Low Mass Eeloo with TAC-LS - kOS script chaos

[DBowman]

Lately I've been doing some Eeloo missions for @Firemetal's low budget Eeloo mission challenge - playing around developing the mission plan is a lot of the fun so I thought I'd try making a mission development diary style report.

I have some reasonable fast transfers I found earlier:

• Kerbin 2400 m/s departure burn
• Eeloo 1400 m/s capture burn
• Flags and Footprints using an ion lander I developed earlier
• Eeloo 2000 m/s departure burn
• Kerbin aerocapture

5 years 380 days

 

This is a mass / deltaV / staging mock up that I think will work ok. (http://imgur.com/cHTofCx.png)

Everything but the departure is straight forward. Nuke burns and staging away tanks and life support containers (1.6 ton of LS 'tankage' for the outbound trip ). I don't think I can (want) to Ion capture since there is 'no sun' it will be running on 2 RTGs which is like 10% thrust which means maybe 6 hours of burn. I don't mind leaving it to run while I do other things, but the navigation is a worry.

The Kerbin departure is the tricky part. Ion powered, Pe kicks then continuous burn powered by the solar arrays. I have a good 'seat of the pants' technique for hitting the final departure burn at the right place and time, I figure it will add 20 - 30 days of kicking time before the departure burn. There will still be 1500 m/s - ( three hours! ) of burn to do while the craft is swinging around Kerbin. Just escaping Kerbin SOI at the right time and then doing one 'straight' burn doesn't sound optimal. I'm not sure how to do better without mod support (some kind of extended time continuous thrust node) but I have some thoughts to try out.

The vacuum craft is about 50% heavier than anything I've space planed before, so that will be interesting also.

My current cutest space plane launched vehicle , the vacuum craft is less than a ton. (http://imgur.com/2CoVgdn.png)

Edited September 20, 2016 by DBowman

[Firemetal]

Looking good! I'm thinking of changing the name to low mass Eeloo challenge since Ions aren't very "low budget". If you can do this with life support and one ship, I'll add a life support leaderboard.

Good luck!

[DBowman]

Low TWR and what to do about it...

We all know that usually propellant dominates mass so getting lower mass means using higher ISP engines, more deltaV per kg of propellant - but high ISP engines are low thrust, which means long burn times ...

It's easy enough to leave MechJeb executing nodes for as long as you want, the problem is what nodes? For those who haven't come across the problem; an ion craft may need to burn for an hour to impart the deltaV required for a transfer burn. You cannot just burn for an hour 'in the right direction' because you will spend most of your time burning at the planet / retrograde / radially...

Half the solution is a series of Pe kicks + a final departure burn. The Pe kicks raise your orbit's Ap as much as you can without escaping, the rest of the required deltaV will be the final departure burn.

 

(Pe kicking & a reference trajectory)

You have to arrange that you will be at Pe when you need to do the final departure burn. You could 'do the math' and back calculate all the Pe kicks, but you can get close enough using some seat of the pants astrogation and laying down a few nodes. The key thing is you can lay down a vanilla departure node from a reference craft (the whole 1400 m/s for example) and your real craft's final departure burn node ( for example the 600 m/s left after Pe kicking a few or many times). On your second last Pe tweak the deltaV to set your last Ap to place your departure node over the reference node - you will be in the right place at the right time to do the final burn.

(adjusting the final departure node - only one kick in this example)

Orbits with Ap past the Mun to near Minmus have a wide range of orbital periods so you have a pretty wide window when you can start Pe kicking - it depends on how much you can kick each time, for a craft that can do 100 m/s in six minutes leaving 20 days early gives you enough time.

The second half of the solution is doing something with the final departure burn. You may still have a lot of burn time left, too much to burn 'at' the planet and miss. Even if you can avoid disaster you are introducing large errors that you have to correct later, and you cannot really tell what the correction will be until you have done the burns. Ideally we'd have a non instantaneous 'maneuver path' rather than a node, but I think I have a reasonable way to avoid burning 'on faith' just with the in game tools.

Remember your final departure deltaV and modify the node to be your kick size. Then lay another identically sized node one burn time ahead, repeat until the subsequent trajectory is 'pretty straight', drop a final node with the residual deltaV.

These two show about 1500 m/s of a 2400 m/s Eeloo burn - you can see now the trajectory is flat enough that you could then use one big 900 m/s two hour burn one hour out from the last node.

Now you have a well defined trajectory and can make corrections by tuning the final node. The bigger you can kick the less nodes you will have to put down. This method should be much better than one burn - your predicted and achieved trajectory should be much closer than 'one big burn'.

 

The imperfections in the method include; to be safe you will leave gaps between burns (not optimal), the gaps will get larger as the craft gets lighter (but calculating the burn times and adjusting each one would be tedious), and the  burns start 'below prograde' and end 'above prograde'.

Oh and importantly since the craft is going slower for longer it's orbit will be more wrapped around the planet than an instantaneous burn. One way you could compensate for that would be moving all the nodes back in time - so tedious  but maybe some KOS script could help...

Edited September 10, 2016 by DBowman

[DBowman]

KOS ...

programmatically created nodes, 200 seconds apart. Scripts can easily shift the times for the whole list also.

[DBowman]

A crazy Mun 'encounter' dialed up with a 'make n identical low thrust burns jammed together' function. 12 times 110 m/s six minutes apart - 1320 m/s to the Mun - but maybe it makes sense to move something really massive with nukes.

Edited September 11, 2016 by DBowman

[Firemetal]

Hmm. Looking good. Is this mission a single launch?

Also on a side note, I think @hazard-ish is doing my challenge. It was at the end of his most recent vid as 3.6 ton Eeloo mission. 

[hazard-ish]

7 hours ago, Firemetal said:

Hmm. Looking good. Is this mission a single launch?

Also on a side note, I think @hazard-ish is doing my challenge. It was at the end of his most recent vid as 3.6 ton Eeloo mission. 

Wow, I had no idea that this was already a challenge! I'll be doing something similar but it might not qualify. Ill post the video here when it's done

[Firemetal]

.

Edited September 12, 2016 by Firemetal
Empty post! Please delete!

[DBowman]

19 hours ago, Firemetal said:

Hmm. Looking good. Is this mission a single launch?

Also on a side note, I think @hazard-ish is doing my challenge. It was at the end of his most recent vid as 3.6 ton Eeloo mission. 

eeek 3.6 .... that's low...

Edited September 12, 2016 by DBowman

[DBowman]

It looks like I can make some good enough low thrust trajectory helper scripts, but enough kOS for a while.

@Firemetal I am looking at a single launch. I put this together pretty quickly - basically a plane with a nuke&ion vacuum stage strapped on top. My dV guess for the closed cycle stage was close. Then I added a remote control in case I want to try to land it (which I've never done...) and adjusted the propellant so it should have 'just a little left over'. It flies ok, as long as you throttle limit the lower R.A.P.I.E.R. as the lower tanks empty. Looks like around 36 ton all up, maybe 6 ton recoverable if I could land it.

The nuke vacuum vehicle 'stages' the fuel and LS tanks by decoupling chunks of it - to try to keep the fairing small. The docking ports don't seem to make horrible drag.

[DBowman]

I've been playing more with kOS to try to make low TWR easy by making some scripts / functions.

The plan was:

1. find an initial impulsive transfer burn the usual way; MJ or alexmoon etc
2. write it down in kOS so you can create it over and over
3. have a reference craft in your departure orbit. You can setup the transfer on that and then with your real craft you target it and can see the trajectory - try to make the real low thrust trajectory look like that impulsive one.
4. Split the real burn in two; one part big enough to get an Ap past the Mun and the rest
5. Split the sub escape part into Pe kicks: PeKNodes( 50, 900, 3, ALLNODES[0] ) - this does 4&5 in one step splits the first node into 50 m/s chunks up to 900 m/s starting 3 orbits ahead of now
6. Adjust the last Pe kick node magnitude 'by hand' to get the departure burn to lie over the reference node: DeltaMagNode( 18, -1) - this adjusts prograde, normal, and radial to make the result magnitude 1 less than before (that's right already 19 nodes...)
7. Now the split the departure burn into chunks so you can see the effect of the planet cranking your orbit around
8. Move all the nodes back in time 'by hand' until the resulting trajectory 'looks like' the reference one: DeltaTNodes( -232 ). Adjust the last node here you can see it needs some extra dV: DeltaMagNode( 42, 100 ).

All looks good and basically works but:

1. I love MJ but I don't think it burns properly for these kind of craft; it seems to do something 1/3 burn time before and 2/3 after the node. I think it should be 50:50 for these small low TWR low propellant burns. I have to look at that.
2. I love MJ but you cannot physics warp and execute all nodes cause it pops out of physics warp after each node. So right now if you want 'hands off' auto pilot you have to run it in real time. I want to fix that.
3. There is something wrong with my node building logic, at least for very very low thrust. Here is a 6 hour burn plan at 0.08 m/s² from Eeloo. IT says 0.5 ms² but it will burn at 15% throttle, limited by 2 RTGs. I should fix this - once I figure out whats wrong, maybe it just looks odd

 

 

[danielboro]

how can i find the kos script?

[DBowman]

11 minutes ago, danielboro said:

how can i find the kos script?

I'll tidy it up, add some comments, and post it. I can just paste in in this thread I guess.

[DBowman]

I finished up the Eeloo mission here ...

One thing I learned is that the Pe kick and the capture are pretty easy to do with very low TRW. The hard part that also generates a lot of errors / deltaV buffer requirement is the departure burn. How much would it help just doing the easy bits?

I compared three options:

1. 16.809 t Just Nuke
2. 13.051 t Nuke & easy Ion
3. 06.474 t Pure Ion

I think it would make much more difference for Duna where the post Pe Kick part is relatively much smaller.

[DBowman]

On 9/20/2016 at 6:34 PM, danielboro said:

how can i find the kos script?

I made it a mod...

MechVal a Low T Auto-Co-Pilot

I tidied up the scripts so I hope they are readable etc - let me know if you use them / have trouble with them.