require a bit of experimentation though -the additionnal thermal data avaible through the menus is difficult to read. and i think the sunlight, by increasing skin temp, can also affect the efficiency. (afterwards, it depends on where you want to go )

[Requia]

6 hours ago, Wcmille said:

This method seems very good for low-velocity hyperbolic orbits, but how do you leave Kerbin if your ejection dV is 2000+? Once you have about 1000 dV put into an orbit, it will start to go hyperbolic.

How do ships with extremely long burn times arrive in a system without wasting DV? If you don't have the TWR for a capture on the first pass, you are not stopping. If you have to start the burn far from the Pe, you are wasting dV. This probably doesn't matter in a Hohmann interplanetary approach, but a high-velocity one-tangent burn requires a lot to stop.

Fall into the gravity well, instead of setting a low PE set a high one then slow down enough well before it that your PE naturally drops, then make a normal capture brake at PE. It's a little inefficient but it worked out fine for me on numerous probes and one huge ship that's similar to what you're flying. Also similar to how real low thrust spacecraft pull it off. For the departure burn its the same thing, you have to eat some inefficiency by making a long departure burn, but the accuracy wasn't too much of a problem (first part of burn was angled between the horizon and node, then finished the burn on the node, then made a ~30m/s correction before leaving Kerbin SOI which was most of the inefficiency).

Because Low TWR vehicles have huge dv capacity the inefficiency just isn't a big deal.

Edited December 31, 2015 by Requia

[Rune]

12 hours ago, Wcmille said:

This method seems very good for low-velocity hyperbolic orbits, but how do you leave Kerbin if your ejection dV is 2000+? Once you have about 1000 dV put into an orbit, it will start to go hyperbolic.

How do ships with extremely long burn times arrive in a system without wasting DV? If you don't have the TWR for a capture on the first pass, you are not stopping. If you have to start the burn far from the Pe, you are wasting dV. This probably doesn't matter in a Hohmann interplanetary approach, but a high-velocity one-tangent burn requires a lot to stop.

'K, let's consider a c3+2km/s escape burn. That's going to be something like 3km/s. The first 1km/s, as you can guess, can be broken down in however many burns as you like, since you won't leave kerbin's SOI. Then comes the one that puts you hyperbolic... and you just keep on going. If you have picked your ejection angle right, you will head out of kerbin on that direction, and you can keep burning all the way to the edge of the SOI, and if those hours ain't enough, you can always keep burning in the same direction once you are in solar orbit.

Capturing is the other way around, except you pick your windows to not get to the target with, say, +5km/s to slow down: you burn all the way in if need be, timed so that you go into capture just as you reach Pe, and lower Ap afterwards at your pleasure. Bonus points if you capture with a gravity assist...

7 hours ago, Wcmille said:

You raise a really good point here -- how does one calculate precisely the number and type of radiators?

Rule of thumb, each Nerv creates about 14MW of heat to get rid off, so a medium radiator will do (maxes out at 19). I have also done ultra-long burns using four curved big ones for each nerv. Take care to have the radiators close to the heating parts, and without low-temperature ones that can blow up on the vicinity.

 

Rune. You do waste dV (or, to be more accurate, you don't benefit from mr Oberth's attentions as much), but it is totally worth it.

Edited December 31, 2015 by Rune

[sgt_flyer]

ok, made a few tests, for a high TWR nerva propulsion pod, i managed to get away with 1 single large radiator panel per nerva (the fixed curved one) for delaying the overheating. - could manage a 18 minutes burn at full thrust before shutting it down to prevent nervaexplosion from overheating. (not sure if it'll do a 15 mn burn with 80 tons of payload, but it should still be close.) - brought my orbit waaay past moho (PE sub 1000.000 km from kerbol, from a single burn from LKO)

 

vacuum full duration test burn roughly 30s after ending the burn - will still need a bit of time to cool down between burns

 

A view of the docking port subassembly (if a 3.75m adapter is ok for your mothership :p) - note, once you add the subassembly to your mothership, you should be able to pick those struts to stretch them higher on your mothership to limit wobbling a lot.

 

tell me if you want to test it (and if you need the launcher i used to put this monster in LKO too ;))

[Wcmille]

13 hours ago, sgt_flyer said:

tell me if you want to test it (and if you need the launcher i used to put this monster in LKO too ;))

That'd be fun. How can I get the craft file?

[sgt_flyer]

@Wcmille Happy new year i'll post a mediafire link with explanations later in the morning  (just came back from new year's eve party 

[sgt_flyer]

@Wcmille

 

Hello - here's the propulsion pod fitted on top of my launcher (this launcher is capable to put up to 900 tons into LKO - so the propulsion pod's 663 tons is not a problem (and if your mothership fits within the fairing, it could even launch the whole thing in one go - i've tested that with an added mass simulator of 80 tons on top of my docking subassembly - feel free to ask me for payload integration ^^) ;)) - i removed the side vernor pods from the propulsion, the pod's attitude control is done through it's large reaction wheel) . http://www./download/8dmeh66a6qua4kl/Nuclear+drive+pod.craft

And here's the dock subassembly to attach to your mothership, so it can dock with the propulsion pod. http://www./download/0zto70ouwcg78zg/NPpod+dock+subassembly.craft

(note, you might want to disable fuel usage on the propulsion pods fuel tanks before launch - to prevent the launcher's upperstage RCS from using the payload's fuel)

note, due to the additionnal radiators and docking assembly, i had to add additionnal fuel tank - so the initial 4000m/s DV manoeuver burn time might be closer to 19mn.

the propulsion pod is installed upside down on top of the rocket, to ensure a correct structural integrity.

Launch sequence : go full throttle and fire the engines

as soon as you reach a speed of 30m/s, start your pitchover manoeuver to initiate the gravity turn. (just tilt the rocket 5° eastwards as fast as the rocket's control authority allows you) once you reach a 5° angle from vertical, simply continue to follow the prograde marker. (you can slightly speed up or speed down the gravity turn by pointing your rocket slightly below or above the prograde marker's center)

you'll want to reach an angle above horizon of 45° near 15000m. stage the rocket boosters as soon as they are empty, and continue to follow the prograde marker. (note, once above 40000m, you can point the rocket straight towards the horizon if you wish)

once the core stage is fully empty, (the core engine burns out before the rest of the stack) stage once to eject the fairing, wait for the parts to clear the rocket a bit, and stage to activate the upper stage.

- lift your apoapsis until your desired orbit height, then coast to apoapsis and circularise with the upper stage. once circularised, you can detach the upper stage and deorbit it. (you can use the N key to use the slightly angled RCS for pushing the upper stage away from the payload after separation, before rotating it and deorbiting it)

switch back to the nuclear propulsion pod, select the probe core and hit 'control from here' so the control is in the correct orientation.

 

[Wcmille]

Thanks everyone for the feedback, suggestions and contributions. I've been continuing to evolve these ideas, and I wanted to send out some thanks, and show what I came up with.

@Temstar -- I hadn't even thought of doing something like that. Definitely see that as a future project for me.

@Rune -- Excellent points; particularly the one about part count. It definitely changed my thinking and turned my focus toward MkIII parts, which greatly reduced my part count. Also took your advice to relax a little bit on the burn time, that helped.

@sgt_flyer -- Very cool design. I really like the tricks for part hiding.

One thing I realized was that if I switched from just Nukes to Poodle + Nukes, it changed my burn dynamic for the better. Since I was dropping so much weight after I'd gone out, I didn't need the same amount of thrust back that I needed to get there. Ultimately, just adding one Poodle made a significant difference.

The base craft is designed to only have 8000 dV loaded and back. I realized I had a transfer window to Moho coming up, so I decided to go there first. The craft won't make that, so I had to drop the carried payload from 60ish tons to about 17 tons and added dockable boosters to the back. This brought me to nearly 11,000 dV, which will make Moho. I'm going to Eve after returning from Moho, so I should be able to take the full payload there.

Here's the album. I hope you like it: