Author Topic: NASA class MMOD as primary threat to commercial crew vehicles  (Read 34812 times)

Offline srtreadgold

  • Engineer
  • Member
  • Posts: 37
  • Houston
  • Liked: 30
  • Likes Given: 102

If that even works (signal-to-noise ratio issues abound here), what would that accomplish?  You look at it and go, "yep, that's a strike."  Then what?

If pieced into the main room duck tape a patch across the hole to stop the air escaping. Pipes can be sealed with a plaster. Broken windows can be covered.

Damage to outside parts of the spacecraft may have to be repaired by a robot or EVA. Such a repair was performed to the ISS solar panels on January 30, 2007. There may be a way to glue or weld heat shield material across a hole the size of Columbia's.

More likely IMO send a replacement vehicle up and use the damaged one only in the very unlikely case of a ISS evacuation. BTW if the pressure hull is compromised it is not only quite easy to fix but will inevitably be detected.

Nothing is "quite easy to fix" in space.

If the pressure shell is compromised that means there exists a hole in the back shell that would also need to be fixed. And these fixes would have to withstand reentry. 
« Last Edit: 08/30/2016 09:28 am by srtreadgold »

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6942
  • Germany
  • Liked: 1906
  • Likes Given: 1969
Nothing is "quite easy to fix" in space.

If the pressure shell is compromised that means there exists a hole in the back shell that would also need to be fixed. And these fixes would have to withstand reentry.

How do you get this idea? The backshell is very well protected by the trunk and protective coating. Anything penetrating the pressure hull will come from the side. A patch is easily applied from the inside.

Offline srtreadgold

  • Engineer
  • Member
  • Posts: 37
  • Houston
  • Liked: 30
  • Likes Given: 102
Nothing is "quite easy to fix" in space.

If the pressure shell is compromised that means there exists a hole in the back shell that would also need to be fixed. And these fixes would have to withstand reentry.

How do you get this idea? The backshell is very well protected by the trunk and protective coating. Anything penetrating the pressure hull will come from the side. A patch is easily applied from the inside.

The back shell is the side. You're thinking of the heat shield, which is mostly protected by the service module. The back shell also protects the vehicle from reentry gasses/temperatures.

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6942
  • Germany
  • Liked: 1906
  • Likes Given: 1969
The back shell is the side. You're thinking of the heat shield, which is mostly protected by the service module. The back shell also protects the vehicle from reentry gasses/temperatures.

OK, thanks.

Offline jgoldader

  • Full Member
  • ****
  • Posts: 728
  • Liked: 276
  • Likes Given: 155
A couple of thoughts-
-Using accelerometers to detect impacts might work above some threshold.  But how to remove the non-impact twitches?  (E.g., collant flowing, astronauts shifting in seats, etc.). You can't have the master alarm going off every 10 seconds.

-Triangulating sounds would probably be a nasty business of modeling, because the microphones are (presumably) attached to metal, and the sound waves would be propagating through the metal structures of the spacecraft.  You'd need to have some sort of acoustical model (edit: which could well be very different than the one used to analyze launch noise) and pinpointing a hit for detailed inspection might not be possible.  I could easily imagine saying, "Oh, we took a hit on the aft," but saying, "It's at PICA chunk 387" or whatever sounds like too much to ask.

-There was a good point made above, that just because the analysis says MMOD is the risk driving the LOCV stats doesn't mean it's that much worse than any number of others.  If MMOD is 1/1500, say, there may be a whole bunch at 1/1510.  We will almost surely not be told thedetailed numbers for proprietary reasons.  I doubt Contractor A wants Contractor B to know their numbers and be able to snipe at them in reports and future proposals.

-The two losses with which I'm most familiar, the shuttles, were from what I think of as known-knowns that turned out to be what I might describe as unknown-knowns (to coin a Rumsfeldian variant).  That is, both failure modes were recognized, but inaccurately characterized to the extent that they were judged as low risks, even though that did not turn out to be the case.  It seems to me that mischaracterized risk is different than a known-unknown.  Does that make sense?
« Last Edit: 08/30/2016 11:12 am by jgoldader »
Recovering astronomer

Online A_M_Swallow

  • Elite Veteran
  • Senior Member
  • *****
  • Posts: 8693
  • South coast of England
  • Liked: 404
  • Likes Given: 184

If that even works (signal-to-noise ratio issues abound here), what would that accomplish?  You look at it and go, "yep, that's a strike."  Then what?

If pieced into the main room duck tape a patch across the hole to stop the air escaping. Pipes can be sealed with a plaster. Broken windows can be covered.

Damage to outside parts of the spacecraft may have to be repaired by a robot or EVA. Such a repair was performed to the ISS solar panels on January 30, 2007. There may be a way to glue or weld heat shield material across a hole the size of Columbia's.

Duct tape / plaster is not exactly what you want keeping reentry gasses/temperatures from breaching your delicate spacecraft.

Repairs have to defend against:
a. vacuum
b. and reentry.

Spacestations do not reenter so they just need defending against vacuum.

The outside of a capsule will need additional heat shield material to defend it or will have to sit around in space until a replacement covering can be launched and fitted.

Filling holes in Space Shuttles was investigated, I do not know if any of that knowledge can be adapted for the new capsules.

Online A_M_Swallow

  • Elite Veteran
  • Senior Member
  • *****
  • Posts: 8693
  • South coast of England
  • Liked: 404
  • Likes Given: 184
A couple of thoughts-
-Using accelerometers to detect impacts might work above some threshold.  But how to remove the non-impact twitches?  (E.g., collant flowing, astronauts shifting in seats, etc.). You can't have the master alarm going off every 10 seconds.

-Triangulating sounds would probably be a nasty business of modeling, because the microphones are (presumably) attached to metal, and the sound waves would be propagating through the metal structures of the spacecraft.  You'd need to have some sort of acoustical model (edit: which could well be very different than the one used to analyze launch noise) and pinpointing a hit for detailed inspection might not be possible.  I could easily imagine saying, "Oh, we took a hit on the aft," but saying, "It's at PICA chunk 387" or whatever sounds like too much to ask.

-There was a good point made above, that just because the analysis says MMOD is the risk driving the LOCV stats doesn't mean it's that much worse than any number of others.  If MMOD is 1/1500, say, there may be a whole bunch at 1/1510.  We will almost surely not be told thedetailed numbers for proprietary reasons.  I doubt Contractor A wants Contractor B to know their numbers and be able to snipe at them in reports and future proposals.

-The two losses with which I'm most familiar, the shuttles, were from what I think of as known-knowns that turned out to be what I might describe as unknown-knowns (to coin a Rumsfeldian variant).  That is, both failure modes were recognized, but inaccurately characterized to the extent that they were judged as low risks, even though that did not turn out to be the case.  It seems to me that mischaracterized risk is different than a known-unknown.  Does that make sense?

Triangulation.

The speed of sound in aluminium is 6320 m/s.

Triangulation can work by differences in volume and the delay between sensors hearing the bang. To detect down to 1 cm 0.01 m (0.39 inches) the surface would have to be sampled 2*6320/0.01 = 1,264,000 times a second.

There are many off the shelf analoge-to-digital chips that will sample 8 bits a million times a second; for instance the Texas Instruments ADS7040, temperature range 40C to 125C, which costs less than a dollar.

To detect MMOD an accuracy of 5 cm may be sufficient. Aluminium is also a (near) worst case problem since it is a very good conductor of sound. At about 355 m/s air is an order of magnitude slower.

A practical system would have to handle several different types of material and filter out false alarms such as astronauts bouncing off the walls. I suspect that liquids going though pipes sounds different from the bang of something hitting the outside of a capsule.

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6942
  • Germany
  • Liked: 1906
  • Likes Given: 1969
Distinguishing a MMOD from normal station sounds is the easiest thing ever. It will have a very short sharp attack to identify. A very simple filter will do that. Even a hammer blow will be much slower.

They have devised materials to repair the Shuttle heat shield to some extent. Finding materials to mend a small hull breach will be many orders of magnitude easier. We are not talking about large ones as they will be exceedingly rare.

Others have already commented on triangulation. That may be somewhat tricky but once an impact is identified as MMOD it is worth the effort.

Offline woods170

  • IRAS fan
  • Senior Member
  • *****
  • Posts: 9052
  • IRAS fan
  • The Netherlands
  • Liked: 6517
  • Likes Given: 2218
Distinguishing a MMOD from normal station sounds is the easiest thing ever. It will have a very short sharp attack to identify. A very simple filter will do that. Even a hammer blow will be much slower.

They have devised materials to repair the Shuttle heat shield to some extent. Finding materials to mend a small hull breach will be many orders of magnitude easier. We are not talking about large ones as they will be exceedingly rare.

Others have already commented on triangulation. That may be somewhat tricky but once an impact is identified as MMOD it is worth the effort.

Emphasis mine.
One word: accessibility

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6942
  • Germany
  • Liked: 1906
  • Likes Given: 1969
Emphasis mine.
One word: accessibility

A hull breach would be mended from the inside.

Offline jgoldader

  • Full Member
  • ****
  • Posts: 728
  • Liked: 276
  • Likes Given: 155
-Triangulating sounds would probably be a nasty business of modeling, because the microphones are (presumably) attached to metal, and the sound waves would be propagating through the metal structures of the spacecraft.  You'd need to have some sort of acoustical model (edit: which could well be very different than the one used to analyze launch noise) and pinpointing a hit for detailed inspection might not be possible.  I could easily imagine saying, "Oh, we took a hit on the aft," but saying, "It's at PICA chunk 387" or whatever sounds like too much to ask.


Triangulation.

The speed of sound in aluminium is 6320 m/s.

Triangulation can work by differences in volume and the delay between sensors hearing the bang. To detect down to 1 cm 0.01 m (0.39 inches) the surface would have to be sampled 2*6320/0.01 = 1,264,000 times a second.

There are many off the shelf analoge-to-digital chips that will sample 8 bits a million times a second; for instance the Texas Instruments ADS7040, temperature range 40C to 125C, which costs less than a dollar.

To detect MMOD an accuracy of 5 cm may be sufficient. Aluminium is also a (near) worst case problem since it is a very good conductor of sound. At about 355 m/s air is an order of magnitude slower.

A practical system would have to handle several different types of material and filter out false alarms such as astronauts bouncing off the walls. I suspect that liquids going though pipes sounds different from the bang of something hitting the outside of a capsule.

I understand triangulation.  But you'd probably end up with diffraction effects, reflections, etc., and the speed of sound depends on the tension in the metal, and...  And you need a system that can handle that in near-real-time.  Hence the spirit of my comment.
« Last Edit: 08/31/2016 10:46 am by jgoldader »
Recovering astronomer

Offline Jim

  • Night Gator
  • Senior Member
  • *****
  • Posts: 32646
  • Cape Canaveral Spaceport
  • Liked: 11541
  • Likes Given: 344

Others have already commented on triangulation. That may be somewhat tricky but once an impact is identified as MMOD it is worth the effort.

That was done with accelerometers and a type that there would not be very many of on a Dragon (if any at all). 

Offline docmordrid

  • Senior Member
  • *****
  • Posts: 5457
  • Michigan
  • Liked: 2846
  • Likes Given: 1
 Debris Resistive/Acoustic Grid Orbital Navy-NASA Sensor (DRAGONS)

http://ntrs.nasa.gov/search.jsp?R=20140017141
DM

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 6942
  • Germany
  • Liked: 1906
  • Likes Given: 1969
And you need a system that can handle that in near-real-time.

No, you really don't need that though it would be good to have. You need a system that can distinguish between MMOD hits and other noises near real time. If you have a hit there is ample time to analyze. The vast majority of its time in space the spacecraft will be docked to the station. So that's the time thad adds most of the risk.

Offline MarkM

  • Member
  • Posts: 35
  • Liked: 2
  • Likes Given: 5
The issue has been looked into in the past: http://www.fox-tek.com/pdf/Aerospace.pdf

I am not sure if it has advanced but I believe that it is used in the petroleum industry currently.

Offline SWGlassPit

  • I break space hardware
  • Full Member
  • ****
  • Posts: 620
  • Liked: 523
  • Likes Given: 75
-Triangulating sounds would probably be a nasty business of modeling, because the microphones are (presumably) attached to metal, and the sound waves would be propagating through the metal structures of the spacecraft.  You'd need to have some sort of acoustical model (edit: which could well be very different than the one used to analyze launch noise) and pinpointing a hit for detailed inspection might not be possible.  I could easily imagine saying, "Oh, we took a hit on the aft," but saying, "It's at PICA chunk 387" or whatever sounds like too much to ask.


Triangulation.

The speed of sound in aluminium is 6320 m/s.

Triangulation can work by differences in volume and the delay between sensors hearing the bang. To detect down to 1 cm 0.01 m (0.39 inches) the surface would have to be sampled 2*6320/0.01 = 1,264,000 times a second.

There are many off the shelf analoge-to-digital chips that will sample 8 bits a million times a second; for instance the Texas Instruments ADS7040, temperature range 40C to 125C, which costs less than a dollar.

To detect MMOD an accuracy of 5 cm may be sufficient. Aluminium is also a (near) worst case problem since it is a very good conductor of sound. At about 355 m/s air is an order of magnitude slower.

A practical system would have to handle several different types of material and filter out false alarms such as astronauts bouncing off the walls. I suspect that liquids going though pipes sounds different from the bang of something hitting the outside of a capsule.

I understand triangulation.  But you'd probably end up with diffraction effects, reflections, etc., and the speed of sound depends on the tension in the metal, and...  And you need a system that can handle that in near-real-time.  Hence the spirit of my comment.

What he said.  Every structural connection, every interface, every free surface, and every rivet, bolt, and screw will reflect transmitted waves, absolutely confounding attempts to triangulate event locations.  Ever wonder why GPS acts weird when you're around a lot of tall buildings?  Same problem -- you have no way of knowing if the signal you received is from a straight line-of-sight or from a reflection.  Triangulation is only simple for simple hardware.  For real hardware that people actually use, it's not even close to being a simple problem.

The only way to get useful data with any reasonable time frame would be to have hundreds of sensors, which requires hundreds of wires, all of which weigh something.  You also need the computing infrastructure to make sense of that data and the power to run it all.  Automated debris strike detection and location is a very low TRL research project at this point and has no business being relied upon in a production spacecraft.

Offline SWGlassPit

  • I break space hardware
  • Full Member
  • ****
  • Posts: 620
  • Liked: 523
  • Likes Given: 75
Emphasis mine.
One word: accessibility

A hull breach would be mended from the inside.

The pressure shell itself is almost never directly accessible by the crew.  There is all manner of interior outfitting in the way.  Wire harnesses, fluid lines, ducts, avionics, stowage, bump shields, other hardware, etc. 

Online Robotbeat

  • Senior Member
  • *****
  • Posts: 28858
  • Minnesota
  • Liked: 9049
  • Likes Given: 5798
Major damage is rare. You don't even have to mend them. Just keep the door closed and take a different vehicle down. The issue you're trying to address here is otherwise-undetected significant damage leading to loss-of-crew.

We're worried about LOC. LOM is more acceptable (provided it's still rare).
« Last Edit: 08/31/2016 03:46 pm by Robotbeat »
Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

Offline SWGlassPit

  • I break space hardware
  • Full Member
  • ****
  • Posts: 620
  • Liked: 523
  • Likes Given: 75
Major damage is rare. You don't even have to mend them. Just keep the door closed and take a different vehicle down. The issue you're trying to address here is otherwise-undetected significant damage leading to loss-of-crew.

We're worried about LOC. LOM is more acceptable (provided it's still rare).

Keeping the door (hatch) closed is a mitigation for PNP (where damage to the visiting vehicle creates a hazard to ISS), not LOC/LOM.  In any case, the need for immediate safe harbor for crew members in the event of conjunctions, leaks, and toxic atmosphere events makes that mitigation step absolutely unacceptable for anything but contingency operations, meaning it can't be considered a hazard control on which to base the risk numbers.

A good rule of thumb when it comes to MMOD analysis and mitigation design: if it's simple and intuitive without heavy background experience, it is most likely wrong.  There are a number of people in this thread who have direct experience with this.  Listen to what they are saying.

Offline Steven Pietrobon

  • Member
  • Senior Member
  • *****
  • Posts: 19592
  • Adelaide, Australia
    • Steven Pietrobon's Space Archive
  • Liked: 7456
  • Likes Given: 997
Keeping the door (hatch) closed is a mitigation for PNP (where damage to the visiting vehicle creates a hazard to ISS), not LOC/LOM.

What is PNP? Is it Probability of No Penetration?

http://acronyms.thefreedictionary.com/PNP
« Last Edit: 09/01/2016 06:15 am by Steven Pietrobon »
Akin's Laws of Spacecraft Design #1:  Engineering is done with numbers.  Analysis without numbers is only an opinion.

Tags: