SEP propulsion for ISS stationkeeping

[Burninate]

I've made several comments in other threads about this idea, but I think it may deserve its own thread, and not in the advanced concept section.  I think this is a reasonable, conservative, near-term option to maintain the ISS, and reduce its consumables budget & ultimately cost.  Though it originated in discussions of what the ISS would do if the Russian Orbital Segment were detached, it's beneficial either way for improving the logistical situation.

I'm going to present it in a bit of a stream of consciousness manner, a prompting question per section and a paragraph per answer, to show my thinking in as much detail as possible and perhaps deflect both Jim's nitpicking and questions that have already been answered:

We keep discussing stationkeeping's fuel requirements, and the need for an expansive propulsion module that gets fueled frequently, in the event Zvezda is removed.  Why isn't VASIMR fixing this?



Chang-Diaz says in his talk that this VASIMR installation is strictly a testing mission, an experiment.  The ISS won't have much power to spare for his device, it's just a prototype that is too big to test easily in ground-side vacuum chambers, because of the substantial number of vacuum pumps necessary to fully evacuate these mass flow rates to the operating pressure on a continuous basis.  NASA ops doesn't want to have to mess with even the altitude changes it might make, but they're willing to donate 2-3kW and a spot on the station to the experiment.  It seems like if he could test it on a free-floating craft easily, he would do it - he says his ideal test site is the Moon's surface, because he could thrust all he wants without influencing position.  On the ISS, the plan is to charge batteries using short-term 'shots' of the device to change the *inclination* infinitesimally, without affecting the altitude, thrusting on alternating sides to maintain position.  Thrust will be measured by thrusting off-CoM and measuring induced rotation of the whole structure.

Okay, so that's out.  Why not just put some bog-standard commercial ion thrusters on it?  These are COTS items that are now standard equipment for GSO stationkeeping.  They're rather small for a 420 ton station, but ion thrusters work perfectly well in an array.  Part of the reason it's taken VASIMR forever to get to this stage is because it's too large to work for reasonable sized satellites without designing a solar array around it (also nobody seems to care for ultra-high ~12ks impulses VASIMR hopes to produce yet, due to power constraints)

Oh, wait - NASA already works with ion thrusters, particularly very-long-lived ones slated for SEP exploration missions.  They'll certainly have some desire to test them in space.  Looks like the current heir apparent is the NEXT thruster, which has ~4100s Isp.  A bit high compared to the 2500s on an example Snecma commercial Hall Effect thruster, since power needs scale with Isp^2, but not prohibitive.  The advantage here is that the commercial ones seem to be aimed at 10k hours of thrusting, while NEXT has been tested ground-side for a five-year-long burn, going through around 870kg of xenon, with little sign of degradation.  Aerojet Rocketdyne are the contractors.

So what kind of impulse is necessary?

Let's see, we can deduce the current drag force on the ISS using altitude, which is charted at http://www.heavens-above.com/IssHeight.aspx .  Let's take a (rare, it seems) month with no burns, November 2013 - the ISS lost 2.7km altitude in 30 days, 417.5km -> 414.8km.  Why doesn't this include both periapsis and apoapsis?  Isn't the thermosphere assymetrical, pulled out to one side by the solar wind?  Well, without one full solar cycle lacking burns, we'll have to make approximations anyway - so assume these are circular figures.  Calctool says  417.5km -> 414.8km takes one from 7662.64m/s -> 7664.17, a 1.53m/s difference.  I made up an independent spreadsheet for a conventional Hohmann transfer equation to see if this approximation had a valid basis (Web calculators have the merit of being hard to screw up with a typo usually), and the sanity check verified - 1.54m/s in two burns would correct for the month of November 2013's drag losses.  At such small periapsis:apoapsis ratios, this is negligibly different from a constant-thrust spiral-out dV for the same maneuver.  30 days is  2592000 seconds, and dividing, the acceleration felt on the station due to drag is 5.941 * 10^-7 m/s².  http://www.nasa.gov/mission_pages/station/structure/isstodate.html lists the ISS mass as 417,289 kg (slightly out of date, but a reasonable approximation, given the lack of new modules).  417289kg * 5.941 * 10^-7 m/s² = 0.2479 Newtons of drag force.

The specs for the NEXT thruster are as follows at full throttle:

* 6.9kW
* >=4170s Isp
* >=236mN
* 70% thruster and 95% PPU efficiency
* Total system mass of 58.1kg, including +-17deg 2 axis gimbal, power unit, feed system, and thruster
* 5.76mg/s propellant consumption
* Propellant consumption can be reduced arbitrarily at the cost of slight reductions in efficiency with roughly the same Isp, down to a deep throttle of 35% thrust
* More thrust is not feasible, mass flow is capped rather than tradeable with impulse as in some other ion thrusters

It works without issue in an array, potentially with added redundancy due to neutralizer which can be used for adjacent thrusters, and it has passed extensive flight qualification testing.

What about the sanctity of the microgravity?  The ISS is a 'microgravity research lab', right?

Several points:

At low frequency, the microgravity is compromised every time anything on the station impacts anything else that's attached in a load-bearing manner to the rest of the station.  Every time someone pushes off from one room to another, every time the exercise machine pumps iron, every time a manipulator arm moves, the rest of the station moves in the equal-momentum and opposite direction.

At much higher frequency, the noise of the station is always substantial, since absent convection-assisted diffusion, localized powered ventilation is required just to prevent people from choking on their own exhaled breath.  Ear plugs tend to be used at times.  Pressurized microgravity experiments feel that noise as a high-frequency oscillating acceleration.

The station's research basis, so far, is primarily engineering/technological development of human spaceflight.  Microgravity science experiments take on a decidedly secondary importance, as critics of human spaceflight have noted.  One of the selling points of commercial crew, remember, was this: "A seventh crew member could potentially add about 33 hours per week to the current amount of crew time devoted to research — a 94 percent increase," according to an IOG audit.  There is hope that we'll find economically useful things to do on the station, but so far the research done is rarely even scientifically useful, outside the field of keeping a group of humans alive and healthy in space.

The current strategy is to boost every few days to every few weeks, depending on availability of servicing missions, and subject the payloads to accelerations that are typically on the order of milli-G's.  These will vary wildly.  The Shuttle was using two of its 3870N RCS thrusters, but had to switch to it some number of its 6x 110N vernier thrusters when the station reached a certain degree of structural complexity.  Zvezda is equipped for 6000N for reboost.  Progress is supposed to use some number (6?) of its 28x 130N thrusters, but apparently on at least one occasion has used 4000N of thrust from two main engine boosts, with a gimbal oscillation problem which spooked the station with sudden extreme vibrations.  Needless to say, this disrupts microgravity.

A lower-intensity thrust (by a factor of 1,000 to 10,000) tailored to roughly correct the aerodynamic drag would eliminate these extreme accelerations, smoothing everything out.  It could also be used theoretically to precisely correct aerodynamic drag, actually improving the microgravity environment of the ISS objectively from all perspectives.

Consider the VASIMR test article: the VX-200 will supposedly generate up to 5N of thrust, 20x the necessary drag correction level, and it will do it in a useless manner (burning normal and anti-normal) to change inclination, at a maximum duty cycle of something like 1:80 (2.5kW/200kW), far less than the 1:20 that would be necessary to compensate for drag at this power level.  That impulse is something the microgravity labs will feel.  Nonetheless, NASA thinks it's worth it to test on the ISS, they evidently think this level of acceleration disturbance will be 'lost in the noise' for their science.  Of perhaps even more import, the test module will be pointed deliberately off-axis in order to measure induced rotation in the station, as a means of validating thrust - and rotation will generate both steady-state and transient loads on microgravity experiments.

In examining the case, I had a thought - what about the rotation of the ISS itself?  It turns out, the full ISS is designed to be maintained in an 'XVV' rotation, with those two end adapters pointed prograde and retrograde, at all times.  That enables items like Earth observation cameras or the cupola, to remain pointed at the ground.

The ESA user guide to the ISS claims:

"The overall ISS design is optimized to fly in the XVV attitude for the following reasons:
*It provides the best microgravity conditions
*It supports attitude re-boosts
*It supports service vehicle docking
*It minimizes aerodynamic drag

Item #1, though, seems facially wrong, unless I'm missing something - rotating the ISS to track the Earth vector will necessarily induce some centripetal acceleration.  How much rotation?  Another web calculator to the rescue!  The ISS is 79m long without Progress or Soyuz docked on the ends.  The ISS orbital period is ~93 minutes, so it must spin at 1/93 = 0.0108 rpm.  This should induce a rotation, felt at the docking/berthing ports at the ends of the X-axis, of 0.0000505 m/(s^2), or 5.2 millionths of a G, and less over the rest of the spacecraft.  Unlike thrusting, this is a *steady* acceleration, a permanent deviation from zero gravity.  This swamps the drag force (or the steady-state thrust required to fight it) by a factor of 100, and completely invalidates criticism on the basis of microgravity disturbance, even if you consider the other points invalid.

But what about location?  Where do we put it?

These high-thrust reboost vehicles are located at the end of the pressurized spine of the spacecraft, with a direct load path down the whole vehicle and a natural orientation towards the center of the spacecraft.

The way thrusting against an object works, the average thrust vector has to be pointed directly away from the center of mass of the spacecraft, or else it will actually induce a rotation on the object, spinning it up.  The Shuttle can't be mounted precisely on-axis, and so it had to use thrusters all over the spacecraft pointed in other directions to correct for rotations (this caused other problems with the potential for solar panels blocking some of those orthogonal thrusters, and it's a bit of a messy hack).  There is some ability to absorb rotations into zero-propellant reaction wheels aboard the ISS, but these are limited (by their maximum rotation speed before they break up, the 'saturation' point).

So if you have to mount the thruster to exactly two ports on the ISS for it to work...

Ahh, but that's not the case.    Two ports on the ISS are pointed *roughly* at the center of mass, true, but this is not necessary or sufficient (they still need to gimbal to hit the precise spot).  The average thrust vector has to be pointed directly away from the CoM, not the thrusting module itself.  This still gives the freedom of mounting anywhere in relation to the CoM, so long as the thrusters are free to rotate to point in the correct direction, and there's nothing blocking them, and the mounting point is solid enough to absorb the thrust without undue bending loads.

For an ion thruster that only generates 250mN, the stresses are *so* tiny in relation to any reasonable load - about 1 ounce of weight on Earth - that there is almost no restriction on bending loads, they could be mounted anywhere on the station that they are free to point away from the CoM, and thrust prograde (as a vehicle, though the actual thruster is aimed at the retrograde).

But they do have to have a free field of view in the retrograde direction.  You can't be pointing the exhaust of an ion thruster at another module and expect it to work.  Additionally, the solar panels and radiators are going to complain if you push 1 ounce of force on them in a direction they're not strong in.  The hard point is that the field of view will rotate around the ISS over the course of an orbit.

So - we need a gimbal anyway, why not mount a thruster like this at the end of a robotic arm?  Anything designed to re-position ten-ton payloads in reasonable time-frames is going to require enough strength and torque to deal with a constant 250mN force at its end effector.   At the end of a robotic arm, a thruster can move around the spacecraft to wherever it has a free FOV, in the intersection of the plane of orbit (it has to point retrograde, but we get to control rotation of the station, adding one DoF).

The X-axis, the pressurized habitable 'spine' (fun word), of the ISS, is kept roughly aligned with the velocity vector.  There are several solutions that come to mind for mounting permanent thruster modules that aren't precisely on that spine.  Option A is simply maintaining an orientation that is offset a few degrees (radial-out direction) from precise velocity vector, and mounting a single ion thruster someplace on the X axis modules, exhaust pointed retrograde.  Anywhere along the XZ plane is fine, though the spacecraft will likely re-orient to precisely line up those points during docking ops, causing a slight wasted inclination component.  Option B is even more flexible, though.  Mount the propulsion as two thrusters, on two arms.  Maintain them parallel, and the averaged thrust vector can still intersect the CoM, but you can put them almost anywhere on the station they're not bumping into the SAWs.  The optimal site for option B, I think, is actually on the ends of the unpressurized truss.  The Alpha Joints do rotate the panels somewhat, but a robotic arm can correct for this.

What about attitude control?  The reaction wheels have to be desaturated occasionally?

The robotic arm just has to position its moment arm and point so that it's off-CoM, and this is easily achieved.

But emergency collision avoidance!
...Is based on quite low risk factors.  According to this, the risk that triggers one is a 1 in 100,000 to 1 in 10,000 probability.
...Requires 30 hours of advance planning to execute
...Typically, the flight team at NASA has three or four days' warning of a conjunction, and if given only ~18hrs, have decided to climb into escape pods rather than burn, in the past.
...Is rare.  16 maneuvers in 15 years
...Is a burn of order ~1-3m/s dV, similar to one of the larger stationkeeping burns, and is usually applied prograde, functioning as an 'early' stationkeeping burn

The thing is, they don't require enormous amounts of thrust, just substantially more than an ion thruster can easily provide.  There is great synergy in a hybrid approach, where the ion thruster and small chemical thrusters are used together.  The difference from the status quo is that fuel is not constantly being sent to the station, because emergency burns are much less common than orbit raising burns, and that the ion thruster can decrease this even further by adjusting (simply by turning off and succumbing to drag if necessary) the orbit for any debris that might pose a threat with more than a week or two notice.

So a chemical thruster is synergistic with an ion thruster.  What sort of chemical thruster then?  What prevents this from requiring a large new propulsion module?

Whether the station burns at 200N for 30 minutes or 2000N for 3 minutes is almost immaterial if followed by a day or two of 'coasting' out of the dangerous phase space.  The 200N option (or perhaps the 20N option) is *much* easier to mount to the station - you can't stick it on the end of a big robotic arm, but mounting option A, anywhere on the X-axis modules, is fine - the station can rotate to thrust prograde, so long as there is gimbal range on the thruster.  There is some number of newtons that are acceptable for any mounting location on the station - you just have to size the thrusters to be that small, and you won't require a whole CBM, structural reinforcement, etc.  You can also, in an emergency where you can't even rotate the gimbal so it can thrust on-axis, just stomach a bit of extra rotation, eating it up with the reaction wheels and then desaturating them efficiently using an ion thruster (or pair of thrusters) later.

There also remains the likely possibility that the ISS will not be separated, and Zvezda will remain for emergency burns indefinitely, or that the thrusters on the US-owned Zarya module could be used on a provisional basis for emergency collision avoidance despite the fact that they are near their EOL.  1-2m/s in CAM burns a year is a lot less stress on them than ~40m/s a year of stationkeeping at the old 350km altitude (the reason we switched to Zvezda/Progress/ATV's engines), or ~19m/s a year of stationkeeping at the new altitude that would be performed using an ion thruster.

We need more power
I believe the ISS has 7kW to spare, frankly. 2-3kW may be all they can spare for the VASIMR experiment if it will produce no useful propulsion, but this would be the *primary propulsion mode* of the whole station, and could be easily prioritized.

Right now, some portion of the US-owned Solar Array Wing power production is slated for the Russian Orbital Segment, since they designed an elaborate supplementary solar array in the ISS planning process, and they never flew that module.  In the event of detachment, the USOS would gain additional usable power.

The ISS presently adjusts its solar array wing orientation separately from the rest of the spacecraft using a complex optimization process designed to mitigate atmospheric drag.  Once there is SEP available, though, power production compromises here translate directly into less thrust available to correct drag.  This gives generous wiggle room to produce a little bit more power, at the cost of slightly increased drag, but to completely correct for that drag with the ion thruster, using the extra power.

But let's assume pessimistically that none of this applies.  How much extra power equipment is required in order to have no impact on the station?

The status quo on the ISS is the use of nickel-hydrogen batteries, which have an energy density of 75Wh/kg and a specific power of 220W/kg.  For 45 minutes (the orbital night) of storage at 7kW, that's 80kg necessary for the energy or 32kg necessary for the power.  Let's arbitrarily double this, to ensure a manageable depth of discharge, to 160kg of batteries per thruster.

NASA has been working on large 200kW/wing space power systems for SEP, with two entrants, megaROSA and megaFlex, both claiming better than 200W/kg specific power, who are in the process of scaling up their designs.  If average power needs during orbital day are 14kW, that's 70kg of solar array per thruster  Let's arbitrarily double this to account for the aging of the panels, and then increase it by 50% more to account for compromises which reduce drag over part of the orbit, to 210kg of solar array per thruster, or ~42kW.  ATK's first-stage demonstrator for their 200kW MegaFlex solar array wing grant, an extension of their UltraFlex line, produces 20kW, has 10m diameter, deploys in pairs to provide 40kW total power, in the same vicinity as our needs, and was built and validated earlier this year.  Actual mass I cannot locate, but let's say for the sake of argument it's only achieved 100W/kg - 400kg for one pair, to serve one 7kW active thruster.

So 560kg power equipment per active thruster.

How much fuel would this use & save?
NASA says , of raising the ISS from 350km to ~400km, "At its current altitude, the space station uses about 19,000 pounds of propellant a year to maintain a consistent orbit. At the new, slightly higher altitude, the station is expected to expend about 8,000 pounds of propellant a year."
Over a 5 year baseline, which the thruster has been tested to, that's 18.1 tons of fuel... but the solar panels will last much longer, and a new thruster is inexpensive in terms of mass (to the point that flying several on the same end effector is entirely practical), so assume this saves you about 1 of the 3300kg capacity Dragon cargo missions per year, indefinitely.

The fuel used for the ion thruster would be xenon, 5.76mg/s per thruster in the default mode of operation, or 181.8kg per year.  Add in some tank mass, and we'll say 1 ton of fuel lasts you 5 years exactly.

How heavy would all this be?

As noted, power generation and storage is covered on ~560kg per ion active thruster
One ion thruster (lasting 5 years) weighs ~60kg.
5 years of fuel per thruster weighs ~1000kg.
The upcoming European Robotic Arm is the most convenient one to choose, as there is an active team on it.  It weighs 630kg per arm

Let's baseline two arms, with four NEXT thrusters each.  This gives plenty of margin for failures, for solar weather, for one thruster failing, et cetera - more thrusters are just cheap, and 8 should guarantee a lifetime in excess of a decade.  Each arm gets 1 ton of fuel, for a system lifetime under current conditions of 10 years, with the possibility of refueling later.  At the base of each arm is a new attachment point for the outer S6 and P6 truss sections (one arm at each end), and a small ~100kg (WAG) structural module holding batteries and and a pair of new ATK MegaFlex 10m solar wings, for a total of 80kw of additional power at launch, potentially.  At the far end of the arm is the quad ion thruster array (let's call it ~20kg of structural mass plus 240kg of thruster assemblies), with only one thruster expected to be in operation at any one time (and probably at half power).  Whether the xenon tank should go at the end of the arm right next to the thrusters, or at the base, is unclear, though the arms are capable of much bigger payloads.

So - with very conservative assumptions, this whole system would be on the order of 2650kg * 2 modules, and would last with considerable redundancy for 10 years of stationkeeping & attitude control.  This does not incorporate a system for emergency Collision Avoidance Maneuvers, but in the event of ISS separation and SEP implementation, a very modest one would serve - a few low power thrusters and a ton and a half of propellant per 10 years on station would net a 1m/s burn once a year.  The fuel on in Zarya's tanks will serve initially - just add new, very low power vernier thrusters.

The whole thing could be launched in perhaps 2 Dragon Cargo vehicles, mass-wise, and we'd be good for a decade.

[Burninate]

-reserved for addendums-

Addendum1:
Xenon is expensive
Yes.  Google claims $1200/kg.  Add the launch price of ~$10,000/kg to ~$30,000/kg (estimates vary) for cargo to the ISS onboard Dragon, though.  Now note that you need 1/15th as much propellant to produce the same impulse, if using NEXT over hydrazine.  Now remember that Dragon is the cheap option.  Krypton or Argon are less expensive, but, for now - does it matter?

Two thrusters, one net thrust vector
Using a two-thruster solution at opposite ends of the Y axis unpressurized truss can free one from worrying about anything in the way, as normal reconfiguration operations go on on the interior of the ISS.  It can also perform a fun trick where the thrusters are aligned 30-60 degrees away from retrograde orientation , in opposite directions (fighting each other), to sustain the necessary thrust at the necessary retrograde vector away from the CoM, without affecting any craft that might want to approach and dock at the retrograde ports.  The cost is temporary cosine losses.

[Jim]

"This swamps the drag force (or the steady-state thrust required to fight it) by a factor of 100, and completely invalidates criticism on the basis of microgravity disturbance, even if you consider the other points invalid."

wrong, because the lab module (experiments) are not at the ends of the station.  The lab is near the CG of the station for that specific purpose.

"So - we need a gimbal anyway, why not mount a thruster like this at the end of a robotic arm?  Anything designed to re-position ten-ton payloads in reasonable time-frames is going to require enough strength and torque to deal with a constant 250mN force at its end effector.   At the end of a robotic arm, a thruster can move around the spacecraft to wherever it has a free FOV, in the intersection of the plane of orbit (it has to point retrograde, but we get to control rotation of the station, adding one DoF)"

not feasible.

The arm brakes are not designed for constant force.
The arms don't have the positional accuracy
Don't want to spray the ISS with SEP exhaust
Arms don't have the reliability for this
The thrust is not enough to destaturate the gyros.
There is no control capability for the ERA on the USOS.
there is no wiring for the ERA on the truss
the ERA is not the SSRMS, it does not move modules.


"I believe the ISS has 7kW to spare, frankly. 2-3kW may be all they can spare for the VASIMR experiment if it will produce no useful propulsion, but this would be the *primary propulsion mode* of the whole station, and could be easily prioritized.

"I believe" doesn't cut it.   Nor does taking power away from the rest of the station viable

Zarya can not be used for any propulsive requirements.  Those systems were deactivated.  Also, the main thrusters are in the wrong direction

Since DAM's are required and need higher thrust than and SEP can provide, it is just easier to maintain chemical rockets and their associate logistics.

Your estimates for one module per Dragon flight are unrealistic.  There is no accounting for packaging and attached hardware in the Dragon. 

[Jim]

I'm going to present it in a bit of a stream of consciousness manner, a prompting question per section and a paragraph per answer, to show my thinking in as much detail as possible and perhaps deflect both Jim's nitpicking and questions that have already been answered:

What nitpicking?  When you are wrong, you are wrong, as pointed out many times

[ncb1397]

Not sure what the context of the posts above is, but it seems to me that SEP could create a more pure microgravity environment by counteracting the drag force that the ISS experiances on a continual basis. So, say the ISS experiances 1 Newton of drag force. A tiny sep thruster would apply an opposite 1 Newton of force maintaining the station's position and reducing marginally the G-force that experiments are subjected to via atmospheric drag. Not sure it would help much given all the other stuff moving on the station(robotic arms for example) but it certainly wouldn't hurt zero-g research and also mantains altitude.

[Burninate]

"This swamps the drag force (or the steady-state thrust required to fight it) by a factor of 100, and completely invalidates criticism on the basis of microgravity disturbance, even if you consider the other points invalid."

wrong, because the lab module (experiments) are not at the ends of the station.  The lab is near the CG of the station for that specific purpose.

While it's true that the other modules do not feel the same force as  the 5.941 * 10^-7 m/s² that the station feels now due to drag, did you notice "factor of 100"? For centripetal acceleration to be exceeded by a ~0.25N translational force, the center of gravity would have to be less than 46cm away from the measurement point.  Again, this translational force is being felt on the ISS *today*, due to drag, and an ion thruster would *counteract* that force ideally.

"So - we need a gimbal anyway, why not mount a thruster like this at the end of a robotic arm?  Anything designed to re-position ten-ton payloads in reasonable time-frames is going to require enough strength and torque to deal with a constant 250mN force at its end effector.   At the end of a robotic arm, a thruster can move around the spacecraft to wherever it has a free FOV, in the intersection of the plane of orbit (it has to point retrograde, but we get to control rotation of the station, adding one DoF)"

not feasible.

The arm brakes are not designed for constant force.
The arms don't have the positional accuracy
Don't want to spray the ISS with SEP exhaust
Arms don't have the reliability for this
The thrust is not enough to destaturate the gyros.
There is no control capability for the ERA on the USOS.
there is no wiring for the ERA on the truss
the ERA is not the SSRMS, it does not move modules.

The ERA is rated for 8000kg at 100mm/s tip speed, with 5mm/s tip positional accuracy.  The ERA's reach is 9.7m.  To accelerate an 8000kg payload from stable at one end of its reach to a maximum speed at the center to stable at the other, the ERA needs to be able to accelerate 8000kg to 0.1m/s in 4.85m span.  To do perform this acceleration in a distance of only 4.85m with constant acceleration, requires a figure for acceleration that can be shown to be (v^2)/2d, or 0.001031m/s^2.  The acceleration requires 97 seconds.  0.001031m/s^2 * 8000kg = 8.25N, a force 17x as large as we are discussing.  So if the ERA can move a mass of ~500kg around and reach 100mm/s at some point, that's functionally equivalent.  If it can reach 0.1m/s in less space, that's also substantially greater acceleration, and more design torque implied.

The ERA was selected as an arbitrary example of a robotic arm, however, with a team actively working on it which could likely increase production to 3 units without much problem.   It's not necessary to use this particular model if it's inadequate, this is effectively a free variable.  Pick any of the other robotic arms on the station and duplicate them if you like, design your own, or design a strong gimbal on a long rigid extending mast - the point is the problem is entirely tractable.  Something to bring the needed thrust vector out of the path of the rest of the station without unduly affecting operations is a solveable problem, particularly if the thrust is split between two units.

Any module we can mount on the station is going to require new wiring for control & capability.  Tractable problem.

Why is 250mN-500mN thrust not capable of desaturating the CMGs, if pointed in the proper direction?  How do you imagine desaturating the gyros works?

"I believe the ISS has 7kW to spare, frankly. 2-3kW may be all they can spare for the VASIMR experiment if it will produce no useful propulsion, but this would be the *primary propulsion mode* of the whole station, and could be easily prioritized.

"I believe" doesn't cut it.   Nor does taking power away from the rest of the station viable

See redundant points if you're skeptical on this one.

Zarya can not be used for any propulsive requirements.  Those systems were deactivated.  Also, the main thrusters are in the wrong direction

Since DAM's are required and need higher thrust than and SEP can provide, it is just easier to maintain chemical rockets and their associate logistics.

"Were deactivated".  Not "Were deorbited".  The main thrusters can be pointed in the right direction by rotating around the Y axis temporarily into a configuration that is offset from but fixed in relation to XVV, a maneuver which does not appear to be uncommon - it's necessary to rotate by 180 degrees anytime the station wants to switch between XVV- and XVV+ to support burns or docking. Rotating to, say, XVV plus 90 degrees, for a brief burn, is something that's feasible for a DAM.

The constraint seems to be that Soyuz/Progress can't be connected to Rassvet or Pirs, and the SAWs must be rotated correctly.  This is very inconvenient, though in an emergency with days of preparation, many things are possible.

See also redundant points - adding new thrusters to connect to Zarya's tanks requires some novel plumbing, but is quite low-mass-requirement...  ISS separation is not that likely to happen, allowing Zvezda and Progress emergency reboosts... etc, etc, etc.

Your estimates for one module per Dragon flight are unrealistic.  There is no accounting for packaging and attached hardware in the Dragon.

This is fair.  I didn't bother to quantify volumetric payload envelopes at this time, but again - tractable problem.  Selection of Dragon was as arbitrary example of the scale of the mass required.

[IslandPlaya]

I for one, welcome the prospect of the ion-thrusters on robotic arms ISS overlord...
Burninate has made a convincing case IMHO. I'm sure there are problems. Point them out and we can see if they can be worked.
Appreciate the detail in the OP. Made good reading.

[RonM]

It sounds like a good idea to me, but I'll let the experts and engineers discuss whether or not it really is a good idea.

I do appreciate all the effort Burninate put in the OP. It's nice to be discussing an idea with numbers to back it up, not just someone's opinion.

[TrevorMonty]

I thought the small G forces from applying thrust continuously would effect the microgravity environment, but as Burnirate pointed out it just counters the deceleration effects of atmospheric drag. The result may be a reduction in microgavity not an increase.
Still going to need chemical thrusts for emergency maneuvers.

[Jim]

1.  The ERA was selected as an arbitrary example of a robotic arm, however, with a team actively working on it which could likely increase production to 3 units without much problem. 

2.   It's not necessary to use this particular model if it's inadequate, this is effectively a free variable.  Pick any of the other robotic arms on the station and duplicate them if you like, design your own, or design a strong gimbal on a long rigid extending mast - the point is the problem is entirely tractable.  Something to bring the needed thrust vector out of the path of the rest of the station without unduly affecting operations is a solveable problem, particularly if the thrust is split between two units.

3.  Any module we can mount on the station is going to require new wiring for control & capability.  Tractable problem.

4.  Why is 250mN-500mN thrust not capable of desaturating the CMGs, if pointed in the proper direction?  How do you imagine desaturating the gyros works?

5.  See redundant points if you're skeptical on this one.

6.  "Were deactivated".  Not "Were deorbited".  The main thrusters can be pointed in the right direction by rotating around the Y axis temporarily into a configuration that is offset from but fixed in relation to XVV, a maneuver which does not appear to be uncommon - it's necessary to rotate by 180 degrees anytime the station wants to switch between XVV- and XVV+ to support burns or docking. Rotating to, say, XVV plus 90 degrees, for a brief burn, is something that's feasible for a DAM.

7.  See also redundant points - adding new thrusters to connect to Zarya's tanks requires some novel plumbing, but is quite low-mass-requirement...  ISS separation is not that likely to happen, allowing Zvezda and Progress emergency reboosts... etc, etc, etc.

1. Wrong.  ERA production was finished long ago.  The manufacturing team is gone.  So, it is a problem

2.  All arms on the station are springboards and are ill-suited for such a task.  They are not rigid and not meant to hold a constant force.   There will be flexing and having two of them makes it worse, there will be control issues.    The problem is not "tractable", it cause more problems and solves none.

3.What new modules?  Anyway, modules are connected from the inside.   More overuse of the word tractable and actually is handwaving away problems.  It is quite the opposite.  ISS was scarred for the existing arms.   If you are going to used the word, have some knowledge to back it up. 

4. Too low of thrust.  Need higher thrust and shorter time periods. 


5.  What points?  You haven't shown that the ISS has excess power nor is willing to give up power.  Giving up power is counter productive, it reduces the reason for the ISS existing.

6.   "Were deactivated" as in never can be used again, not just turned off.  The FGB was never intended for reboost.  And it is idiotic to think that "wrong direction" in the context of this discussion means the thrust vector as it relates to direction of flight.  The FGB thrusters are in the "wrong direction" because they are pointed at the ISS structure and would impinge on it. 

7.  If Zvezda and Progress are still in use, then SEP for reboost is a waste of resources.

[IslandPlaya]

Thanks Jim! Phew!
Need more details though

[Burninate]

1.  The ERA was selected as an arbitrary example of a robotic arm, however, with a team actively working on it which could likely increase production to 3 units without much problem. 

2.   It's not necessary to use this particular model if it's inadequate, this is effectively a free variable.  Pick any of the other robotic arms on the station and duplicate them if you like, design your own, or design a strong gimbal on a long rigid extending mast - the point is the problem is entirely tractable.  Something to bring the needed thrust vector out of the path of the rest of the station without unduly affecting operations is a solveable problem, particularly if the thrust is split between two units.

3.  Any module we can mount on the station is going to require new wiring for control & capability.  Tractable problem.

4.  Why is 250mN-500mN thrust not capable of desaturating the CMGs, if pointed in the proper direction?  How do you imagine desaturating the gyros works?

5.  See redundant points if you're skeptical on this one.

6.  "Were deactivated".  Not "Were deorbited".  The main thrusters can be pointed in the right direction by rotating around the Y axis temporarily into a configuration that is offset from but fixed in relation to XVV, a maneuver which does not appear to be uncommon - it's necessary to rotate by 180 degrees anytime the station wants to switch between XVV- and XVV+ to support burns or docking. Rotating to, say, XVV plus 90 degrees, for a brief burn, is something that's feasible for a DAM.

7.  See also redundant points - adding new thrusters to connect to Zarya's tanks requires some novel plumbing, but is quite low-mass-requirement...  ISS separation is not that likely to happen, allowing Zvezda and Progress emergency reboosts... etc, etc, etc.

1. Wrong.  ERA production was finished long ago.  The manufacturing team is gone.  So, it is a problem

You have a point I suppose - perhaps cloning ERA would be no easier than cloning Canadarm2 or designing a novel arm for this purpose, I don't know.  ERA was merely a choice of perceived convenience.

2.  All arms on the station are springboards and are ill-suited for such a task.  They are not rigid and not meant to hold a constant force.   There will be flexing and having two of them makes it worse, there will be control issues.    The problem is not "tractable", it cause more problems and solves none.

There is some amount of constant force they will hold.  I do not see how it could be less than 250 milliNewtons.  I do not see how if they are designed to move around 8 tons, 250 milliNewtons could make them flex in prohibitively unpredictable, or structurally damaging ways.  I also do not see how their flexing might be impossible to adjust for in the software's control loop.

3.What new modules?  Anyway, modules are connected from the inside.   More overuse of the word tractable and actually is handwaving away problems.  It is quite the opposite.  ISS was scarred for the existing arms.   If you are going to used the word, have some knowledge to back it up. 

Attachment points is something I don't have a lot of knowledge on, which is why I have been assuming (for attachment option A only) that some kind of rigid attachment point on the habitat modules existed, and for option B, with two arms, that some structural overhead might be required at the ends of the unpressurized truss for attachment.

Turns out I was right, from WP:
"Officially known as the Space Station Remote Manipulator System (SSRMS), it is self-relocatable and can move end-over-end to reach many parts of the Space Station in an inchworm-like movement. In this movement, it is limited only by the number of Power Data Grapple Fixtures (PDGFs) on the station. PDGFs located around the station provide power, data and video to the arm through its Latching End Effectors (LEEs). The arm can also travel the entire length of the space station truss using the Mobile Base System."

So mounting something directly to a pressurized module, at least, turns on whether there are at least some spare PDGFs available, which are not necessary for the existing arms to use during normal operations.  If you could illuminate me on that availability, I would appreciate it.

4. Too low of thrust.  Need higher thrust and shorter time periods. 

Again - for what?  What exactly is the restriction here?  Use numbers.  Is there some thrust-limited application I'm missing?  The station has been flipping around XVV- to XVV+ without propellant using the CGMs since 2006.

5.  What points?  You haven't shown that the ISS has excess power nor is willing to give up power.  Giving up power is counter productive, it reduces the reason for the ISS existing.

I haven't.  Which is why I first noted that the ROS detachment would add available power to the USOS, then pointed out that SEP would free up the existing solar array wings to increase power production by compromising less on low-drag-optimized orientation, then pessimistically, assuming neither of these were available, outlined a two-part addition to the power systems aboard the station which would increase power production by ~80kw (enabling the use of up to two active ion thrusters at EOL, expecting to need ~1) and provide an opportunity for NASA to test new solar array technology.

If not A, then B.  If not B, then C.  If not C, then D.  You've got me on A - I have no evidence to back that up other than the amount of power they're willing to sacrifice for the VASIMR experiment.  So go on to B, C, and D.

6.   "Were deactivated" as in never can be used again, not just turned off.  The FGB was never intended for reboost.  And it is idiotic to think that "wrong direction" in the context of this discussion means the thrust vector as it relates to direction of flight.  The FGB thrusters are in the "wrong direction" because they are pointed at the ISS structure and would impinge on it. 

I expect you have more expertise than me on this, which is why this is so frustrating.  What module would the Zarya thrusters impinge on, in the event of ISS separation?  I suppose I made a mistake here talking about Progress being in the way, because *Pirs and Rassvet and Zvezda wouldn't be part of the station anymore* in the contingency we're discussing.

7.  If Zvezda and Progress are still in use, then SEP for reboost is a waste of resources.

Less fuel mass sent to the station is more other things sent the station, or fewer Progress flights.  If need for Zvezda diminishes to 1m/s per year instead of ~20, its engines will last longer and the ISS might not require external reboost capability at all.

[Space Ghost 1962]

4.  Why is 250mN-500mN thrust not capable of desaturating the CMGs, if pointed in the proper direction?  How do you imagine desaturating the gyros works?

4. Too low of thrust.  Need higher thrust and shorter time periods. 

Hold on. Too much shorthand here.

Yes you can design a electric thrust system that can desaturate CMGs. It's done on other spacecraft with much less thrust.

No its unlikely it can be retrofitted to the ISS.

The strategy for desaturation is different. With few large thrusters you want short bursts. So you can keep the nutation effects to a minimum. But actually for longer lived CMG's, more frequent smaller pulses result in less wear.

You can actually do an entire ACS with many thrusters and no CMGs. You use the spacecraft thrust structure as a spring, and use chaotic coupling knowing the deformation of the structure, Can be done even with relatively fast slewing for imaging too, maintaining a rock solid positioning in an inertial frame.

Requires considerable FEM examination of the problem before understanding if the problem is tractable.

It's been proven on small spacecraft, and in simulation on large space telescopes.

Saying it can't work on a space station isn't right. It wouldn't surprise me if it could. And provide a lower cost, better microgravity environment.

Might be an interesting project for IR&D funding, if not being done already. Not ready for operation.

For understanding other strategies in ACS that are related:
A Brief Survey of Attitude Control Systems for Small Satellites using Momentum Concepts

[Burninate]

The strategy for desaturation is different. With few large thrusters you want short bursts. So you can keep the nutation effects to a minimum. But actually for longer lived CMG's, more frequent smaller pulses result in less wear.

You can actually do an entire ACS with many thrusters and no CMGs. You use the spacecraft thrust structure as a spring, and use chaotic coupling knowing the deformation of the structure, Can be done even with relatively fast slewing for imaging too, maintaining a rock solid positioning in an inertial frame.

Requires considerable FEM examination of the problem before understanding if the problem is tractable.

It's been proven on small spacecraft, and in simulation on large space telescopes.

Saying it can't work on a space station isn't right. It wouldn't surprise me if it could. And provide a lower cost, better microgravity environment.

Might be an interesting project for IR&D funding, if not being done already. Not ready for operation.

So let's work backwards - I'm having trouble grokking this.  If FEM examination of the problem says it's not tractable...  what does that mean?

Forget more or less frequent pulses - an ion thruster would presumably operate at steady-state thrust, with no vibration.  ~250mN is the current force on the station due to aerodynamic drag.  This is ~1 ounce of weight equivalent - not enough to snap a rubber band.  The default mode of operation would be to fire up one thruster on each side at 50% power (50% mass flow rate), generating a combined thrust of about 250mN, matching that drag force.  With the full baseline incorporating additional power, a reserve capacity to turn things up to 500mN steady-state is expected even after considerable age-related degradation in the solar panels - or even more if you're willing to run more than one ion thruster at a time (there's enough power at launch for ~1N thrust from four ion thrusters for the full duration of orbit, or ~2N thrust from 8 during the day, but this goes unused in my planning).

This 250-500mN can be pointed in any direction, can be applied asymmetrically, et cetera, without changing the station's XVV orientation - just move the arms around at the end.  The simplest *efficient* way to desaturate a gyro involves some combination of positioning the thrusters well away from the CoM and burning prograde as normal, or running only one of the sides, at full power, and using the gyros to 'fight' the moment that generates.

[Space Ghost 1962]

Love to get into it. But please do us all a favor and read the link provided.

It gives the essence of ACS operations, including the issues and requirements for desaturating CMG's.

Also a lot of innovative ways small sats deal with this. As well as the control theory.

Jim happens to be right here in the coarse grained obvious. But you need to educate yourself on why before you can go further.

As to the FEM model, lots of considerations have to be right. Otherwise you have too many thrusters in too many places. Which is a hint.

Suggest you put the torques and moment arms into consideration given the equations near the end of the paper. Its not that hard.

Think of the desaturation impulse as like fixing/grabbing the spacecraft while the gyro reacts to it. Simple fix.

Now imagine many carefully calculated "fixations" from thousands of different, damped load paths that add up to do the same.

You get them from many precisely timed precalculated thrusts, across the station. Phased carefully.

[Burninate]

Love to get into it. But please do us all a favor and read the link provided.

It gives the essence of ACS operations, including the issues and requirements for desaturating CMG's.

Alright, done - I appreciate the reference.  I still don't have the faintest idea what you're talking about. An ion thruster does in fact act as a 'mass expulsion system'.  I have described how assymetrical or off-axis thrust constitutes 'body-fixed offset roll/yaw thrusters'.

"The duration of the desaturation impulse is a function of the amount of momentum to be dumped. This is typically about one percent of nominal wheel momentum.ll For a 500-kg satellite in geosynchronous orbit. the required torque for desaturation is about 0.01 N· m-s. The number of thruster cycles expected over the lifetime of the satellite is a potentially limiting item. "

Thruster cycle life is not an issue for ion thrusters, particularly ion thruster controlled for yaw using CoM offset & orientation.  There's nothing here that suggests some minimum torque figure (and also, unit Newton Meter Second in this context is torque * time, not torque alone), only that the strategies that a 500kg spacecraft can use to employ conventional hypergolic RCS thrust blocks (tiny, high thrust, low cycle life, I guess non-gimballing / non-throttling devices) involve precision pulsing, and that the minimum duration for this pulse is set by thruster cycle life.  I don't see why that has to carry to this solution, which would employ fractional newtons at a very large number of seconds per day (all of them), and an adjustable number of meters moment arm, with no constraint on cycle life known.

Suggest you put the torques and moment arms into consideration given the equations near the end of the paper. Its not that hard.

Think of the desaturation impulse as like fixing/grabbing the spacecraft while the gyro reacts to it. Simple fix.

Now imagine many carefully calculated "fixations" from thousands of different, damped load paths that add up to do the same.

You get them from many precisely timed precalculated thrusts, across the station. Phased carefully.

Yes, I understand how desaturation impulses work, we're on the same page there, but that doesn't lead to this conclusion.  Differential throttle authority, gimballed attitude of the thruster and translation across the range of movement should provide abundant torque, and control over that torque, to desaturate the CMGs as the spacecraft flies.  It does not require precision pulses, only the precision to control the amount of torque, and sufficient mechanism to generate that torque using a moment arm in all three dimensions.  I thought I'd shown that.

[Todd Martin]

Burninate's OP is one of the best I've seen on NSF as far as laying out the technical requirements for a major ISS upgrade. 

One point that was never described is our ability to refuel or replace a Xenon tank.  Which is cheaper or easier?

The concerns over desaturation, oscillations, and robotic arm suitability are remarkable in that the answers are not already provided in a NASA study.  ISS has been in orbit since 1998, it is a shame no one has seriously looked at using SEP in all that time!  Perhaps this is because reboost came in part from Shuttle or that today these costs are mostly paid for by Russia.  Perhaps this is because ISS has lacked a long enough time commitment to justify the upgrade. 

[AnalogMan]

Here's an early study of SEP on ISS:

Electric Propulsion for International Space Station Reboost: A Fresh Look
January 2002

Abstract

Electric propulsion has recently been revisited for reboost of space station due to its high fuel efficiency. This paper focuses upon the propulsion system and orbit analysis trades undertaken at the beginning of a study to show the relative performance of potential electric propulsion system. A code was developed to analyze continuous low thrust reboost of space station with various electric propulsion systems at various power levels. Analysis showed that a major portion of reboost of space station can be made using electric propulsion systems with 0.5 N of continuous thrust. 1.0 N of EP thrust can provide almost the entire reboost mission. Three electric propulsion systems at various total power levels were chosen for further investigation: N2H4 arcjets at 5 kW, xenon Hall at 10 kW, and xenon ion thrusters at 20 kW. They were chosen for their ability to reduce the internationally launched chemical reboost fuel by 50% or more.

Also this note on the scope of the study:

"While the use of electric propulsion on space station can have other benefits such as microgravity environment enhancement and space station charge control, only the benefit of reduction in chemical reboost fuel was considered in this work."

[Space Ghost 1962]

Love to get into it. But please do us all a favor and read the link provided.

It gives the essence of ACS operations, including the issues and requirements for desaturating CMG's.

Alright, done - I appreciate the reference.  I still don't have the faintest idea what you're talking about.

Good. Here's another:

How often must the ISS desaturate its control moment gyros?

It does not surprise that you don't understand me. I am glad you are interested enough to struggle with the links I am feeding you. Having read them means we won't be talking by each other.

Please note the amount of momentum stored, the restrictions on the use of the CMG's in ACS, the location of the CMG's,  the current momentum distribution (e.g. moment arm) from the CMGs, and the pattern of desaturation dumps.

My point is that the station (and HSF vehicles in general) work on a different scale than smaller spacecraft. It is possible with the smallest of spacecraft to do ACS with the smallest of forces (radiation pressure, gravity gradient, magnetic, ...). They don't dock with multiton objects. Ever wonder about low impact docking system benefits among others?

An ion thruster does in fact act as a 'mass expulsion system'.  I have described how assymetrical or off-axis thrust constitutes 'body-fixed offset roll/yaw thrusters'.

Actually waste dumps can desaturate if used in this manner - just conservation of momentum.

"The duration of the desaturation impulse is a function of the amount of momentum to be dumped. This is typically about one percent of nominal wheel momentum.ll For a 500-kg satellite in geosynchronous orbit. the required torque for desaturation is about 0.01 N· m-s. The number of thruster cycles expected over the lifetime of the satellite is a potentially limiting item. "

Thruster cycle life is not an issue for ion thrusters, particularly ion thruster controlled for yaw using CoM offset & orientation.  There's nothing here that suggests some minimum torque figure (and also, unit Newton Meter Second in this context is torque * time, not torque alone), only that the strategies that a 500kg spacecraft can use to employ conventional hypergolic RCS thrust blocks (tiny, high thrust, low cycle life, I guess non-gimballing / non-throttling devices) involve precision pulsing, and that the minimum duration for this pulse is set by thruster cycle life.  I don't see why that has to carry to this solution, which would employ fractional newtons at a very large number of seconds per day (all of them), and an adjustable number of meters moment arm, with no constraint on cycle life known.

Small spacecraft can use continuous magnetic/gravitational/solar wind drag as well. Not unlike small fixed thrusters.

They might be able to work gradually on the extremes of a heavy station to the effect of desaturating the CMGs by countering the statistical torques that the CMGs are fighting to begin with. This is one way of using a smallsat technology on a huge station. Other approaches are even more esoteric I'm alluding to.

Suggest you put the torques and moment arms into consideration given the equations near the end of the paper. Its not that hard.

Think of the desaturation impulse as like fixing/grabbing the spacecraft while the gyro reacts to it. Simple fix.

Now imagine many carefully calculated "fixations" from thousands of different, damped load paths that add up to do the same.

You get them from many precisely timed precalculated thrusts, across the station. Phased carefully.

Yes, I understand how desaturation impulses work, we're on the same page there, but that doesn't lead to this conclusion. 

Good. In a nutshell, the conclusion is in applying chaos theory to transiently amplify the effect of multiple EP thrusters on the periphery of the station such that enough control authority is maintained for docking related desaturations that would otherwise require higher thrust, and doing this in a way that enhances the microgravity environment above current practice.

Mostly an application of esoteric math to the same control equations. In effect preloading certain "springs".

This is necessary to close the loop on all the cases for CMG stabilized ACS supplemented by EP. Then you'd see if you could design/test/simulate such a system, tell if it was cost effective (remember, multiple locations not gimballed), then a partial test on a portion of the ISS to confirm the dynamics, then a build/deploy, then a phase over between new/old looking for surprises/operational issues, then a phaseout and discard of prior systems. And this would be just ACS.

Differential throttle authority, gimballed attitude of the thruster and translation across the range of movement should provide abundant torque, and control over that torque, to desaturate the CMGs as the spacecraft flies.  It does not require precision pulses, only the precision to control the amount of torque, and sufficient mechanism to generate that torque using a moment arm in all three dimensions.  I thought I'd shown that.

You've suggested an approach yes. Just like I did. I based mine on an actual practice on another class of spacecraft, while yours was in the abstract. I then went back to the means by which the class it worked on was similar/different (theory/usage), found that this meant that adapting might mean working things differently, and slowly developed a process that might lead to a process and design that could be assessed and validated. You are taking the approach of attaching a thruster with a gimbal on an arm to supply a control input for the purpose of alleviating a issue, then scaling that to handle the need - this adds many new issues that may seem small one by one, but which works against the total system including the ISS as designed.

My suggested approach extends a known system, and while it does add demand to ISS (structural stress, relative stability), it is in a well known area with well known codes that can be assessed with confidence, especially the load paths.  My means of dealing with docking is to preload stress anticipating a docking, with the load paths the same as before, and follow the transients in real time. No plume impingement issues because of preplanned thruster positioning/angles.

I can't tell anything about your approaches load paths, plume impingement of thruster, interaction with center of mass, control bands and resulting control authority, especially during a docking manoeuvre.

And my approach is still too abstract for consideration. But because of the approach I've taken, it may allow others enough to consider it because they can see the potential "trade space" - the beginnings of a rational discussion.

Trying to be brief but not cryptic.

[IslandPlaya]

Burninate.
An invisbile man sleeping in your bed?
Who you gonna call?
Space Ghost 1962 busters!