Aethon

Orion. It's not just a constellation anymore.

473 posts in this topic

Reading the recent stories about how they are considering a manned EM-1, it made me realize that EM-2 is hardly better than EM-1 as a first crewed launch. EM-2 has a different block SLS, a different upper stage, and the EM-1 Orion is not actually an all-up version, it has no life support, no waste management, even the computers, etc have no crew controls. So it is not really a usefuul test of anything. You'd think if safety was the primary driver, they'd fly an otherwise flight article crew vehicle as the EM-1 unmanned test. Make sure the life support can run for 3 weeks, etc. 

Seems like the only thing actually held constant might be the booster and the LES between the 2 flights.

1 person likes this

Share this post


Link to post
Share on other sites

So EM-2 is supposed to reply the EM-1 Orion vehicle (cool), so it's a good test in that respect, but you'd think they'd want to sort out the crew systems in flight in some fashion instead of the current plan. Man-rating seems as much art as science, lol. You'd think it would require a clone of crew flight article being fully tested.

Share this post


Link to post
Share on other sites

Instead of a manned EM-1, they should do an EFT-2 with a Delta IVH and a systems-complete Orion with full ECLSS. No need to man-rate the D IVH:

  • Fly unmanned in LEO for a week, including propulsion tests and maneuvers.
  • Dock to the ISS, be boarded by crew and tested for a week in LEO.
  • Undock and free-fly for a week, including manned propulsion tests and maneuvers.
  • Dock to the ISS, disembark crew.
  • Undock and reenter.

You might even be able to squeeze an EVA from Orion in the schedule somewhere.

That plan would need some logistics, including having the two fully-functional IDAs installed at the ISS, and rotating the astronauts trained on Orion into an ISS expedition. But it would be cheaper and safer than a manned EM-1.

Of course, the long pole here is a "systems-complete Orion" with a functional ECLSS. I doubt that will be ready before EM-2 anyway, so speculation about a manned EM-1 is a bit moot.

Edited by Nibb31
10 people like this

Share this post


Link to post
Share on other sites
1 hour ago, Nibb31 said:

Instead of a manned EM-1, they should do an EFT-2 with a Delta IVH and a systems-complete Orion with full ECLSS. No need to man-rate the D IVH. Orion could docked to the ISS, be boarded by crew and tested for a week in LEO, doing various manoeuvers, disembarked crew at the ISS again, and reentered. The ISS would be a safe-haven if something went wrong with the vehicle.

That plan would need some logistics, including having the two fully-functional IDAs installed at the ISS, and rotating the astronauts trained on Orion into an ISS expedition.

In many ways I think that would be much cooler than the Apollo 8 lite mission being proposed for EM-1. To me, using the ISS as a genuine space base to conduct orbital missions and test flights from sounds a lot more futuristic than a free-return around the Moon. The latter would be more visually impressive for the general public though I suspect.

2 people like this

Share this post


Link to post
Share on other sites
7 hours ago, Nibb31 said:

Instead of a manned EM-1, they should do an EFT-2 with a Delta IVH and a systems-complete Orion with full ECLSS. No need to man-rate the D IVH:

  • Fly unmanned in LEO for a week, including propulsion tests and maneuvers.
  • Dock to the ISS, be boarded by crew and tested for a week in LEO.
  • Undock and free-fly for a week, including manned propulsion tests and maneuvers.
  • Dock to the ISS, disembark crew.
  • Undock and reenter.

You might even be able to squeeze an EVA from Orion in the schedule somewhere.

That plan would need some logistics, including having the two fully-functional IDAs installed at the ISS, and rotating the astronauts trained on Orion into an ISS expedition. But it would be cheaper and safer than a manned EM-1.

Of course, the long pole here is a "systems-complete Orion" with a functional ECLSS. I doubt that will be ready before EM-2 anyway, so speculation about a manned EM-1 is a bit moot.

I agree. As a PR thing, I think that the Moon is much more interesting, but from a testing standpoint, this is exactly what I was thinking. It seems like their current plan results in flying EM-1, but wasting the mission by not actually testing the actual spacecraft.

Share this post


Link to post
Share on other sites

A question out there for anyone who keeps up with the specifics more than I do:

Is man-rating SLS/Orion any different than man-rating Dragon 2, or CST-100? When, for example, is the max-Q LES test for Orion? They did a pad abort test at White Sands, but I cannot find a video of a max-q abort, nor any indication of a launch before EM-1---or did I miss it somehow?

They have to do this, right, or do only Boeing and SpaceX (and any other commercial crew vessels) have to do this?

1 person likes this

Share this post


Link to post
Share on other sites

Apparently the max q abort test will be before EM-2 using a peacekeeper missile. 

How do they get their 1:500 loss rate numbers (requirement for commercial crew) from SLS after just one flight to man rate the LV (that's 4 times better than Shuttle managed, up and down).

Share this post


Link to post
Share on other sites
1 hour ago, tater said:

Apparently the max q abort test will be before EM-2 using a peacekeeper missile. 

Launching a 15 t craft with a 4 t payload rocket?

1 person likes this

Share this post


Link to post
Share on other sites
1 hour ago, tater said:

How do they get their 1:500 loss rate numbers (requirement for commercial crew) from SLS after just one flight to man rate the LV (that's 4 times better than Shuttle managed, up and down).

LOC/LOM rates are calculated from the failure rates of individual subsystems, which are calculated from the failure rates of components, etc...

Doing all these calculations is the major component of "man rating" a vehicle.

Edited by Nibb31

Share this post


Link to post
Share on other sites
4 hours ago, kerbiloid said:

Launching a 15 t craft with a 4 t payload rocket?

For the LES test, you only need to go suborbital until MaxQ, i.e. you only need to reach some 15km of altitude. And considering that a Peacekeeper has a TWR of 2.5 (twice that of Saturn V), i.e. much bigger than what SLS probably will have, you can add a considerable payload before it won't be able to reach MaxQ at the desired altitude and speed.

1 person likes this

Share this post


Link to post
Share on other sites
4 hours ago, Nibb31 said:

LOC/LOM rates are calculated from the failure rates of individual subsystems, which are calculated from the failure rates of components, etc...

Doing all these calculations is the major component of "man rating" a vehicle.

The mind boggles at how you would calculate such a thing.

I also look with disdain upon the "mean time between failure" numbers bandied around in the consumer electronics industry.

What does "Man rating" mean in practice?

Using components that have proven themselves on unmanned flights?

Additional redundancy, at the minor systems level -  backup electric circuits and valve actuators/sensors/microprocessors - that would not be used on a cargo launcher?

Extra stages of inspection of every assembly action - every time an engineer plugs a connector cable in from one module to another, two others come along to inspect his work?

Thorough non-destructive testing of every component coming from suppliers, be it a fastener , cable or strut, before attempting to fit the part?

More generous safety margins on component spec, erring on the side of cost and weight?

How much safety comes from the above measures, as opposed to 

a)  Mission Architecture

Ensuring there is a viable abort mode for every part of the flight.    One of the issues with the space shuttle being the lack of launch escape system .    

b) Program Management

Avoiding complacency in the face of known risks - Challenger launching outside of the operating temperature parameters of the SRB seals,  Columbia continuing its mission despite foam strike being witnessed on launch. Apollo 1 conducting a ground test with a pure oxygen cabin atmosphere, at much higher pressure than would have been used in space.

In the case of SLS, the issue of whether to launch manned or unmanned first, and what degree of change to the launch vehicle stack config justifies another unmanned test flight.

1 person likes this

Share this post


Link to post
Share on other sites
5 hours ago, Nibb31 said:

LOC/LOM rates are calculated from the failure rates of individual subsystems, which are calculated from the failure rates of components, etc...

Doing all these calculations is the major component of "man rating" a vehicle.

I understand they are calculated, but it seems like comparison with reality would also be in order. The actual shuttle failure rate was substantially higher than that, yet they continued flying. It's a reasonable design goal (1:500 launch failures, 1:500 reentry failures), but it seems like given even the most optimistic launch cadence for SLS/Orion, they will not have enough real data for a long, long time.

Share this post


Link to post
Share on other sites
31 minutes ago, Tullius said:

considering that a Peacekeeper has a TWR of 2.5

Then TWR 2.5 * 4 / 15 ~= 0.7

So, maybe if just the first stage of  MX, replacing other stages with Orion.

Edited by kerbiloid

Share this post


Link to post
Share on other sites
7 minutes ago, kerbiloid said:

Then TWR 2.5 * 4 / 15 ~= 0.7

So, maybe if just the first stage of  MX, replacing other stages with Orion.

Peacekeeper weighs approx. 90 tonnes, i.e the calculation is something along the lines of (90+4)/(90+15)*2.5=2.24. So there is still plenty of TWR to spare at takeoff.

Edited by Tullius

Share this post


Link to post
Share on other sites
3 hours ago, AeroGav said:

The mind boggles at how you would calculate such a thing.

I also look with disdain upon the "mean time between failure" numbers bandied around in the consumer electronics industry.

What does "Man rating" mean in practice?

Using components that have proven themselves on unmanned flights?

Additional redundancy, at the minor systems level -  backup electric circuits and valve actuators/sensors/microprocessors - that would not be used on a cargo launcher?

Extra stages of inspection of every assembly action - every time an engineer plugs a connector cable in from one module to another, two others come along to inspect his work?

Thorough non-destructive testing of every component coming from suppliers, be it a fastener , cable or strut, before attempting to fit the part?

More generous safety margins on component spec, erring on the side of cost and weight?

Pretty much all of the things that you mentioned.

Components have MTBF ratings and known failure modes. Not all of those failure modes cause a LOM/LOC. There is also built-in redundancy and redundant systems.

This is all why "man rating", like most certification processes, is such an expensive thing, even though it's mainly just a lot of paperwork. Think ISO9000 on steroids. This is also why the aerospace industry uses aerospace-certified components and frowns upon SpaceX's conduct of buying"off-the-shelf" struts and fasteners.

 

3 hours ago, tater said:

I understand they are calculated, but it seems like comparison with reality would also be in order. The actual shuttle failure rate was substantially higher than that, yet they continued flying. It's a reasonable design goal (1:500 launch failures, 1:500 reentry failures), but it seems like given even the most optimistic launch cadence for SLS/Orion, they will not have enough real data for a long, long time.

Absolutely. It's theoretical of course (I doubt there will ever be 500 flights of SLS), but you've gotta work with something, right?

2 people like this

Share this post


Link to post
Share on other sites

No, I agree completely, I was just musing on the difference between man rating the current commercial crew vehicles and launch vehicles, vs SLS/Orion.

Under the (perhaps misguided :wink: ) assumption that SpaceX stops tweaking F9 and launches a definitive vehicle, the LV for Dragon 2 will have multiple tests under its belt, with loads of data. That data will include failures that were addressed before return to flight, then at some point they have X launches of the definitive crew version, then they will do an unmanned test, then at some point manned tests (the same for CST-100).

Whatever number X is, it's certainly going to be larger than SLS's number of tests for EM-2, and both will be vastly lower than the theoretical value of 500 (I realize this is a probability, not a real number) . Presumably the 500 number is a response to shuttle, and shuttle was done without these metrics. I'm also curious at what point they might (if ever) un-rate a vehicle. It would be interesting to apply the same criteria use for 1:500 to Shuttle as designed, without considering the actual failure modes, and see what the risk metrics show. If back-analyzing Shuttle doesn't provide theoretical risk levels concordant with the actual risk, then those theoretical tools might not be very useful.

Edited by tater

Share this post


Link to post
Share on other sites
24 minutes ago, Nibb31 said:

This is all why "man rating", like most certification processes, is such an expensive thing, even though it's mainly just a lot of paperwork. Think ISO9000 on steroids. This is also why the aerospace industry uses aerospace-certified components and frowns upon SpaceX's conduct of buying"off-the-shelf" struts and fasteners

Agreed, but in the consumer electronics industry ISO 9001 doesn't guarantee all that much.    Any large company (as opposed to a small, family run repair shop) will have to have a system to log what it's engineers do anyway.   Someone creates some flowery sounding documents describing how things are done , and there you go , you're ISO 9001 compliant.  Employees may have never read these process control documents, let alone adhere to them.   Just so long as the inspector can be shown that they exist.     Even then, design tolerances, employees caring about the quality of their work matters more than rigidly following some document written by someone who isn't even an engineer.

As for MTBF,    build 1000 circuit boards containing dodgy capacitors (or simply under-specced ones, or that are being subject to high temperature).  Test these 1000 circuit boards for 1000 hours each and get zero fails,  you can advertise the component of having a MTBF of "better than a million hours".   In practice, these boards start dying after 1200 hours operation and by 1600 hours the mortality rate has risen to 80%.   

Finally,  I recently had to memorise an incredibly dull presentation on soldering quality control.   What a Level 3 (mission critical) certified solder joint looks like, in terms of miniscus, overlap, fillet etc. etc. etc.. zzzzzz

I suppose I'd want my space suit life support to have great looking solder joints on its pcbs too, but it's by no means the whole story.   The components could still be overstressed , under specced, poor quality plagued by noise emi etc. or poor design  but  at least it's shiny !

But yes, Aerospace is a different ball game,and it's not just about paying lip service.

Share this post


Link to post
Share on other sites
28 minutes ago, tater said:

No, I agree completely, I was just musing on the difference between man rating the current commercial crew vehicles and launch vehicles, vs SLS/Orion.

Under the (perhaps misguided :wink: ) assumption that SpaceX stops tweaking F9 and launches a definitive vehicle, the LV for Dragon 2 will have multiple tests under its belt, with loads of data. That data will include failures that were addressed before return to flight, then at some point they have X launches of the definitive crew version, then they will do an unmanned test, then at some point manned tests (the same for CST-100).

Whatever number X is, it's certainly going to be larger than SLS's number of tests for EM-2, and both will be vastly lower than the theoretical value of 500 (I realize this is a probability, not a real number) . Presumably the 500 number is a response to shuttle, and shuttle was done without these metrics. I'm also curious at what point they might (if ever) un-rate a vehicle. It would be interesting to apply the same criteria use for 1:500 to Shuttle as designed, without considering the actual failure modes, and see what the risk metrics show. If back-analyzing Shuttle doesn't provide theoretical risk levels concordant with the actual risk, then those theoretical tools might not be very useful.

I heard somewhere that Apollo was designed to a failure rate of 1 in 1000. So I would be surprised if the Shuttle wasn't designed to some similar figure.

The problem is not really the calculation of risk, you have to do it (there is no way of doing it otherwise). And generally, you overestimate the risks (as in: You guess that an engine fails 1 in 1000 launches, but you enter a failure rate of 1 in 500 into the calculation).

The problem is that you are going to miss some possible failure modes or misidentify the actual failure rate of a component. Nobody really bothered about the risk created by the O-ring design until after Challenger. The probability that the Shuttle's heat shield could be fatally damaged by foam from the external tank was also underestimated until Columbia.

In hindsight, NASA estimated the risk of failure for the Shuttle during the first flights to be 1 in 9, while the later ones were 1 in 90 (i.e. barely acceptable). But that is in hindsight and doesn't help with the development of a spacecraft, other than showing that targeting such an extreme reliability as 1 in 500 or more is a good idea.

Doing lots of test flights or man-rating a cargo rocket, makes it easier to estimate the risks, since you can analyse the results of those flights. But this in turn also means that you can use sharper estimates on the risk of different failures, and thereby reduce the amount of over-engineering to be really sure that you achieve the desired reliability, and thereby making the development cheaper.

2 people like this

Share this post


Link to post
Share on other sites

http://www.spacepolicyonline.com/news/gao-slams-nasas-cost-estimating-for-orion-sls

(this link was posted at NSF, and I hadn't read it before, even though it is old)

Regardless of what the cost per launch might be in the indeterminate future, the total dev cost of this large hog is now hovering around 19-22 billion $. They won't get near even shuttle total cost per launch (program $/launches) until they fly this thing dozens of times. 

Edited by tater

Share this post


Link to post
Share on other sites
1 hour ago, tater said:

http://www.spacepolicyonline.com/news/gao-slams-nasas-cost-estimating-for-orion-sls

(this link was posted at NSF, and I hadn't read it before, even though it is old)

Regardless of what the cost per launch might be in the indeterminate future, the total dev cost of this large hog is now hovering around 19-22 billion $. They won't get near even shuttle total cost per launch (program $/launches) until they fly this thing dozens of times. 

Even if they would get the price down to shuttle levels for a number of launches equal to the one of the shuttle, they will never get the number of missions. Okay, NASA may send one probe (due to the payload capacity of SLS, of the billion dollar class) per year for the next 30 years with SLS, do a dozen missions to the Moon and 2-3 missions to Mars, before the Senate decides that they should shut down the program, since they achieved all goals (cf. Apollo). But that is still at most some 75 missions (and that is still a very generous amount), i.e. half of the shuttles launches.

But you also need to put it into perspective: the shuttle needed much more flights to achieve its goals: 22 Spacelab missions, 10 to Mir, 5 for Hubble, 37 to ISS, and numerous other science missions or satellite transports. Every goal that the shuttle achieved or tried to achieve relied on numerous launches, as otherwise they wouldn't be worthwile.

Would it be a problem, if one Mars mission costs as much as the entire American contribution of ISS (37 shuttle missions, all of the American modules etc.), i.e. some 100 billion dollars?

1 person likes this

Share this post


Link to post
Share on other sites

Posted (edited)

33204101741_4feb9a2c38_o.jpg?itok=f9Pbsg

 

NASA performs Orion abort parachute test.  Sorry, can't find it on YouTube yet.

http://www.space.com/35979-nasa-orion-capsule-parachutes-down-after-abort-sequence-simulation-video.html

 

And the first SLS second stage arrives at the Cape.

 

IMG_1452_7_SLS-ICPS_Ken-Kremer-700x432.j

 

http://www.universetoday.com/132259/1st-sls-2nd-stage-arrives-cape-nasas-orion-megarocket-moon-launch-2018/

Edited by Aethon
4 people like this

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now