Stories
Slash Boxes
Comments

SoylentNews is people

posted by janrinok on Wednesday September 04 2019, @07:31AM   Printer-friendly

About a week ago, the 18th Space Control Squadron, US Air Force, relayed warning data to the European Space Agency.

The data indicated that there was a non-negligible collision risk between ESA's Aeolus satellite and Starlink44, an active SpaceX satellite, at 11:02 UTC on Monday, 2 September.

As days passed, the probability of collision continued to increase, and by Wednesday, August 28, ESA's Ops team decided to reach out to Starlink to discuss their options. Within a day, the Starlink team informed ESA that they had no plan to take action at that point. By Thursday evening, ESA's probability threshold for conducting an avoidance manoeuvre had been reached, and preparations were made to lift Aeolus 350 meter in orbit. By Sunday evening, chances of a collision had risen to 1 in 1000, and commands were sent to the Aeolus satellite, which triggered a total of 3 thruster burns on Monday morning, half an orbit before the potential collision. About half an hour after the collision prediction time, Aeolus contacted base, and normal measurement operations could continue.

What the SpaceX satellite was doing in ESA's Aeolus orbit is not clear.

ESA has taken the opportunity to point out that, given SpaceX plans to put up 20,000 of those things, handling monitoring and avoidance semi-manually, and by mail, is no longer practical.


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 4, Insightful) by The Shire on Wednesday September 04 2019, @02:09PM (5 children)

    by The Shire (5824) on Wednesday September 04 2019, @02:09PM (#889529)

    If you've ever worked in any sort of support role, no matter how critical, you'll know that this is exactly the sort of issue that can and does crop up all the time. Their incident response system did work initially. I'm sure a response ticket was generated, the entities involved discussed it, and both agreed no action was needed so the ticket was closed. The later request from ESA under that same ticket probably fell through because the system had already closed or deprioritized it with a "no action needed" status. This is where the bug likely lies and it's something that can happen in any complex system. I'm sure it didn't help that this occurred during a US holiday as well.

    I do fault ESA for not using additional channels to contact the Starlink team however. Hell, they could have DM'd Musk himself.

    It's worth pointing out that (b) SpaceX does know where their satellites are, (c) they have a working collision probability system same as ESA and the USAF (even this high risk assessment put the two satellites as passing 2km away from each other), (d) they do have the capability to move starlink satellites when necessary, (e) starlink satellites and really any satellite in that low orbit will come down fairly quickly without constant control - and there is tremendous debris up there which active satellites are always maneuvering around anyway, and (f) there is a coordination system with other agencies as was demonstrated by the initial contact with ESA.

    Starlink is new, this is the first set of satellites put in place. If you think all such endeavors don't have growing pains like this then you've never worked in the tech sector. This comm bug will be fixed, the whole process reanalyzed and made better and they will move forward.

    Now, if you're really looking for something to get angry about take a look at when China tested a satellite destroying satellite. They didn't announce the test, and the debris from the test posed an enormous hazard for everything in that orbit. And of course just last year they lost control of their entire mini space station, the Tiangong-1, and had to just sit there and hope it didn't hit anything before finally deorbiting. That's how you demonstrate incompetence.

    Starting Score:    1  point
    Moderation   +2  
       Insightful=2, Overrated=1, Underrated=1, Total=4
    Extra 'Insightful' Modifier   0  
    Karma-Bonus Modifier   +1  

    Total Score:   4  
  • (Score: 2) by quietus on Wednesday September 04 2019, @04:01PM (4 children)

    by quietus (6328) on Wednesday September 04 2019, @04:01PM (#889574) Journal

    I've got no beef with either SpaceX or ESA.

    I do have quite a bit of experience with working in a support role, at L3. Something like this would immediately be signaled as a prio-1, and followed very closely, until estimated collision time, and then a bit after. If that didn't happen, that's a clear lack of professionalism. That's simply not a matter of a bug in a ticketing system -- which, given SpaceX' history, should have been cleared out long time by now.

    Not being on the ball in space can easily cost your company at least tens of millions of dollars (or half a billion in Aeolus case) in direct damages, not to mention the damage in reputation. In such circumstances, you do not take a holiday as a company.

    If Starlink is new, you'd expect even more due diligence and extensive testing.

    If SpaceX knows where their satellites are, and have a working collision probability system, they should have contacted ESA themselves once their system indicated the threshold level of 1 in 10,000 was breached, let alone the threshold of 1 in 1000.

    Orbits, or orbit bands, are not haphazard things you can choose to your liking: once Starlink44 came within range of Aeolus' orbit, they should have fired its thrusters and move it back to its designated place. That Starlink44 is moving at all into any other active satellite's orbit is a very, very, very, very bad sign in itself. That SpaceX didn't take such action is quite simply not believable.

    There is clearly no coordination system: if there was, both companies would have followed up each other (in)actions. SpaceX clearly didn't do that, or they would have mentioned those further communications. In that respect, also note the remark in ESA's description where they mention that SpaceX reacted within a day. That's not a compliment: that's hard criticism.

    • (Score: 2) by The Shire on Wednesday September 04 2019, @06:25PM (3 children)

      by The Shire (5824) on Wednesday September 04 2019, @06:25PM (#889622)

      Orbits, or orbit bands, are not haphazard things you can choose to your liking

      I think you overestimate the accuracy with which ANY company knows their satellite position to be especially in high drag LE orbits like Starlink occupies. In all cases they have an approximate spherical area where they believe it to be. Even so, when the probability of the two satellites positions had an overlap possibility of 1 in 50,000 they discussed the situation and both companies decided no action was necessary. So it's not like they aren't watching and communicating with each other. And it wasn't until much later that the AF systems came up with an update that placed the satellites within 2km of each other that ESA opted to make a maneuver.

      And the idea that no one knows why Starlink "was in the ESA's Aeolus orbit" is misleading. It never was. The assigned orbits simply brought the two in close (2km is considered close for a satellite) proximity. Had the ESA done nothing it would still have been a 1 in 10,000 shot that everyones position data was off enough that they would collide. ESA is the only one who says it was a 1 in 1000 chance. The AF only put it at slightly above 1 in 10,000.

      Starlink can do better, and you can bet they will after this. But honestly, there was never any real danger.

      • (Score: 2) by quietus on Wednesday September 04 2019, @07:22PM (2 children)

        by quietus (6328) on Wednesday September 04 2019, @07:22PM (#889654) Journal

        What you are saying is that NASA can track 20,000 pieces of space debris the size of a softball and 500,000 pieces the size of a marble, can manoeuvre their satellites away from these, but has not really an idea where their satellites are at any moment to about, what, hundreds of meters?

        Further, linky to the Air Force probability data, please.

        • (Score: 2) by The Shire on Wednesday September 04 2019, @09:37PM (1 child)

          by The Shire (5824) on Wednesday September 04 2019, @09:37PM (#889704)

          What I'm saying is that when NASA says they're tracking 20,000 pieces of debris they really mean they have a good ballpark idea of where it is, like within a few kilometers. This is why collisions are reported in terms of probability rather than being definitive. They only know the position of a satellite is somewhere within a given area - not exactly, nor do they have a high precision idea of its speed. Why would ESA move their satellite a 1/4 mile if they knew precisely where and how fast everything was.

          • (Score: 2) by quietus on Thursday September 05 2019, @10:46AM

            by quietus (6328) on Thursday September 05 2019, @10:46AM (#889954) Journal

            Within a few kilometers, eh?

            GPS and Galileo satellites transmit satellite orbit data as a set of 16 quasi-Keplerian parameters, known as the ephemeris. Two further parameters are used, both of which are considered constant: the Earth rotation rate, and the Earth's gravitational constant. Long story short, Keplerian motion/orbit is determined by solving Kepler's equation, a calculation which is solved iteratively, resulting in millimetric accuracy.

            Satellites depart from pure Keplerian motion due to a combination of nonuniformity of the Earth's gravitational field, the gravitational fields of the Sun and Moon, solar radiation pressurre, and other effects. These are approximated by the remaining ephemeris parameters: the mean motion correction (resolution 10E-13 rad/s), rates of change of the inclination and longitude of the ascending node (both also with a resolution of 10E-13 rad/s), and six harmonic correction terms, 4 of which have a resolution of 10E-9 rad/s, and 2 of which have a resolution of 0.03125m.

            Source: Principles of GNSS, inertial, and multisensor integrated navigation systems, second edition, Paul D. Groves, published by artech house (2013). Pages 330-339.

            I notice you didn't provide the 'AF' data.