Platypus Dive Research Vehicle

[Hotel26]

This is the Platypus Dive Research Vehicle.  I will give an overview of what is inside and how it works.

 

Pilot Control Overview

1. fast-flood main ballast tank
2. fast-vent main ballast tank
3. fine-flood main ballast tank
4. find-vent main ballast tank
5. trim for'ard (moves mass from the aft tank to the for'ard tank)
6. trim aft (moves mass from the for'ard tank to the aft tank)

Craft-File Editing: Resources

The distinctive monoprop tanks are actually, from fore to aft: XenonGas, Oxidizer and MonoPropellant.  The first of those two tanks was edited to include a new Resource section for the desired contents and with capacity expanded as desired.  I retained the original MonoPropellant section but zeroed amount and maxAmount and inverted all the booleans (True to False and False to True) to nullify them.  It's possible to simply remove them but for some kind of stock tanks, KSP defaults the 'known' resource back into existence if it does not appear at all.

The for'ard trim tank is 24,000 units of XenonGas and the aft trim tank is 600 liters of MP.  Respective densities are 0.1 kg/liter and 4, giving a 1-to-40 ratio in terms of mass.

The 'standard' C7 Brand Adapter LFO tanks have been modified to contain a pure 800 liters of LF each.  They power the Wheesley for conventional propulsion.  I did just remove the Oxidizer section in these tanks and expand the capacity in liters of LF to be the sum (360 + 440), such that the total did not expand.

Worth making a distinction here that this Adapter is a standard tank in the machine used for a conventional purpose, so I never 'cheat' when 'monkeying' (research term; look it up) with the contents of such 'for convenience'.  The ballast system tanks are 'open season' because they are part of the 'magic'^[TM] side here.  There are mods for this kind of thing, but I generally find it desirable to avoid e.g. config files that others might then have to download.

Craft-File Editing: Engines for resource production

Apart from the conventional propulsive Wheesley, the following engines can be identified: two Twitches, a Puff and a Dawn.  Their ModuleEnginesFX sections have been modified to set independentThrottle True and independentThrottlePercentage to a calculated (hugely) negative amount.  The Twitch supplying OX for fast-flood of main ballast is geared to -40404.0391%, for example.  The fine-flood Twitch uses -12121.212%.  The OX is pushed to the central main ballast (yellow) tank.  (The catch is that LF is also produced by this technique!)

A worked example is worthwhile.  I wanted to produce 250 liters of OX per second.  The Twitch is rated at 1.125 liters/sec (LF + OX) fuel consumption.  We want a percentage thrust.  Hence:

-250*20/11*100/1.125 = -40404.04.04

and you will realize that the multiplication by 20/11 (LF:OX being in a consumption ratio of 9:11) is intended to scale the OX component up to the total fuel production of the Twitch.

The Dawn produces XenonGas for the for'ard trim tank and the Puff produces MP for the aft trim tank.

Main ballast buoyancy control is relatively uncomplicated.  Two Twitches producing LF/OX at different rates are employed for flooding and the OX goes directly to the single main ballast tank.  Venting simply opens one of two drain valves on the OX tank, selecting the appropriate rate.  The wrinkle, however, is that, while flooding the OX tank, LF is being pushed into the twin fuel (LF) tanks.  Two opposed LF drain valves on one of the LF tanks, set in Vessel-draining mode, are opened to exactly vent the same amount of unwanted LF as is being produced by the running Twitch as a by-product.  There are two sets of these LF valves, matching the two flood rates.

Trim adjustment uses a similar technique for different reasons.  Trim For'Ard produces XenonGas for the for'ard trim tank, adding mass.  An equivalent mass has to be removed from the aft tank, so this is done by venting MP via a drain valve.  Trim Aft works in reverse.

So a note about the initial priming of the trim ballast subsystem.  When those tanks (Xe and MP) are empty, Either of the Trim modes will fill one tank while venting an already-empty tank: a no-op.  Thus, you can prep for ocean-going by filling one trim tank, while loitering on the surface.  Filling the tank takes a while, but the advantage of completely filling it is that that total mass in the trim tanks is now exactly half of their total carrying capacity.  That gives a) a known quantity consistent across every dive, and b)the greatest 'leverage' on trim balance via the greatest trim range.  After priming, the opposite Trim command can be employed to balance fore and aft equally or, better yet, when the sub has been tuned to known quantities, that known balance can be dialed in, ready for dive operations.

Submersible parameter tuning

A trim ballast subsystem is possibly not so necessary on small, simple craft but ought to alleviate a lot of work at design time for large and/or complex submarines.  Furthermore, dynamic trimming for differing fuel loads or to compensate at speed for unbalanced 'aerodynamic' drag may become important.

A primitive guide to trim while in motion and with SAS engaged is to observe the plane angle and trim until is it neutral.

The overall best and easiest way to tune is to take a submersible in its initial sea trials to the seabed (as an easy way to arrest/reset motion and attitude.  Vent main ballast until the craft slowly rises.  Patiently true out the vertical speed to zero.  Release SAS and then observe attitude drift: bow (nose) up or down.  Be patient but it is quite easy to find settings in which the sub hangs motionless in the water, with zero degrees attitude.  Note the numbers, main and trim quantities.

Planning and Design

I ran a spreadsheet on everything to make it easier to adjust settings when anything unexpected forced a change.  I'll expand on this in the near future...

Issues and Future Work

KSP truncates drainRate values into the range 1-20%.  This is the annoying kind of 'fact-checking' that modern programmers must do to comply with corporate standards for making bland, boring, unimaginative, creativity-suppressing products.    Where would we (I) be without the imaginative lack of concern displayed by Squad programmers when they wrote the code that trusted the human-editable craft file when it specified a vastly negative throttle setting, hey??

OK, so this forced me to have a faster fine-control vent rate than I wanted.  Particularly as this was ultimately determined also by the LF fuel capacity for LF/OX production reasons.

Adding valves (doubling up) can conceivably lift the 20% limit (although not very relevant to subs unless you wanted emergency evacuation in less than 5 secs).

I noticed that the drain rate on one of multiple tanks, when set in Vessel-drain mode, appears to take the percentage on the tank (part) it is connected to, and disregards the total capacity of that kind of tank across the vessel.  This suggests that a small, auxiliary tank can be added in some circumstances, with a vessel drain and a rate set, calibrated on the size of the small tank, such that an effective drain rate significantly less than 1% can be achieved.  (I have assigned an underling on my staff to look into this, but he is known for laziness; well it will be his last chance!)  This will work when like tanks are being filled/emptied uniformly.

Another item to note relates to why trim ballast on Platypus is this strange platyputtian mix of Xenon and MP.

The original intention had been to use MP uniformly across the two trim tanks and in the main ballast.  Filling the for'ard trim tank would require producing MP for it, and the aft and the main, in the desired amount scaled across the total capacity of the three tanks.  The main would simultaneously drain what it received and the aft trim would drain twice what it received.  Unfortunately, fuel produced is not stored in the ratio of the individual tank's total capacity, but in the ratio of the tank's remaining capacity (i.e. emptiness).  (A desirable way to do things, really, and exactly what my mod, Telemagic, does: to each according to its need and from each according to its ability.)

Platypus unintentionally flies.  Badly.  With its puny wingspan (staying that way), its approach speed for a conventional landing would resemble a ballistic projectile.  It can currently be flown to a dive site and enter the water, nose-first, under chutes at about 9 m/sec.  Chutes repacked, and after a submerged vertical speed run, it can get airborne again and could be flown to KSC and ditched offshore and then taxied onto land.  It is probably not hard at all to change the chute arrangement for it to perform a vertical gear-down landing on land (easy).  The superable challenge would then be to effect the belly landing on water without breakage.  I think a v3 update will add this capability shortly.

Acknowledgements

The first time I heard of the engine negative-gearing for resource production idea was from Dr @swjr-swis.  (See the furtively beautiful Anion.)

When I ran into the first set of issues with the original objectives (listed in Pilot Control Overview), I ruminated on what would be necessary to overcome them (and would eventually find an additional issue (unwanted production of LF)) but decided to publish Platypus in a slightly-restricted form ("publish or perish" re grant schedules, you know).  Gratitude again to Dr @swjr-swis for the "push" to go the whole mile.

Edited September 6, 2023 by Hotel26

[Wizard Kerbal]

How does it work?

[Hotel26]

Platypus was a prelude for work commissioned by Admiral @chadgaskerman for his Project 432 and TEK-46 (the latter just because I like it).

   

I have in mind to go for a simpler set of controls (at first) for these large submarines.

Inspection of the TEK-46 shows it already equipped with fore and aft ballast (ore) tanks; 3 at each end.  My initial plan will be to convert these to a set of two monopropellant tanks at each end: one main-ballast size and the other trim-size.

Then add a Fast Flood/Vent control for the main ballast tanks (fore & aft).  Then Slow Flood/Vent controls independently for fore and aft trim.

Fine adjustments to total ballast can be performed by e.g. additional flooding into fore and aft tanks simultaneously.  Then manually transferring some amount to/from main ballast.   Possibly, the trim tanks will be over-sized for trim permitting a main ballast quota for adjustments, without the necessity of transferring to the main tanks -- as long as capacity for trim flood/vent remains for the primary function of boat trim.

Priming the boat for dive operation would entail:

• an initial flood of main tanks to provide sufficient ballast for
• transfer  to trim tanks to bring them to 50% each^[1], then
• flooding main to overall neutral buoyancy condition, then
• diving to a shallow depth and leveling off to
• trim the boat for neutral attitude

[1] performed equitably by flooding for'ard trim to 100%, deducting from both main tanks; then transferring (manually) 50% from for'ard trim to aft trim

If there is a forseeable "pressure point" in this control design, it is whether fine amounts of ballast can be transferred accurately for small adjustments.

It is attractive to start with a simple design and then augment it only when and how required.

There is an argument, in fact, to begin with just fore and aft main ballast and determine whether manual pumping between them is accurate enough for trim.  I will, in fact, start with this scheme and possibly using heterogeneous sizing to pre-dial some "one-size" trim into the boat for initial dive.

This thread is certainly open to comments, as well as for showing project progress.

UPDATE: the TEK-46 does work, as is.

Edited September 11, 2023 by Hotel26

[chadgaskerman]

10 hours ago, Hotel26 said:

Priming the boat for dive operation would entail:

• an initial flood of main tanks to provide sufficient ballast for
• transfer  to trim tanks to bring them to 50% each^[1], then
• flooding main to overall neutral buoyancy condition, then
• diving to a shallow depth and leveling off to
• trim the boat for neutral attitude

Personally I think so long as you can find the centre of boyancy and trim to that, you can fairly easily add a lot of inert ballast so it's just barely positively boyant, and then only have to use a small amount of active ballast to drop it into negative boyancy. finding the centre of boyancy is a lot easier on something with a low part count due to there just being less points of force being exerted on the craft.

Essentially the key to making a submarine is being able to locate the centre of boyancy and then add ballast, after that you can trim with hydrodynamic surfaces and/or ballast.