Inspiration Kerbal

[von Ziegendorf]

Moho cycler was the general concept for that mission - inspired by the Messenger probe. But it turned out not to be as simple. You can see that correcting orbit after a Pe=228km costed 114m/s. Unfortunately had to skip 4th Moho flyby in order to be able to return home If not I would have been ~2 days late for date with Eve.

The original Kerbin-Eve-Moho orbit was found using TOT, then I was flying around to catch up with Eve in the same position when I made the first encounter. And during that I saw opportunity for ending the mission with another triple flyby rather than simply fly directly to Kerbin.

[donfede]

This has been a most enjoyable challenge, and also quite educational (I've learned more about orbits, and this was my introduction to the Inspiration Mars mission). Below is my attempt to make a presence on the leaderboard, with the first posted Free Return Trajectory to Dres.

I was initially seeking FRTs by randomly trying maneuver nodes at each celestial encounter; results were minimal. Reading success reports from others, I made it a point to create an encounter orbit that returned on an integer value of Kerbin's period around Kerbol(106 days). Starting with the efficient orbits suggested by alexmoon's porkchop plotter, I would then raise the apoapsis to create a "collision orbit" that passes a body twice; in my mission I met Dres after the apoapsis. The Dres encounter I ended up using (Departure year2, day 134, 13:39; Phase Angle ~93deg ; Ejection Angle ~100deg; Ejection delta-v 1656ms), was only late by ~6 days. This was correctable with a little RCS burn (~150ms) after reaching Dres' SOI. Alas, the Kerbin departure and arrival were both in the dark.

As this challenge is already in progress, scoring should probably not be altered. However, Time Of Flight was clearly a significant factor in RL planning of FRT. Note the porkchop plot below, used in the Inspiration Mars mission. It has *many* FRTs (as we've been finding ), but only a couple options with usable TOF (circled in red; from http://www.inspirationmars.com/Inspiration%20Mars_Feasibility%20Analysis_IEEE.pdf).

Photos from my entry:

Javascript is disabled. View full album

Mods used: TriggerTech's Kerbal Alarm Clock

Craft are stock; designed v0.23, mission on v0.23

kinsk01-KMk2 craft (habitat module, sent up on first launch)

kinsk01c-KMk2 craft (crew launch craft, based on previous effort from another challenge)

draft scoring:

b (multipier) = 2

P (C/A to Dres in km) = 523

dv = dv1 + dv2 + dv3 = 1999.26

LV909 Isp vac = 390

RCS Isp vac = 260

dv1 (transfer module) = LN(25.89/16.06) * 390 * 9.81 = 1826.96

dv2 (habitat module) = LN(13.14/12.31) * 260 * 9.81 = 166.42

dv3 (KRP) = LN(4.35/4.34) * 260 * 9.81 = 5.87

D (duration of total round trip in days; apologies if I borked this ) = de - ds = 332

ds = year 2, day 134, 13:39

de = year 3, day 101, 21:02

U (units of monoprop used) = u1 + u2 = 208.25

u1 (habitat module) = 270 - 62.51 = 207.49

u2 (KRP) = 30 - 29.24 = 0.76

M (mass at launch) = 209.08 + 11.42 = 220.5t

m1 (habitat module launch) 209.08t

m2 (crew launch) = 11.42t

S (score) = (5000 * 2 * (5 - log10(523)) ) - 1999.26 - (3 * 332) - 208.25 - 220.5 = 19390.97.

Not a great score, but I like my craft (as I bet the kerbals did ), the design and mission were fun, and I broached some new ground. It is the first post to find a Dres FRT , though I've no doubt many more exist. I'm slowly looking over alex's code, hoping it can be modified to point out these FRT opportunities (much like was done for the Mars mission above).

best,

kdonfede

--

"Adding K to every word..."

[PLAD]

Well done donfede, Dres requires a lot of MCC and time to get to. I particularly like your habitat and the mission architecture, it's got a nice touch of realism.

I've been wondering how the TOT works, and what paths it might throw away when searching for the best one, and I think I've come up with a way to use a pork-chop plotter to find flyby and return missions. See the pictures below. My thanks to the author of the plotter I use in the example, which was linked to earlier in this thread. And I haven't read more than a few hundred threads on this forum so far, so my apologies if someone found this already.

Javascript is disabled. View full album

It uses the fact that during a flyby the speed you leave the planet at will be the same speed you arrived at, relative to the planet. So finding an ejection speed that matches the insertion speed (at the same altitude!) guarantees that the flight is theoretically possible. The only thing that makes some of the solutions impossible is that they require a flyby periapsis lower than the surface of the planet (or its atmosphere).

Every point on the original pork-chop defines a line in the second pork chop (I call those pork slices, ha). Some slices have no points that have the desired ejection dv, but a lot have 2 or more. You could make a 3-d graph up from the original chop, you would end up with a bunch of points indicating where the free-returns are. I suspect that they would appear as closed ellipsoidal blobs whose surfaces defined all possible solutions. Is this what those 3-d graphs that TOT flashes while it is searching are?

For multiple flybys you would take one of the points found on the 2nd pork chop and use it to make a 3rd plot. The odds of finding something are surprisingly high if you start in the zone of the first pork chop where the departure speeds are not too far above the minimum, since when it's too high the target planet can't change the path much, or too close to the minimum, where there isn't enough energy to give a large range of orbit sizes.

Edited March 17, 2014 by PLAD
its clearer if I specify the FLYBY periapsis

[Jasonden]

It is the first post to find a Dres FRT , though I've no doubt many more exist.

Dres -- score! I like the approach of getting the orbit resonant with Kerbin and then looking for a Dres encounter. That should work well with any low-mass target, I should think. Maybe the asteroids in 0.23.10 as well?

[PLAD]

Hooboy, this one has grabbed my free time lately. First a mention on my earlier post- I agree with von Ziegendorf- multiple flybys have scientific value, witness Mariner 10. So restricting points to 1 score per planet is not optimal. As v-Z and donfede touch upon, total flight time is the important thing when a life form is on board, and the current scoring deals with that.

Donfede's post above gives a pork chop plot of departure date versus flight time for Earth-Mars-Earth free return missions. I've been slowly making one like that for Kerbin-Eve-Kerbin. Doing it by hand is too slow to try for long, but I think it can be done with relatively small modifications to a pork chop plotter. Here is the algorithm I'd use. Note that only 2 pork chops need to be made, one could then put those two in memory and most of the program is comparing the points in those two 2-dimensional arrays to find free returns. The next step is to determine flyby periapsis requirements and rule out the impossibly low ones, then finally plot the results as below.

Javascript is disabled. View full album

I bet some of the sub-100 day free returns are possible.

I ran TOT for K-E-K paths launching in this time frame and it gave me one that leaves Kerbin Y1:143:21 at 1136 m/s, flies by Eve 50 days later, and returns to Kerbin 144 days after leaving. It clearly searched for the solution with the lowest initial dV from LKO. There are solutions at least 30 days shorter, but needing at least an extra 100 m/s.

Are the values for the Kerbol system orbits correct as given in the wiki? Better yet has someone got a VSOP-type set of terms for their orbits? I'm guessing there are no perturbations involved in the Kerbol system motions at all so there would only be one set of terms for L B, and R for each planet.

Note that this method would apply to any free return flyby missions, you could keep using the arrivals at each planet to quickly search for paths to the next planet. We need a way to deal with paths that flyby the same planet twice though.

Enough for now. This is an instructive challenge!

[donfede]

...I've been slowly making one like that for Kerbin-Eve-Kerbin. Doing it by hand is too slow to try for long, but I think it can be done with relatively small modifications to a pork chop plotter. Here is the algorithm I'd use. Note that only 2 pork chops need to be made, one could then put those two in memory and most of the program is comparing the points in those two 2-dimensional arrays to find free returns. The next step is to determine flyby periapsis requirements and rule out the impossibly low ones, then finally plot the results as below.

http://imgur.com/a/iozSj

...

Excellent writeup and information PLAD - thanks for sharing! There must be potential to morph alexmoons porkchop plotter into a tool to identify FRTs. Absorbing your pics/notes now; look forward to playing with alexmoon code tomorrow

Edited March 23, 2014 by donfede
cleanup quote