[WEB] [1.0.5] KSP Optimal Rocket Calculator v1.20

[problemecium]

It pleased me to find that according to the calculator I'd already (pretty much) found the optimal solution for a single-stage orbital lifter ^^

[Pecan]

It's finding better payload-ratios than I usually do. Even letting it run for a while it's quicker than building in the VAB and launch testing too :-)

I've just let it run for some 9,000 designs (4,000 valid) to provide a stock, 8t-payload, asparagus staged, Kerbin-atmosphere launch vehicle, optimised for mass. It beat me by around a ton (48.9 vs my 50).

Some observations that I hope will help:

i) In a couple of the top designs it says to use decouplers, fuel lines and tanks in multiples of 4, instead of 2. I assume this is because it had calculated that the stage should have 4 engines but it is simpler and lighter to cluster these, rather than build them as separate stacks. In at least one case this also allows the fuel-tanks to be better optimised too.

ii) It seems to quite favour slack-tank (engineless) stages between other, powered, ones which is quite interesting. I'd always assumed they should be the first stage(s) jettisoned to get rid of the dry weight, but I can see the logic of the choices it finds. In one case, however, it's slack-tanks were entirely Round-8s and Oscar-Bs, neither of which will attach to radial decouplers. Not a problem when it comes to building it, as it happens, but an extra 'twist' in KSP.

iii) In the case above the total fuel in the stage was ~60 (can't be bothered to calc the Oscar-B fractions ^^). It seems unlikely that such a small, slack-tank, stage would be optimal on anything but the smallest vehicles. Since the previous and subsequent stages both included more substantial amounts of fuel this design seems odd.

iv) And finally; the best design - by mass - included a couple of 24-77 engines on the outermost (first to be jettisoned) stage. Haven't tested yet but I don't think these will prove to be worth their weight, even for the short time that stage lasts.

Good job. Now I really have to start work on my YARD (Yet Another Rocket Designer (with a nod to YACC)).

[quietsamurai98]

Suggestion: Rockets could be saved, then be used as a payload for another rocket.

For example: I want to go to the Mun and back.

The optimizer first makes a LMO return rocket, which then goes on a Munar Lander, which goes on a LKO->LMO transfer module, which goes on a Lifter.

These all are saved, but have independent algorithmic inputs.

[GaryCourt]

In a couple of the top designs it says to use decouplers, fuel lines and tanks in multiples of 4, instead of 2. I assume this is because it had calculated that the stage should have 4 engines but it is simpler and lighter to cluster these, rather than build them as separate stacks. In at least one case this also allows the fuel-tanks to be better optimised too.

Yea, the tool currently doesn't support clustering. Something to add to the todo list.

ii) It seems to quite favour slack-tank (engineless) stages between other, powered, ones which is quite interesting. I'd always assumed they should be the first stage(s) jettisoned to get rid of the dry weight, but I can see the logic of the choices it finds.

It's likely trying to maintain the TWR in this case.

iii) In the case above the total fuel in the stage was ~60 (can't be bothered to calc the Oscar-B fractions ^^). It seems unlikely that such a small, slack-tank, stage would be optimal on anything but the smallest vehicles. Since the previous and subsequent stages both included more substantial amounts of fuel this design seems odd.

I imagine that had you given the program enough time, it would have eventually weeded this one out for a design with one less stage.

iv) And finally; the best design - by mass - included a couple of 24-77 engines on the outermost (first to be jettisoned) stage. Haven't tested yet but I don't think these will prove to be worth their weight, even for the short time that stage lasts.

Well, you do use most of your fuel in the first few seconds of flight. However, it's likely a TWR thing.

Good job. Now I really have to start work on my YARD (Yet Another Rocket Designer (with a nod to YACC)).

Thanks! Happy rocketing!

[GaryCourt]

Suggestion: Rockets could be saved, then be used as a payload for another rocket.

For example: I want to go to the Mun and back.

The optimizer first makes a LMO return rocket, which then goes on a Munar Lander, which goes on a LKO->LMO transfer module, which goes on a Lifter.

These all are saved, but have independent algorithmic inputs.

I'm not quite sure how this could be implemented without making the interface too complicated.

In the mean time, you can open the app into multiple tabs and transfer the data (mass) manually.

[Pecan]

Further to my previous post - the 24-77s improved the design (by 10m/s) so I was wrong, again, and the tool was right :-)

[breakthrough]

I was curious how it would handle part count and got this

Attempt #41
Stages: 2
Part Count: 49
Cost: 55650
Mass: 243.850
Delta-V: 4857
Time: 270.4
Stage 	Part 	Cost 	Mass (t) 	Delta-V (m/s2) 	TWR 	Time (s)
0


    Payload: 37t

	0 / 0 	37.000 / 37.000 	0 / 4857 	0.00 	0.0 / 270.4
1


    TT-38K Radial Decoupler x7
    Rockomax X200-16 Fuel Tank x7
    LV-T30 Liquid Fuel Engine x7

	22750 / 22750 	71.925 / 108.925 	2623 / 4857 	1.41 	135.2 / 270.4
2
Asparagus 	

    TT-38K Radial Decoupler x7
    FTX-2 External Fuel Duct x7
    Rockomax X200-32 Fuel Tank x7
    LV-T30 Liquid Fuel Engine x7

	32900 / 55650 	134.925 / 243.850 	2234 / 2234 	1.26 	135.2 / 135.2

What exactly are the radial decouplers attached to when it returns a stage made entirely of them? Similarly, are all 7 stacks of the asparagus stage feeding onion style into the stage above or is it actually dropping 1 stack at a time spiraling around?

[GaryCourt]

I was curious how it would handle part count and got this

What exactly are the radial decouplers attached to when it returns a stage made entirely of them?

In stage 1, they would all be attached to your payload. In stage 2, each stack would be attached to a stack in stage 1.

Similarly, are all 7 stacks of the asparagus stage feeding onion style into the stage above or is it actually dropping 1 stack at a time spiraling around?

All 7 stacks would drop at the same time.

[breakthrough]

In stage 1, they would all be attached to your payload. In stage 2, each stack would be attached to a stack in stage 1.

All 7 stacks would drop at the same time.

So both the payload and part counts are inaccurate since I would have to have non payload parts underneath my craft that invariably have landing legs and other equipment mounted radially. Still, it is interesting to see how my launchers compare to the optimal ones it generates.

[rkman]

(Never mind, i think i figured it out: the calculator does not include payload delta-v.)

I am getting strange results. For a 10 ton payload, 7000m/s delta-v, launched from Kerbin, the calculator comes up with a design that is about 3 times as heavy and has 3 times the thrust per stage, compared to a design a created myself.

Neither uses asparagus staging. I have no mods installed that affects rocket performance (such as FAR). Rocket stats obtained with mechjeb.

my design:

payload 10.7 tons

delta-v 7010m/s

total mass 118.55

(launch from Kerbin)

stage 1

x200-32

x200-8

skipper

stage 2

jumbo 64

jumbo 64

1 mainsail + 2* lv-t30

Optimal Rocket Calculator:

payload 10 tons

delta-v 7017m/s

total mass 302.55

(launch from Kerbin)

stage 1

FL-T200 Fuel Tank

FL-T400 Fuel Tank

Rockomax X200-8 Fuel Tank

Rockomax X200-16 Fuel Tank

Rockomax 'Skipper' Liquid Engine

stage 2

FL-T100 Fuel Tank x2

Rockomax X200-16 Fuel Tank x2

Rockomax Jumbo-64 Fuel Tank x2

Rockomax 'Mainsail' Liquid Engine x2

stage 3

Rockomax Jumbo-64 Fuel Tank x4

Rockomax 'Mainsail' Liquid Engine x4

When i build the design the calculator gives me, it ends up with 8500m/s delta-v. It has an excessive 2.4 launch twr where the calculator says it is a little over 2. The mass is 256 tons versus a little over 300 that the calculator tells me.

Edited February 1, 2014 by rkman

[breakthrough]

I take back what I said earlier, if I take stage 1 and just build a cluster with the same number of engines and fuel tanks it should be the same difference right?

Following that line of reasoning, it would be nice if there was an option to force stage 1 to have at least one gimbaling engine.

[GaryCourt]

it would be nice if there was an option to force stage 1 to have at least one gimbaling engine.

That's a good idea. I'll add it to the list.

[GaryCourt]

Just pushed out another big update:

* Added engine clustering to search.

* Added min/max burn time restriction.

* Added maximum mass restriction.

* Added maximum part count restriction.

* Added branch parts from Stock.

* Added KSPX 0.2.5 pack.

* Added KW Rocketry 2.5.6 pack.

* Now outputs just the parts in a stack, with a separate multiplier on how many stacks are required.

* Fixed bug with decoupler size matching.

* Now displays payload fraction.

* Changed max symmetry to maximum number of stacks.

Enjoy!

[Surefoot]

Yay for mod parts !! (now waiting for Novapunch parts, then life will be complete)

Just one nitpick, it seems to only produce radial stacks around one single stage under the payload, and further radial stacks around these radial stacks. While this is fine for stock game, it's not good at all when using FAR. My suggestion would be to limit the amount of radial stages through an option, so it would force the calculation of subsequent lower stages.

[GaryCourt]

Yay for mod parts !! (now waiting for Novapunch parts, then life will be complete)

If you'd like to make a donation to the project, I could get that done right away.</shamelessplug>

Just one nitpick, it seems to only produce radial stacks around one single stage under the payload, and further radial stacks around these radial stacks. While this is fine for stock game, it's not good at all when using FAR. My suggestion would be to limit the amount of radial stages through an option, so it would force the calculation of subsequent lower stages.

You can set the number of stacks to 1 to prevent radial stack from being created.

[Justin Kerbice]

great tool but I need some more explanation of the assumed building structure.

with these options:

Mass=6.798, DV=8000, g=Kerbin, mpc=50, mstag=3, mstack=2, clust, and the 3 last checkboxes checked

(feature idea: dump options used in a packed format that an be usefull for "debugging")

This design

Attempt #14

Stages: 3

Part Count: 29

Cost: 57165

Mass: 198.986 (3.42%)

Delta-V: 8072

Time: 431.0

Stage Part x Cost Mass (t) Delta-V (m/s2) TWR Time (s)

0

Payload: 6.798t

0 / 0 6.798 / 6.798 0 / 8072 0.00 0.0 / 431.0

1

FL-T200 Fuel Tank

FL-T800 Fuel Tank x2

LV-T30 Liquid Fuel Engine

4275 / 4275 11.375 / 18.173 2484 / 8072 1.21 152.1 / 431.0

2

TR-18A Stack Decoupler

FL-T100 Fuel Tank

FL-T400 Fuel Tank

FL-T800 Fuel Tank x6

TVR-1180C Mk1 Stack Tri-Coupler

LV-T30 Liquid Fuel Engine x3

14390 / 18665 33.762 / 51.935 2594 / 5588 1.27 149.3 / 278.9

3

TT-38K Radial Decoupler

Rockomax X200-8 Fuel Tank

Rockomax X200-16 Fuel Tank

Rockomax X200-32 Fuel Tank

Rockomax Jumbo-64 Fuel Tank

Rockomax 'Mainsail' Liquid Engine

x2 38500 / 57165 147.050 / 198.986 2994 / 2994 1.54 129.6 / 129.6

I build this and as you can see, decoupler are very weak so I got just a big firework

How are we suppose to deal with "FL-T800 Fuel Tank x6" or the stage3x2, there is no hint about location (tankx6 = "naive style" as I do or in a star pattern glued together or with a structural thing which is not a decoupler) ?

Thanks.

[Kerbart]

Very small point: the unit listed for entering dV is m/s² but should of course be m/s

[GaryCourt]

How are we suppose to deal with "FL-T800 Fuel Tank x6" or the stage3x2, there is no hint about location (tankx6 = "naive style" as I do or in a star pattern glued together or with a structural thing which is not a decoupler) ?.

The one thing this tool does not do is ensure the design is structurally sound. That is an exercise left up to the user (since, as you mentioned, there are multiple ways of building it). Since FL-T800's are the same mass and fuel as the X200-8, you could swap them out for more stability. Turning off the "tank size must be the same as the engine size" (last option) would likely have done that for you.

(feature idea: dump options used in a packed format that an be usefull for "debugging")

That is a good idea, I'll add it to the list.

[GaryCourt]

Very small point: the unit listed for entering dV is m/s² but should of course be m/s

Man, that's embarrassing. I've had it wrong all this time.

[Justin Kerbice]

Found a potential bug:

4 stg max, 1 stack max, clust, 01100 (cb 0 unchecked, 1 checked)

Attempt #96

Stages: 2

Part Count: 14

Cost: 15425

Mass: 50.630 (4.84%)

Delta-V: 7010

Time: 368.4

Stage Part x Cost Mass (t) Delta-V (m/s2) TWR Time (s)

0

Payload: 2.45t

0 / 0 2.450 / 2.450 0 / 7010 0.00 0.0 / 368.4

1

TR-2V Stack Decoupler

FL-T100 Fuel Tank

FL-T200 Fuel Tank

Rockomax X200-16 Fuel Tank

Toroidal Aerospike Rocket

6385 / 6385 12.203 / 14.652 4003 / 7010 1.22 207.9 / 368.4

2

TR-2V Stack Decoupler

FL-T100 Fuel Tank

Rockomax X200-8 Fuel Tank

Rockomax X200-16 Fuel Tank

Rockomax X200-32 Fuel Tank

TVR-1180C Mk1 Stack Tri-Coupler

LV-T30 Liquid Fuel Engine x3

9040 / 15425 35.977 / 50.630 3007 / 3007 1.30 160.5 / 160.5

The aerospike is not "stackable" as other engines.

And another issue, the tri-coupler + 3 LV-T30 make rocket go sideway, whereas with a Mansail, i's ok. What do you think regarding that ?

Edited February 13, 2014 by Justin Kerbice
1 remark added

[GaryCourt]

Found a potential bug: The aerospike is not "stackable" as other engines.

In my mind, that's a bug in KSP. But yea, I'll fix that.

And another issue, the tri-coupler + 3 LV-T30 make rocket go sideway, whereas with a Mansail, i's ok. What do you think regarding that ?

The LV-T30 doesn't have thrust vectoring, so you can't control your angle with it like some of the other engines. You need another control mechanism such as winglets or SAS.

[pantheis]

I just found this and figured, what the heck, I'll give it a go. I think there may be a bug with trying to figure out a rocket to launch really heavy payloads. I have a 470 ton station I'm going to launch all at once, but this doesn't seem to be able to find any possible options. I've checked the KW parts (as I have those), increased the max stages to 8 and max stacks to 6, tried checking/unchecking engine clustering, asparagus fuel routing, etc, and all that happens is the "invalid" count is screaming up by just under 100 a second.

If I drop the payload mass down to 100 tons, I start getting multiple rocket options.

Ideas?

[B787_300]

Could you possibly add options for only 1.25m, 2.5m, 3.75m parts so that you dont have to go through and manually unclick every part?

[GaryCourt]

I just found this and figured, what the heck, I'll give it a go. I think there may be a bug with trying to figure out a rocket to launch really heavy payloads. I have a 470 ton station I'm going to launch all at once, but this doesn't seem to be able to find any possible options. I've checked the KW parts (as I have those), increased the max stages to 8 and max stacks to 6, tried checking/unchecking engine clustering, asparagus fuel routing, etc, and all that happens is the "invalid" count is screaming up by just under 100 a second.

If I drop the payload mass down to 100 tons, I start getting multiple rocket options.

Ideas?

At that weight, the solution would involve just too many parts. The program limits itself to 8-16 tanks and 4 engines per stack per stage. You're going to have to break that bad boy into smaller pieces.

[pantheis]

At that weight, the solution would involve just too many parts. The program limits itself to 8-16 tanks and 4 engines per stack per stage. You're going to have to break that bad boy into smaller pieces.

Or, I'll just build the rocket myself, which I did. Why launch multiple things when I can go fully kerbal and just launch 500+ tons of space station all at once?