Showing posts with label failure. Show all posts
Showing posts with label failure. Show all posts

Saturday, 21 December 2019

The stars did not align well for Starliner, it seems

click map to enlarge

Yesterday's Boeing CST-100 Starliner Orbital Flight Test was a true nailbiter. This blogpost briefly reitterates what happened, and what could have happened had they not been able to eventually raise the orbit.

Launched atop an Atlas V rocket, this uncrewed inaugural test flight of the new Boeing Starliner crew transport vehicle should have gone on its way to a docking at the ISS today, followed by undocking and landing at the White Sands Missile Range a week from now. The map above which I prepared pre-launch from information in the Starliner Press Kit and Starliner Notebook, shows what should have been the launch track and some keypoints on that track. As we now know, it went wrong at one of these keypoints.

Launch was at 11:36:43 UT. The Atlas V and Centaur upper stage performed fine, the Centaur inserting the Starliner in a 76 x 191 km suborbital trajectory some 12 minutes after launch. Three minutes later, the Starliner separated from the Centaur.

Next, 31 minutes after launch near 12:08 UT, it should have fired its own thrusters, in order to raise perigee and in this way circularize the orbit, becoming truely orbital.

And that went wrong.

Due to a misfunctioning Mission Elapsed Time clock, the Starliner's orbit insertion burn did not go as planned. Initially, an "attitude problem" was reported as well. The next half hour or so was nailbiting, as Boeing and NASA were not quite coming forward with information, apart from the ambiguous comment that the Starliner had "stabilized" its orbit (which is extremely ambiguous wording).

Those of us who know about orbits, realised that if no orbit insertion burn took place, the Starliner would continue on a suborbital trajectory, and reenter with or shortly after the Centaur upper stage (see also the end of this post, where I modelled this). The Centaur reentry was expected to occur south of Australia at about 12:30-12:35 UT (see the map in top of this post), and as the clock approached that time, it became really nailbiting: was the Starliner crew module still on orbit, or breaking up and burning up over the Indian Ocean?

Eventually, it became clear that, many minutes after the original burn time, Boeing did manage to do a burn that raised perigee from 76 to 180 km.

As an interesting sidenote: during the post-launch press conference, Boeing's Jim Chilton seemed to suggest (at 7:25 in below video) that following the timer anomaly, they tried to uplink new commands, but were faced with delays caused by the relay satellite(s) used (TDRS). It also transpired that on a crewed flight, the crew itself would have intervened in this stage:





Orbital data released by CSpOC provided the first unambiguous information to the world about the whereabouts of Starliner. A pre-burn orbit appeared first, showing a 76 x 191 suborbital orbit. As this was pre-burn, this still did not say much about Starliner's state. But shortly after that, a 186 x 221 km orbit was published, somewhat later followed by a new 180 x 221 km orbit. These showed that Starliner had reached a safe orbit around the earth.

The diagram below shows the altitudes of apogee and perigee of the orbit published so far (21 december 12 UT): currently it is in a 241 x 265 km orbit.

click diagram to enlarge

The amount of fuel spent in the emergency manoeuvres after the planned burn did not occur, was thus that it was no longer feasible to reach and dock to the ISS. Over the night, a new burn or series of burns therefore raised the orbit to 241 x 265 km, 58.4 degree inclined, lining it up for a landing at White Sands Missile Range on Sunday 22 December.

The current orbit (epoch 19355.3601887) results in a landing opportunity at White Sands between 12:45-12:55 UT on Sunday 22 December, approaching the range from over the eastern Pacific, as can be seen in the map I prepared below:


click map to enlarge
This is based on the current (epoch 19355.3601887) orbit. If new orbit adjustments happen, the projected time of landing might change slightly (e.g. a lowering of the orbit would make the Starliner speed ahead a bit, resulting in a slightly earlier landing time).

[UPDATE 21 Dec 22:15 UT: NASA has announced that the landing will be around 12:57 UT]


What if the orbit raise had failed completely?


Starting from the first, pre-boost orbit released, 76 x 191 km, I used GMAT to model what would have happened. I find that the Starliner, had it continued in that orbit, would have reentered over Polynesia around 12:50 UT,  about 1h 15m after launch, with its first revolution still uncompleted:

click map to enlarge

Tuesday, 9 January 2018

Fuel dump of Zuma's Falcon 9 Upper Stage observed by a Dutch pilot over east Africa (and rumours that Zuma failed)

image (c) Peter Horstink, used with permission
click image to enlarge

The spectacular image above was taken by Peter Horstink, the Dutch pilot of a Boeing 747-400 freighter aircraft (Martinair Holland flight MPH8371 from Amsterdam to Johannesburg), around 3:15-3:20 UT on January 8. The aircraft was flying at 35000 feet just north of Khartoum, Sudan, at that moment, which can be seen in the foreground (the yellow lights). The image above is the first one out of four images taken by Horstink.

The spectacular green-blue "spiral" on the image is, given time and location and character, almost certainly the Falcon 9 Upper Stage from the launch of the classified Zuma satellite that day (see my earlier post here), depressurizing and venting fuel at the end of its de-orbit burn. Some 15-25 minutes later, it would re-enter in its designated re-entry zone in the southern Indian Ocean (see map below).

Horstink described his observation as follows (translated from his Dutch e-mail report):

"It started with a greenish light in the top of my front window. At first I thought it was a reflection from some lightsource behind me, but it turned out not to be. At about 218 UTC [this later turned out to be a mistake and must read 3:18 UTC: the aircraft passed Khartoum near 3:25 UTC - ML] with a  very clear sky and with Khartoum in our sight, a point of light (like a star but somewhat bigger) moved from above us to South of us. It moved slower than a usual satellite but clearly did move. I estimate we saw it for 2-3 minutes. The waning moon crescent at that time was almost right above us. The object was surrounded by a greenish glow in the shape of spiral arms, like a spiral galaxy. Two of them, which didn't seem to move much. The total  size of the phenomena was about three times the diameter of the moon."
(note that when measured from the photographs using the star background as a reference, the actual diameter of the spiral cloud is about 11 degrees. The cloud is at ~8 degrees elevation over the horizon, near azimuth 155 degrees. The two bright stars to the right of the cloud are alpha and beta Centauri ).

The map below gives my estimated trajectory for the Falcon 9 Upper Stage, with apogee at ~900 km. It fits the area of the sighting, the launch hazard zone direction and the de-orbit zone position (from Maritime Broadcast Warnings) in the Indian Ocean. The Falcon 9 Upper Stage should have re-entered into the atmosphere between 3:30-3:45 UT, about 30 minutes after the  window from the Maritime Broadcast Warning opened at 3:00 UT.

click map to enlarge

The sighting points to a somewhat higher orbital altitude for Zuma than I had anticipated before the launch: with hindsight, I had too much of an idée-fixe that the orbital altitude would be similar to that of USA 276. The Falcon 9 sighting over East Africa suggests an altitude over double as high, in the order of 900-1000 km rather than my original 400 km estimate.

The sighting does confirm the 50-degree orbital inclination of the orbit. A new estimated elset based on this revision of the orbital altitude is here.

The map below shows the (very) approximate position of the aircraft at the time of Peter Horstink's observation relative to the Falcon 9 trajectory (times in UT, January 8th 2018). The aircraft was flying on a heading of 170 degrees, and Horstink gives his position as "between waypoint Alpox and Khartoum VOR" which corresponds to about 16.38 N, 32.35 E. The Falcon 9 Upper Stage was coming down at an altitude in the range of 200-400 km at that time. Approximate positions for the Falcon 9 Upper Stage are indicated in 2-minute intervals:

click map to enlarge

Horstink made the image above and below plus a few more with a handheld camera, from the cockpit of the aircraft.

image (c) Peter Horstink, used with permission
click image to enlarge
Horstink's observation is not the only observation. Very similar photographs showing a spiral in the sky have been coming from the ground in Sudan, e.g. in this tweet:




image from the ground, from Sudan (author unknown)

On rumours that Zuma failed

The sightings from Sudan near 3:15-3:20 UT are significant, as in the late afternoon of the 8th, rumours appeared on Twitter of a Zuma launch failure. These rumours then were picked up by some news outlets, e.g. here and here.

I have no idea about the veracity of these rumours, and so far SpaceX has said the mission was "nominal" (indicating no problems with the Falcon 9), while Northrop-Grumman and the US military haven't given comments (they never do about classified mission status). They could very well just be rumours, perhaps born out of a misunderstanding of events in the launch seen from the ground by lay observers.

For the moment, unless the US Government comes with some statement, I think it is wise we should treat it as "just rumours", and not necessarily true.

The sighting of the Falcon 9 Upper stage venting 2 hours 15 minutes (1.5 orbit) after launch, bear significantly on the discussion, as it seems to confirm the remarks by SpaceX that the mission was nominal. Of course, for SpaceX the mission ends at orbit insertion.

At any rate, it shows that at least the Upper stage achieved orbit (so it was definitely not a launch failure where the rocket failed to achieve orbit), and it makes sense that the payload then did as well.

So if something went wrong, if at all (a big "if" - I am skeptical), then there are three options left:

(1)  Zuma was inserted into orbit, but it is in the wrong orbit (too high, too low); 

(2)  Zuma was inserted into orbit, but is "dead", i.e. non-responsive;

(3)  Zuma achieved orbit with the Upper Stage, but failed to detach from the Upper Stage, and next de-orbitted with the Upper Stage near 3:30-3:45 UT.

JSpOC ("NORAD") did enter an object from this launch into its master catalogue on January 9th, as object nr. 43098, COSPAR 2018-001A, name USA 280. They designated it "PAYLOAD" (and the USA 280 designation would point to this as well). As usual for classified missions, they do not give details on the orbit.

screenshot showing the JSpOC master catalogue entry for a "PAYLOAD" named USA 280 associated with the launch

This suggests something achieved orbit long enough (i.e. over more than one orbit) to be detected and added to the catalogue.

While this does not necessarily mean the object is still in orbit (and it could in theory reference the Falcon 9 Upper Stage, with the "PAYLOAD" designation then in error), it does fuel my skepticism towards the truth of the rumours.

If Zuma is on-orbit but did fail, the situation becomes reminiscent of the USA 193 saga - an experimental satellite launched in December 2006 that failed after orbit insertion, and a year later was shot out of the sky with an SM3 missile, which has become infamous as "Operation Burnt Frost".

With regard to the observed fuel dump/depressurization: this is normal for most launches and does not necessarily indicate something's wrong.

Rocket stages always carry excess fuel, as you don't want the engine to cut out prematurely by running out of fuel. So it always has a sufficient fuel margin. Once its work is done, this excess fuel is often vented, also known as "depressurization".

[update] An earlier example of such a spiral resulting from a Falcon 9 venting fuel after launch into LEO, is this one from a SpaceX Falcon 9 test launch of a DRAGON in 2010. So this event over Sudan is not unusual. [end of update]

Depressurization and fuel venting avoids the risk of the rocket stage blowing up, for example as a result of static electricity building up in the rocket stage. You do not want your rocket stage to blow up, as it creates an uncontrolable swarm of debris and includes the risk that particles are ejected into orbits where they do not decay quickly,  adding to the space debris risk.

The spiral pattern results when the rocket stage is spinning, perhaps as result of the fuel vent.

At the moment, Zuma is not visible from the Northern hemisphere because all passes are in daylight or earth shadow. This will change 1-2 weeks from now, depending on the exact orbital altitude. The sighting from Sudan does confirm the orbital plane the object should be in (that is: unless it did a manoeuvre into another orbital plane after separation from the Falcon - but I doubt that). So we have to wait now untill a new object is observed in this orbital plane.

The hunt is on!


UPDATE: some news sources are now claiming sources within the US military and US Government confirm the failure, saying the second stage of the Falcon 9 "failed" and stage and satellite crashed into sea.

This does not tally with the observations over Sudan, which show the Upper Stage did reach orbit. So my skepticism remains. If there is some truth to it nevertheless, it could point to option (3) above and subsequent misinterpretation in the press.

UPDATE 2: the adapter mating ZUMA to the Falcon 9 Upper Stage was not made by SpaceX, but by Northrop-Grumman itself (which is somewhat unusual). So if ZUMA did not separate from the Falcon 9 (and did a dive into the Indian Ocean with it), the blame is not on SpaceX but on Northrop-Grumman. In that case, the SpaceX declaration that the Falcon 9 performed "nominal" is correct, even if Zuma did not separate from it.

UPDATE 3 (17 jan 2018): It turns out that a ~52 degree inclined, ~660 km altitude orbit also fits the constraints of the de-orbit area and being over East Africa at the right time. So we are adding that option to the search efforts. I did a partial plane scan of the 50-degree orbital plane two days ago.

click map to enlarge

Ackowledgement: I thank Peter Horstink for his report, for providing additional information on request, and for the permission to feature his images on this blog. I thank Govert Schilling for bringing me into contact with Peter Horstink. The photographs with this post are (c) Peter Horstink.

Friday, 8 July 2016

MUOS 5 stuck in GTO

The website Spaceflight.com has broken the news that something has gone wrong with the orbit raising manoeuvres of MUOS 5. They have therefore been halted for the moment. A formal statement by the US Navy on this all is here.

So MUOS 5 at this moment appears to be stuck in the aproximately 15240 x 35700 km Geostationary Transfer Orbit (GTO) in which Paul Camilleri and me observed it between July 3 and 5 (see my previous post).

It is the white orbit in the plots below (replaced with new versions July 9):




Orbit in TLE form:

MUOS 5                                               15242 x 35703 km
1 41622U 16041A   16186.93646397 0.00000000  00000-0  00000+0 0    08
2 41622   9.8319 324.4682 3211964 178.4686 182.8307  1.52727671    09

rms   0.003   from 14 observations Jul 3.46 - Jul 5.57 (arc 2.1 days)