This is not a NASA Website. You might learn something. It's YOUR space agency. Get involved. Take it back. Make it work - for YOU.
Russia

Soyuz Launcher Recovery Plan Starts To Emerge

By Keith Cowing
NASA Watch
October 18, 2018
Filed under
Soyuz Launcher Recovery Plan Starts To Emerge

Roscosmos to make three unmanned launches before next manned flight, TASS
“Russia’s space corporation Roscosmos will carry out three unmanned launches by the end of the year before the next manned mission will be put in space, Roscosmos’s executive director for manned space programs, Sergei Krikalyov, told a news conference on Wednesday. “The Soyuz rocket will be launched only after the inquiry has identified the causes of the emergency and measures have been taken to prevent such situations in the future. Under the existing rules there must be at least one unmanned launch before the flight of a manned spacecraft. We have plans for at least three launches (before the next manned mission due in early December) from the Kourou space site, the launch of an unmanned spacecraft and of an unmanned spacecraft Progress. The confirmations will be more than enough to put the next crew in space,” Krikalyov said.”

NASA Watch founder, Explorers Club Fellow, ex-NASA, Away Teams, Journalist, Space & Astrobiology, Lapsed climber.

26 responses to “Soyuz Launcher Recovery Plan Starts To Emerge”

  1. Saturn1300 says:
    0
    0

    Members of Roscosmos commission and investigators have identified those
    who could be responsible in the damage to one of the sensing devices on
    board the Soyuz-FG carrier rocket, which could have led to the failed
    launch of the Soyuz MS-10 manned spacecraft, a source at the Baikonur
    space center told Sputnik on Thursday.

    • Jeff2Space says:
      0
      0

      Anytime something goes wrong, the Russians launch an investigation, immediately assign blame, fix the one thing that went wrong, and start flying again. IMHO, fear and intimidation seem to be an odd way to encourage “quality assurance” on such a large, complex, program.

      The sooner we start flying commercial crew the better. Unfortunately, NASA seems to have swung their “safety” pendulum to the opposite side of the scale to the point that the “paperwork” (in this seemingly waterfall process) is taking longer to finalize than the actual hardware. When the only other alternative is Soyuz, I’m not sure this is such a great way to go. At this point I’d like to see them switch to an iterative approach which would include flying what we have a.s.a.p. and then evolve the hardware and processes as we go.

      • Robert Jones says:
        0
        0

        Having a backup is good but my fear is that the commercial crew vehicles may well cost more than Soyuz, be less reliable than Soyuz, and/or be less safe as compared to Soyuz. Experiments with ISS “unmanned” should have been done years ago. Soyuz flown unmanned should have been done years ago.www.robert-w-jones.com

        • fcrary says:
          0
          0

          I think two of those concerns have already been resolved. We know how much commercial crew will cost NASA, since the contracts have already been signed. How much it costs SpaceX and Boeing is another matter. As I understand it, the contracts covered the final phase of development and certification, plus six flights (24 seats) from each company at the same $70 million per seat price the Russians were charging.

          As far as an unmanned Soyuz is concerned, the Progress resupply vehicle is a unmanned version of Soyuz. Their family trees split off a couple generations ago, so they are no longer as similar as they used to be, but they are still basically the same vehicle.

          • Daniel Woodard says:
            0
            0

            If they are getting $70 million per seat, SpaceX at least will be making a significant profit. Of course if they could use propulsive landing and avoid getting the capsule wet, it might permit re-use.

          • Steve Pemberton says:
            0
            0

            Ocean landings don’t seem to be much of a hindrance to Dragon reusability. Cargo Dragons are already being reused, and apparently crew Dragons will also be reused, except not on NASA crewed flights. NASA was apparently open to the idea but SpaceX has decided that NASA crewed flights will always be on a new Dragon, after which the capsules will be reflown on cargo flights.

          • ThomasLMatula says:
            0
            0

            Cargo Dragons don’t have the thruster pods with the SuperDracos in them. Water and engines never mix well which is why it was intended to be recovered on land and not at sea.

          • Steve Pemberton says:
            0
            0

            Are we sure that water landings is the reason that crew Dragons won’t be reused? I suspect the decision was more complex than just that. Getting the SuperDracos wet might add some extra work and time in refurbishment but might not have been insurmountable. I tend to think that the driving factor on a lot of these decisions is that there just aren’t going to be that many crewed Dragon flights, at least for NASA. Using previously flown Dragons on crewed flights would have required another round of paperwork to get NASA’s approval and SpaceX may have decided that for the limited number of flights it just wasn’t worth it.

          • ThomasLMatula says:
            0
            0

            Most likely it wasn’t, whichis why Elon Musk has written Dragon and Falon off and moved on to the NASA free BFR/BFS. It will be great to watch those vehicles fly. Imagine a rocket as large as the SLS first making a verticle landing.

          • fcrary says:
            0
            0

            If the crew contract is really $70 million per seat, which a brief literature survey makes me think it is, SpaceX will do extremely well. Cost plus contracts were intended to prevent profiteering. But this could be the best deal I’vs heard of since my great-great grandfather got rich from blockade running during the Civil War and profiteering on cotton exports during the Reconstruction. And the SpaceX deal would still be a cost-effective deal for NASA.

  2. ThomasLMatula says:
    0
    0

    Given the long history of the Soyuz this is most likely an issue of quality control, which means the basic solution is expanded and more detail inspections. These flights therefore serve mostly to build confidence that the more detailed inspection is working.

    That said, I think it would be wise to use at least one of them to replace the defective Soyuz docked to the station with new one that has been very carefully inspected, even if it delays the crew rotation.

  3. Jeff2Space says:
    0
    0

    Unfortunately, if you look at the failure rate of the Soyuz launch vehicle family over the last 10 to 20 years it looks like a systemic quality control problem not just “a random failure”. This many failures in such a long-lived “proven” system is quite troubling. And frankly, I’m surprised it took this long to impact a crewed flight.

    • Michael Spencer says:
      0
      0

      If you look at the success rate of the Soyuz launcher you discover an astonishingly robust rocket system.

      • Jeff2Space says:
        0
        0

        Yes it was once “an astonishingly robust rocket system”. Again, look at the data for the last 10 to 20 years. Look beyond the Soyuz FG (crewed variant) which has only had one failure. Soyuz 2 is the latest variant. Russia plans to eventually shift crewed launches to Soyuz 2.

        Soyuz 2 – 7 failures or partial failures in 79 flights or a 9% failure rate!
        https://en.wikipedia.org/wi

        Those are *not* respectable numbers for Soyuz 2.

        Here’s a graph of all R-7 launches:

        https://en.wikipedia.org/wi

        Note the successes and failures since 1997 when the recent grouping of failures seems to have started (note the most recent failure hasn’t made it into the graph yet). The failure (plus partial failure) rate over that period of time works out to 5%. That’s not at all respectable.

        • fcrary says:
          0
          0

          Unfortunately, it isn’t that easy. I pulled the numbers, and played with the statistics. Even summing over eight year intervals, I couldn’t get anything statistically significant. After the initial years, the R-7 had a failure rate of around 2.5% (1979-2010) and from 2011-2018, it’s been 6.6%. But the number of events doesn’t make that statistically significant. It’s more correct to say the failure rate was under 8% from 1979-2010, and under 20% from 2011-2018. That isn’t encouraging, but technically, it could just be bad luck.

          • Michael Spencer says:
            0
            0

            Another case of “Seems like…”

            And, another example for the Professor’s contention that statistics is poorly understood.

  4. Daniel Woodard says:
    0
    0

    The actual science of quality control as C. Edwards Demming formulated it does not mean just adding more people to watch procedures being followed, as it does for NASA. Nor does it mean waiting for a failure abnd blaming someone, as Russia seems to be doing. Quality control means making quantitative measurements, both nondestructively (i.e. dimensions of mechanical parts) and destructively, including testing a sample of production components to failure. Most critically it means continuing statistical analysis of the measured control parameters, whether they are physical measurements, results of destructive tests, or failure rates in life cycle testing. This critical part of the process is what makes it possible to determine whether a process is under control _before_ a failure occurs. If the process is under control, the measured parameters will be normally distributed with a small standard deviation. If a portion of the process begins to slip, the variability in a measured parameter will increase, usually because some measurements become outliers, falling outside the normal distribution. At that point (usually before any actual failures) it is essential to identify and correct the problems in the production process that has cause the loss of control.

    In Demmings’ day the critical parameters were plotted by hand, as was still the case when I studied industrial engineering in the Seventies. Today of course it could be done more easily, unfortuantely with all our electronics no one actually takes the time to understand the underlying principles.

    • ThomasLMatula says:
      0
      0

      Unforunately that is true, very few understand the basics of SPS let alone apply it, probably because few really grasp the basics of probablity and statistics.

      • Michael Spencer says:
        0
        0

        I have to think that engineering students are required to learn something about statistics, no?

        • Daniel Woodard says:
          0
          0

          I would guess they have basic training in statistics, but a lot of engineering training tends to be OJT and if a manufacturer doesn’t maintain a corportate practice it is forgotten. Demmings was hired by the War Department in WW2 because manufacturing quality was so low. After the war, he tried to sell his ideas to US industry but they were selling all they made and did not want to spend money on something that would not immediatel impact profits. He found a more receptive audience in Japan, which was trying to build an export industry after the war and had problems with poor quality.

        • fcrary says:
          0
          0

          I’d expect a good engineering school to require _something_ about statistics, but perhaps not an entire semester-long course. Berkeley didn’t require that for physics majors when I was a student there. (Although the physics department did play a game with not officially requiring an excessive number of courses, even though there was a de facto requirement if you wanted to pass some of the officially required courses.)

          One easy way to get a gap in a student’s knowledge is to allow specialization at the junior and senior level. For example, requiring six courses, selected by the student, out of a list of ten. I never learned much particle physics or any general relativity because I picked plasma physics instead. Engineering departments are no different. I don’t have trouble imagining engineering majors getting limited exposure to statistics. Maybe they took a computer course instead, covering things like finite element modeling and computational fluid dynamics.

          But some of the specifics Dr. Woodard mentioned are a little bit subtle. Consideration of outliers, for example. An introductory course in statistics likely to start with the assumption that a set of measurements (or objects or whatever) are normally distributed (i.e. a gaussian or bell curve distribution.) It would probably cover the central limit theorem, that any random process with enough samples/events/whatever will produce such a distribution. It probably would not cover what happens when that assumption is incorrect.

          If data show it isn’t correct (if there are too many outliers) that could be evidence of a non-random process and a red flag in quality control. Or it could mean reporting uncertainties and measurement errors using standard techniques will be an underestimate.

          E.g. something reported at three sigma significance supposedly means only a 0.3% chance it’s an incorrect measurement; and five-sigma means odds of 0.6 in a million. But that’s only true for a normal distribution of measurement errors. If there are a significant number of outliers, those odds do not apply. Unfortunately, that’s the sort of thing that gets left out of many intro statistics courses.

          • Michael Spencer says:
            0
            0

            I was not required to take a full semester’s course in statistics until graduate school; the requirement supported thesis research, obviously.

            In the case of many engineering disciplines I suppose I assumed that statistics is more directly related to the work, and thus needed earlier.

            (Plus, it’s an interesting subject).

        • mfwright says:
          0
          0

          Engineering students may take one course in statistics but several in calculus. CSPAN showed a lecture (I forgot speaker’s name) and the speaker said he has traveled the country and yet to find anyone that has used an integral at work. From what I see statistics are used all the time in one form or another. Whether it be trying to make sense with lots of data, or data used to manipulate political opinion.

          • fcrary says:
            0
            0

            I use integrals at work all the time. But I see your point. Once the solution is known, you don’t really need the calculus to use it. The real applications, I think, are dealing with novel problems and in teaching why the existing equations are true (and what underlying assumptions went into them.) If you’re satisfied with designing and building according to black-box tools, then a good grip on statistics is probably more important.

  5. JJMach says:
    0
    0

    Not to make too light of an ugly situation, but did anyone else read the headline and wonder if Roscosmos was thinking of modifying the Soyuz rocket with feet for landing? “Launcher recovery” threw me for a second.

    Yes, I get that this is too big of an ask for the old bird the moment I thought it. Curious if they do have any plans in that direction, since that’s what all the cool kids are up to these days.

  6. mfwright says:
    0
    0

    Reading the book “Starman” about Gagarin and it said Korolev based his Soyuz on the proposed GE Apollo configuration, http://www.astronautix.com/… (one of many pages on the internet). Prior to the fatal flight of Soyuz 1, engineers identified more than 200 faults but Brezhnev wanted a notable spaceflight during the 50th anniversary of the Russian Revolution.

    Obviously there has been many improvements to Soyuz over the 50 years of service, maybe this configuration is quite robust and can tolerate sloppy quality control.