of course, there are still many points that are garbled in the machine translation, so I am very much looking forward to Pandaneko's help!

Akatsuki failure report by JAXA (17 December 2010)

Contents

1. Akatsuki mission outline

1.1 Mission objective
1.2 Project's objective
1.3 Development policy

2. Probe outline

2.1 Comms antenna diagram
2.2 Attitude control system
2.3 Propulsion piping system
2.4 Attitude control policy during OME
2.5 Orbit policy
2.6 development schedule
2.7 Development - within JAXA
2.8 Development - Domestic and international cooperation

3. Failure outline

3.1 Main events after launch
3.2 Orbital insertion operation plan (VOL-1) and reality
3.3 Graound tracking stations
3.4 Ground tracking stations in time line
3.5 Telemetry data (unexpected) around VOI-1
3.5.1 Main telemetry data about attitude control around VOI-1
3.5.2 Accerleration record around VOI-1
3.5.3 Attitude record around VOI-1
3.5.4 Angular velocity record around VOI-1
3.5.5 Accerleration and angular velocity during latter half of VOI-1
3.5.6 Propulsion system pressure record around VOI-1
3.5.7 Propulsion system record after VOI-1
3.6 Events record around VOI-1
3.7 Probe behaviour estimated from telemetry data

4. Current status of the probe

4.1 Orbit information after VOI-1 control
4.2 Measurement devices and their health

This particular PDF is about 3.5 mega and has a lot of pictures and graphs. As I do not wish to refer to them with numbers I first translated the contents page so that I can copy and paste and upload every single (numbered) page with pictures etc. That way, I think it will be easier to digest what is what.

Also, that will be much less burden on the system, I think.

There is a much shorter ducoment, but it will take a while before I get around to it. It is going to take some time even before I translate the main document.

Pandaneko

3. Outline of the failure
Akatsuki tried its orbital insertion manouvoure on 7 December by the OME retro firing (VOI-1)
However, as will be stated later OME firing was stopped and only about 20% decceleration was achieved and due to that Akatsuki overshot Venus gravitatinal zone and entered solar orbit.
It has been confirmed that Akatsuki is now in a solar orbit of about 200 earth days, nearest point is about 90,000,000km, and furthest point 110,000,000km from Venus.
This particular orbit may allow another close contact in 6 years time.
What follows is the summary of activities Around VOI-1.

3.1 Main events after launch
21 May: Launch and insertion into transfer orbit
June: Initial checking of devices
28 June: OME test manouvoure 1, 13 seconds of bruning, 12.1 m/s against planned value of 11.7 m/s
July: Dynamical model estimate (with NASA and JPL) in order to decide on the details of precision orbital insertion
August: Comms performance evaluation for revived ? distance measurment method 2
September:
October:
8 November: RCS firing for delicate orbital manouvoure (1, for 21 seconds)
22 november: Same as above (2, for 2.1 seconds)
1 December: Same as above (3, for 0.4 seconds)
7 December: VOI-1 , but failed due to insufficient
Note 1: Test manouvoure: Firing for confirming OME orbital health, operational sequence confirmation, propulsion allignment confirmation. This allows total propulsion performance test, but due to short firing unable to obtain fuel supply characteristics
Note 2: Revived ? distance measurment method, removes noise from the received telemetry data and it is sent back to earth.

With section 3.2 I have a problem. This page is an image and no matter how I tried C&P I simply failed. Fortunately, these boxes are numbered. So, my translation referes to these number on the original JAXA Japenese PDF file, which was their status report to the Space Activities Commission (SAC).
3.2 Orbit insertion (VOI-1), planned and what actually happnened
Hereafter box numbers follow.
1. 06:10 on 5 december: Commands pre-sent
2. 07:50 on 6 December: Attitude changed to orbitalmanouverouring eigine (OME) firing, and anntena was switched to MGA
3. 08:49 on 7 December: OME burn started
4. 08:50:43 on 7 December: earth eclipse interrupted comms
5. 7 December: Planned: 09:01:00 OME to be stopped, reality: 08:51:38 OME stopped
6. 7 December: 09:12:03 Earth eclipse came to an end
Planned: Comms re-starting, reality: comms not recovered
7. 7 December: Planned: Transfer in the shade from 09:36 to 10:40, reality: 09:55 to 11:04 battery operated transfer in the shade
8. 7 December: reality: radio signals received from LGA
9. 7 December: 10:59 to 12:09, planned: Antenna direction to Earth, switching to HGA, reality: antenna switched to MGA, comms not re-established
10. 7 December: 16:00 on orbital data acquisition, reality: 23:32 on orbital data by LGA

Phew...
Pandaneko

3.3 Ground based tracking stations' activities
1. Usuda 64 m DS antenna tried, immediately after VOI-1, but anntenna data prediction ( frequence and direction) was not right and they could not capture any ignal.
2. After that, an off-set was imposed on the frequency and re-tried. About one hour and half we received a very weak signal, then tried again without frequency modulation, but the signal was so weak, and even repeated attempt at frequency modulation failed to get telemetry data.
3. After that, reception was transfered to NASA/DSN and we repeated anntena status predictions. Madrid managed to get weak signals, but we could not trace and telemetric data.
4. Reception was then switched to Goldstone
5. The day after VOI-1, Usuda's 64m and Canberra tried and we then succeeded to get telemetry data and carried out data reproduction.
Thank you very much, NASA DSN Canberra! Much appreciated!

Pandaneko

Thank you, Pandaneko

The level of detail in the JAXA report is great, because we see not only the development of the problem; we also get a feel for how difficult it is to reestablish communication with a probe this far away.

We look forward to further updates!

lol!

thanks Pandaneko, that slide looked like being one of the most interesting and it was too bad not being able to copy and paste into the translator!

With 3.4 I have difficulties as this is an image. Basically, this table decribes events on 7 and 8 December, that is ground station trackings.

The lefthand column, top one is time and date, next one down is ground statiions engaged, next one down is main events by these, the bottom one is Akatsuki anntena used.

Hereafter I will have to be descriptive.

Ground stations:

First, I will start with ground stations. From left to right, but by going down in between as you see on the table.

Usuda (located near my mountain cottage in central highland area of the main land), Uchinoura (this is the usual launch site for ISAS probes, down south), Madrid, Goldstone, and up again to Usuda and Uchinoura, in that order.

Main events:

Here, I see raws, going from left to right, so:

1st raw (from left to right)

08:49 OME firing, 14:04 One way capture, 22:36:00 Two way capture

2nd raw

08:51 Firing stopped, 17:08:59 One way capture, 23:32:00 Two way Doppler capture, 05:50:00 Two way telemetric data capture

3rd raw

08:55:15 Safe hold mode, 20:50:22 One way capture, 01:34:00 Atittude corrected for best reception and safe hold mode with HGA pointing to Earth, 10:09:00 Data reproduction

4th raw

10:26:17 One way capture, 15:48:00 One way capture every 10 miniutes

5th raw

13:10:00 Weak signal received by Usuda and Uchinoura stations, weak signal reception and a spectrum analyser was used to confirm unmodulated signal

The bottom raw must be clear enough, LGA, MGA, and HGA business. However, there are explanations about signal capture.

1. One way capture: Unmodulated signal reception
2. Two way capture: Uplink synchronised unmodulated signal reception
3. Two way Doppler capture: Data for orbit decision making obtained
4. Two way telemetry capture: Telemetry data obtained

Pandaneko

This evenig here, I also tried to translate 3.5 as well, but unlike before I could not do usual C&P. 3.5 is a document, so I thought it would be easy, but

I had to do area selection first and copy, and even that did not allow me to copy the whole page in one go and I gave up. I did try segmented C&P but even that did not work. Perhaps, tommorrow might be different, something wrong with the system? I do not know...

Pandaneko

3.5 Un-anticipated telemetry data around VOI-1
In the attitude maintain mode and safe hold mode after VOI-1 part of telemetry data concerning probe attitude is not recorded (Refer to section 3.5.1 for obtainable data type and time zone)
In this section we show data (type and time) which was un-anticipated around VOI-1.
> according to control mode history probe went into attitude maintain mode 158 seconds after OME burn and 375 seconds later into safe hold mode (See 3.5.1)
> re deceleration we expected increased deceleration after OME firing because of lighter mass due to fuel consumption. However, in reality deceleration decreased slowly and changed rapidly after 152 seconds. (See 3.5.2)
> Attitude angle history shows that the probe is pointing in the right direction until after 152 seconds after OME burn (See 3.5.3) and this agrees with the angular velocity history (see 3.5.4)
> In particular, angular velocity after 152 seconds of burning changed from increase to decrease (155.5 seconds to be exact) (see 3.5.6)
> Re propulsion system pressure history fuel tank pressure continued to decrease during OME burn (See 3.5.6)
> Fuel tank pressure recovered slowly after 158 seconds of OME burn (see 3.5.7)

Pandaneko

3.5.1 Control mode and major telemetry data around VOI-1 (Control mode history)

This is an image page. This page contains two major structures, a table and a set of bands. Hereafter, raws refer to those comments from left to right.

Raw 1: (From left to right except empty spaces)

1. Stable control mode
2. 152 secs, decrease in decerlation and lost attitude
3. 158 secs, transfer into attitude maintain mode (OME stopped) (This, strictly speaking between 157.625 and 158.625 seconds)
4. Irrelevant and not important

Raw 2: Orbit control mode

Raw 3: Attitude maintain mode

Raw 4: Safe hold mode

Raw 5:

1. - 3 seconds, ??? link started
2. 0 second, OME firing started
3. 3/5 seconds, transfer into safe hold mode

(Time in seconds relative to OME firing)

Hereafter, translation of the bands, second section from the left is reported.

1st band: Reproduced attitude system telemetry, deceleration, IRU rate (8 Hz)
2nd band: Estimated attitude angle and angular velocity (0.5 Hz)
3rd band: HK telemetry reproduced, IRU rate, tank pressure and temps (0.5 Hz)
4th band: Real (whatever it may mean) (1/32 seconds)

I have a horrible feeling that the original document may be being altered as days go by, initially, its size was 3.2 Mega, but now it is down to 1.8.

Pandaneko

This report is indeed quite detailed - JAXA must be commended for its openness here.
Pandanenko's translations are very welcome.

As with other foreign literature, I can at least look at the pictures - these are rather telling.
First note the architecture of the propulsion system
Click to view attachment

The history of these parameters (P1...P4) should be something as follows. P1 (the helium tank pressure) should
fall during the burn, as the pressurant fills the space above the fuel and oxidizer as these are used up. P2 is the
pressure upstream of the fuel and oxidizer tanks and is controlled by the regulator (which I see is redundant RG1
and RG2). This pressure basically determines the flow rate of fuel and oxidizer and thus the thrust.
P3 and P4 are the pressures downstream of the fuel and oxidizer tanks - these should not be too different from
P2 (and thus each other).

But now look what happens during the burn
Click to view attachment

P1 drops (not sure whether the amount by which it drops is reasonable - this will depend on the respective
volumes of the pressurant and fuel tanks). Odd that it seems to drop well beyond (to 3000s) the duration of the
(truncated) burn itself - 152s, we'll come back to this
P2 behaves more or less as it should - stays constant as the regulator meters the helium flow to fill up the space
as fuel/ox is used. Similarly P4 (oxidant downstream) follows P2
P3 - the downstream fuel pressure - is the anomaly. It drops substantially during the burn (and implies the fuel
flow also reduces). This could be a downstream leak, or an upstream constriction (which, since P4 tracks P2 pretty well,
must be between P3 and P2 - the CV-F valve or the pipework around there seems a likely culprit. Either way, the flow
from the regulated P2 is not keeping up.

Once the burn ends, P3 creeps back up - suggesting that gas is managing slowly to get to the fuel tank - it is not a
complete blockage. This is consistent with P1 dropping as the regulator allows some gas through to make up the
ullage volume opened up by the fuel usage. With knowledge of the tank and pipework volumes, one should be able
to calculate the leak flowrate.

So, the engineering forensics seems to make sense - regulator and everything upstream of there was fine, oxidizer
side was fine, but somewhere between regulator and fuel tank exit there was a constriction.

As the fuel pressure dropped, so did the fuel flow (by ~30%, to judge from the pressure record).
Thus the thrust declines. Normally, for a constant regulated pressure,
the thrust should be constant, and thus the acceleration of the vehicle will increase with time as the mass of the spacecraft
falls as the fuel/ox is used up. Yet here the acceleration drops, so the thrust must be declining along with the fuel pressure
Click to view attachment

Then at 152s, the acceleration drops suddenly, implying a massive loss in thrust. Not only that, but the attitude history
shows that something suddenly starts to torque the spacecraft around. This history
Click to view attachment
shows the rate about the X-axis (blue curve, orthogonal to the motor thrust axis) increases from 0 to 12 deg/s in about a third of a
second (presumably at this point the ACS threw a fit and cut off the burn and initiated safing). 12 deg/s is ~0.2 radians/sec
so to get there in 0.3 sec implies an angular acceleration of ~0.6 rad/s2. Taking the moment of inertia of the ~500kg
spacecraft as a cube of dimension 1.4m gives ~160 kgm2, and thus the torque must have been 0.6*160 = 100 Nm. If
the torque acted 0.7m from the center of mass (much more than that would be off the body of the spacecraft), then
a thrust component orthogonal to the line to center of mass must have been 130N, if my midnight math is right.

For a 500N thruster, this is a pretty severe misalignment. Note in the acceleration history above, at this point the axial
acceleration drops from 0.8 to 0.5 m/s2, suggesting the thrust component in that direction fell by 30%. If we take the
thrust as perhaps closer to 350N total than 500, since the fuelpressure was down, then if the thrust suddenly were
directed about 30-45 degrees to the side (if a large part of the nozzle broke off) this would give a component normal to
the axis of about 100N or so, and would drop the axial component by 30%, as required to match the torque and
acceleration record. None of this proves anything, of course - perhaps a meteorite hit the nozzle and the
fuel pressure just happened to be declining beforehand - but a coherent story is that the low fuel pressure caused
poor fuel flow; usually fuel is injected into the combustion chamber to provide film cooling; the poor fuel flow
interfered with this cooling action and a hot spot developed, cracking the nozzle and causing the thrust diversion,
which led to the attitude excursion, in response to which the spacecraft correctly safed; had it not, the spacecraft would
be spinning at some absurd rate - even without the autonomous termination of the burn, the mispointing would have
exceeded 90 degrees in a couple of seconds and so VOI wouldnt have succeeded)

Rather an interesting puzzle to solve (not that I'd not rather be working on Venus data from Akatsuki instead).

The exercise
shows (1) the value of engineering telemetry in diagnosing what happened, (2) that the fault protection on Akatsuki worked
well and prevented the situation getting worse. (3) If all the above is correct, any further
use of the main engine will cause large torques, as well as significant thrust losses. Furthermore, (4) unless the
constriction is cleared, thrust will fall appreciably for burns longer than a few tens of seconds.

It might nonetheless be
possible to use the engine in shorter bursts (and the pressure drop effect will become less troublesome as the fuel is
used up and so the ullage volume is larger), and by spinning the spacecraft during the burn to even out the torque.
I don't know whether VOI in 6 years will be possible given these constraints - presumably that is exactly the analysis that is being
done by JAXA now : as Hayabusa showed, if it is possible, JAXA will find a way.

(sorry for the long post)

"Sorry for the long post"?

Au contraire, Ralph; thank you for your generous, extremely interesting analysis of this data!

As you say, the situation is rather dire, but this seems to be when JAXA is often at its very best. Your 'spin-stabilization during burn' suggestion definitely seems like the way to go, here.

Wonderfull work Ralph and the spin idea does rock! I wonder if with enough engineering analysis they could build in some RCS firing to counter the off axis thrust as well as do spinning. Or could the RCS be used as a sorta kinda ion engine. Are the main prop tanks cross fed to the RCS?

I think JAXA shouldn't be counted out yet.

3.5.2 Decerlation history around VOI-1 (data rate: 8 Hz)

Here again, this is an image, but easier to understand as there are not many sections to translate.

Box outside the table says: OME started

Box just above it says: Attitude control started by RCS

Box just above it says: Settling by RCS

Here after I will translate the remaining boxes clockwise from the top.

Box underneath the long blue line says: Probe deceralation gradually dropped

Next box down says: Decerlation suddenly dropped and then increased to 0.62 m/s*2

The last box here says: Stable after 156 seconds

All these times are relative to the start of OME.

Pandaneko

3.5.3 Attitude angle history around VOI-1 (data rate : 0.5 Hz)

This is an image and simple to understand.

Blue is X- axis angle, red is Y -axis angle, and yellow ? is Z- axis angle. Bottom right box says: Rapid attitude change after 152 seconds

3.5.4 Attitude angle history around VOI-1 (data rate: 0.5 Hz)

Here again, this is an image and easy enough.

Blue line is the angular velocity around X- axis, pink is Y- axis, and yellows is Z, and the long box says: Stable attitude control for 152 seconds, and the bottom right box says: Angular velocity becomes 5 degrees per s*2 from 152 seconds on, rotation

Pandaneko

I am absolutely delighted about this, simply because I am not exactly sure about how to interpret what I have been translating so far and specialists' comments are just exactly what I want!, so thank you very much! By the way, the second document (shorter) to be translated is apparently how to cope with the failure (JAXA proposal), and of course, I will be going back to sections before failure status.

By the time I finish these off, we may be hearing about Hayabusa grains from the second can!

Thank you very much once again and a happy Christmas and new year for my colleagges on the forum!

Pandaneko

And happy holidays (and thanks!) to you, P, for providing us with these translations! Most of us wouldn't have a single clue about what was going on without you, man.

Quick question for Ralph (and possibly you as well, Pandaneko): Was there any indication of abnormal performance by the ACS itself during the burn? Reason I ask is that is that the plumbing between the P2 & P3 points is used to pressurize the hydrazine tank. I'm assuming that the 'ground-state' tank pressure was (and is) enough to run the thrusters, but it would be interesting--and perhaps confirmatory?--to know if the ACS was fully responsive during the anomaly.

IIRC one of the interviews/JAXA remarks was that no, the RCS thrusters were working (which is why I had earlier
discounted the regulator as a problem). But since these are monoprop, there are no mixture issues to worry
about - lower feed pressure likely translated into lower thrust, but then the duty cycle of the thrusters would just
be increased to compensate. I suspect there may be ACS diagnostics that can explore this - as you say, the
thrusters might be expected to underperform.

As for the suggestion to use the RCS to counter the torque from the asymmetric nozzle, that might be barely
possible (the RCS thrusters are 23N each, plus a couple of 2N ones) whereas 100N+ of orthogonal thrust is being
generated by the main engine. But even if it is possible thrustwise to counter the torque, it would be horribly
inefficient, costing a lot of fuel. Spin would be the way to go, I think

I hope spinning is a viable option considering the (vastly) increased RCS hydrazine expenditure that would be required for other methods.

Question: Since Akatsuki was of course not designed as a spinner, would we nevertheless expect that its internal mass distribution is still relatively symmetrical (radially, at least) in order to minimize RCS operation during OME burns?

the biggest problem that I see with a spinner is that the large flexible appendages (aka solar panels) would make it unstable. See for example what happened to the first US satellite Explorer 1.
I guess this would require some frequent stabilization burns by the RCS.

That wont matter if the moment of inertia about the thrust axis is the maximum (on Explorer 1 it was the minimum,
which is unstable in the presence of energy dissipation). The solar panels might even help in that regard.

There could be other issues like propellant management in the presence of spin, or attitude determination performance.
We'll see.

Something I am wondering about is whether a fuel pressure drop was noticed during the initial test burn of the engine -
it was only a few seconds, so by analogy with VOI the drop would have been too small to affect engine performance, but
the pressure sensors should have shown something.

I think if they spun it up s-l-o-w-l-y (and de-spun it the same way, of course), this might not be a problem. What concerns me is if the center of mass is not aligned with the spin axis it's going to precess & probably also wobble; the magnitude of those effects will vary with the amount & moment arm(s) of the non-centered masses.

However, if the precession/wobbling is minimal & well-understood, the spacecraft should still be controllable. Whether or not the ACS can cope with it within the limits of propellant expenditure to go on & complete a scientifically useful mission @ Venus would be the next big question to answer.

3.5.5 Decceleration and angular velocity during latter hapf of VOI-1

Again, this is an image, and data rate is 8 Hz.

1. Vertical axis on the left is angular velocity
2. Vertical axis on the right is decceleration

Blue line : Angular velocity around X- axis
Pink line: Around Y-axis
Yellow line: Around Z-axis
Black line: Decceleration curve

Again, time axis is relative to OME burn starting

3.5.6 Propulsion system pressure history around VOI-1, and data rate 0.5 Hz

1. Pink: Regulated pressure (P2)
2. Green: Fuel tank pressure (P3)
3. Dotted line with blue dots: Oxidant pressure (P4)
4. Blue line with dots: Gas tank pressure (P1)

Scale on left: Mega Pascal for P2, P3, and P4
Scale on right: Mega Pascal for P1

Horizontal time line: 0 sec= OME burn starting

3.5.7 Propulsion system pressure history after VOI-1 and data rate is 0.5 Hz

Here,

1. Green: Fuel tank pressure (P3)
2. Blue: High pressure gas tank pressure (P1)
3. Purple: Oxidant pressure (P4)
4. Pink: Regulated pressure (P2)

Scale on left: Mega Pascal for P2, P3, and P4
Scale on right: Mega Pascal for P1

Pandaneko

3.6 Events history around VOI-1

Here, we show events record from the telemetry data around VOI-1 as follows.

1. Minus 3 seconds: RCS attitude control starts (3.5.2)

2. Minus 3 seconds to 0 second: Settling by RCS (3.5.2)

3. 0 second (08:49:00 on 7 Dec, JST) : OME burn starts (3.5.2)

4. 0 to 152 seconds: Attitude control with +- 2 degrees accuracy (3.5.3)

5. Fuel tank pressure (P3) gradually decreases from 1.47 Mega Pascal to 0.95 Mega Pascal (3.5.6)

6. Decceleration gradually decreases from 0.91 to 0.82 m/s*2 (3.5.2)

7. At 152 seconds the deccerlation rapidly goes down (3.5.2 and 5)

8. Between 152 and 156 seconds deccerlation drops to its minimum value of 0.52 m/s*2 and then increases to 0.62 m/s*2 (3.5.2 and 5)

9. Angular accerlation around X-axis is 5 degrees/s*2, rotating, and maximum attitude 42 degrees, maximum attitude rate is 11 degrees/s (3.5.3 and 5)

10. During 156 and 158 seconds accerlation was stable at 0.62 m/s*2 (3.5.4 and 5)

11. Rotation about X-axis goes down from 11 degrees to 8 degrees/s (3.5.5)

12. At 158 seconds control was shifted from RCS orbit control mode to attitude maintain mode by the reaction wheels and at the same time OME valves were closed and burn stopped (3.5.1)

13. Oxidant tank pressure (P4) went up in stepps (3.5.7)

14. After 158 seconds in unison with the OME stoppage the fuel tank pressure (P3) started to increase slowly (at 158 seconds it was 0.95 M Pascal, after 2000 seconds it was 1.28 M Pascal, and after 6781 seconds it was 1.36 M Pascal (3.5.7)

15. At 375 seconds control shifted from attitude maintain mode to safe hold mode (3.5.1)

16. After 9660 seconds regulated pressure (P2) went down and became equal to the fuel tank pressure (P3) (3.5.7)


Pandaneko

3.7 Probe behavoiur estimated from telemetry data

Here, we have give an outline around VOI-1 with relevant telemetry data. There is no other telemetry data which shows irregularities. What follows are those which showed unexpected behaviours.

1. Fuel tank pressure which is supposed to be maintained at a constant pressure immediately fater OME firing continued to go down slowly

2. Rapid attitude change happened after 152 seconds from OME burn and at almost the same time probe deccerlation rapidly changed

3. There is a recod which shows attitude control mode change in response to OME burn stop after 158 seconds of OME burn

4. After 158 seconds from OME burn oxidant tank pressure started to increase in steps and fuel tank pressure started to go up slowly

Pandaneko

4. Current status of Akatsuki

Based on the telemetry data coming in from the probe we have examined;

power generation by the solar pannels, battery voltage, power consumption by each of devices, temperatures of these devices, reaction wheel revolution etc and all these subsystems are working normal.

Attitude control mode and fuel tank pressure, these are back to normal. Attitude is stable and HGA is pointed to earth and normal operation continues with regular communications with ground stations..

In order to check out the health of observing devices we captured a Venus image at a distance of 600,000km
on 9 December.

Also, in order to ensure good health of these devices for 6 years we have started to list up possible problems.

Pandaneko

4.1 Akatsuki orbit after VOI-1

Only 20% of the pnaned deccerlation was achieved by VOI-1 the probe failed to enter Venus orbit and escaped from Venus gravity and entered its solar orbit.

Current orbit information is 203 earth days revolution, nearest 9 thousand times 10 thousand km, furthest 11 thousand times 10 thousand km.

Since Venus's solar rotation takes about 225 earth days there is a close encounter in 6 years from now.

Pandaneko

There is a graph, but that is obvious enough,

Thanks yet again for your diligent efforts to keep us informed, Pandaneko!

It sure sounds like they're preparing to try again in 2016. After all the analysis is completed--and it sure looks like Ralph's take on it is pretty close if not spot-on--it will be interesting to see what sort of 'Venus round 2' maneuver testing they may conduct during the cruise.

4.2 Confirming observation devices' ability

LIR：Wave length 10μm
UVI：Wave length 365nm
IR1：Wave length 0.9μm

In order to check camera systems we used above three cameras at 09:00 (JST) on 9 December. LIR cannot capture stars we thought Venus was the right target.

At this point the distance was 60 times 10 thousand km. The angular size seen from ASkatsuki was about 1.2 degrees and Akatsuki is vewiing Venus from the night side. Since LIR captures thermal radiation the image includes Venus' night side, too.

This exercise showed that they were all healthy.

With 2μm camera (IR2）、thunder and atmosphelic light detection camera（LAC）、ultra stable oscillater (USO） we will be chaking them later on.

Pandaneko

There is a much shorter second document and I translate this next, before going bak to the large document.

This is titled;

Investigating causes and trying to come up with solutions

17 December 2010 by JAXA and ISAS

Pandaneko

Contents:

1. Overall plan and scope for finding causes and solutions

2. Looking into the causes of the failure

2.1 About the tree top of FTA analysis
2.2 FTA (Fauklt Tree Analysis)

3. Summary of the 1st report

Pandaneko

1. Overall plan and scope for finding causes and solutions

(Hereafter, Pandaneko will skip obvious statements. So, this page goes like;)

Akatsuki failed, so investigation started by setting up teams.

Note 1: Method used is Fault Tree Analysis (FTA)


Note 2: Wrting out detaled tree structures and reflecting on the evaluation results during development stages we will try to come up with an overall plan and scope for our investigation

Note 3: We will investigate the possibility of renewed insertion

(Please note that there are boxes on this page, but I will come back to them later, as I have to go out now)

Pandaneko

Boxes on this page step down from left to right. This particular report is concerned with the first box up top left (only). So, presumably there will be at least 3 more reports in the next year, I think. Anyway, boxes are as follows.

1st box says: Confirming the situation and laying out facts, creating FTA and wash out estimated causes

2nd box says: Narrowing down on causes and setting up an investigation plan

3rd box says: Investigation plan (analysis and testing), horizontal spread of the FTA tree structure, looking at background elements, tactics for re-insertion

4th box says: Report summary, reflecting on actions needed, investigation (analysis and testing), estimating possible causes

Pandaneko

3
得られたテレメトリデータから、FTAの手法で原因考察を行う。FTA解析の頂上事象を設定し、原因の推定を行う。
2. Looking into the possible causes of Akatsuki failure

2.1 Canopy events for FTA

In this particular report we talk about seperating out possible causes and irrelevant events.

Mulfunctions of Akatsuki may be summerised as a failure in inserting the probe into the right orbit. However, for reasons below we define the failure, for the purpose of FTA, as burn stop by detecting abnormal attitude.

1. Rapid attitude change was detected at 152 seconds from OME start. The systme is so designed that 5 seconds or longer of rapid attitude change will automatically activate attitude control mode change. (1-2 Sec 3.5.3)

2. At 158 seconds attitude control shifted from orbit control mode to attitude maintain mode. (When this happenes OME fuel supply valves are automatically closed) (1-2 Sec 3.5.6)

Therefore, we estimate that burn was stopped.

3. After 158 seconds, oxidant tank pressure went up in steps (1-2 Sec 3.5.6)

We therefore estimate that oxidant side valves were closed and as a result flow was stopped.

4. Incremental increase in deccerlation up to 158 seconds and orbit determination result are almost consistent.

(Orbit determination tells us that velocity increment was 135 m/s. On the other hand the deccerlation calculus up to 158 seconds (1-2 Sec 3.5.2) is 133 m/s and they are almost consistent)

It is also estimated that there was no significant propulsive power generation after 158 seconds.

5. Deccerlation between 156 and 158 seconds was almost constant (1-2 Sec 3.5.5)

OME burn was continuing up to 158 seconds and there is no evidence of propulsive system causing the burn stoppage.

Pandaneko

2.2 FTA analysis (starting here)

This is an image page and shows a lot of boxes. With this particular page I am not going to translate all those boxes now, because I need to explain the structure of the page first. So, box contents translation will follow after this explanation page.

Basically, this page starts out from A, the canopy of the tree as they know it right now (How do they know?), I think, shaded boxes are likely culplits.

You see a neat column of boxes on the right grouped together. With each of these boxes the first entry is either a cross, or a triangle, cross being negative, and triangle being possible. These are verdict boxes from top to bottom.

My translation scheme is as follows.

The boxes on the left (i.e. boxes to the left of verdict boxes on the right) , the whole entries are raw structured and talk about what might have happened (There are gaps, of course, depending on which raw we are looking at and I will simply ignore these gaps and say, something like (from left to right) box1, box2, box 3 and then >>> verdict, followed by the reason), from top to bottom. Each raw always leading to the verdict and reason box further to the right.

So, what I will do is to translate what is siside the boxes from left to right, from top to bottom and make sure I end up with verdict marker, let me use >>>, and the format of the verdict box after >>> will be the verdict (cross or triangle or whatever), followed by the reason for each verdict.

Would that be OK with colleagues? I find this page very, very interesting...

Pandaneko

Pandaneko, I think that it's very safe to say that any translation/interpretation/analysis you offer to us will be gladly accepted. Please feel free to define your own protocols for doing so; you certainly don't need to ask our permission.

This news just in, 15:11 27 December 2010 by the Yomiuri local here.

JAXA reported to SAC that Akatsuki's failure was due to the mulfunctioning of the valve used in the fuel pressure piping.

They were looking at 5 possibilities such as insufficient cooling of the main engine etc etc, and all of these cases were judged to come from this valve. As a result there was a burn anomaly in the main engine, it is estimated.


This particular valve was a standard item used for space use and there was no alterlation specially made for Akatsuki. They did not know if there was a problem as the test burn in June was so short.

JAXA will examine the reason and carry out the damaged nozzle simulation experiment and find out the possibility of re-insertion in 6 years.

Pandaneko

Very significant information, thanks! (Test burn duration/circumstances was one of Ralph Lorenz's questions.) They do seem to be pursuing a very logical anomaly identification/resolution strategy, and not to be repetitive but this is one of the very great goods that can come of unfortunate occurrences in UMSF.

In terms of engineering, we learn very little from successes in comparison to failures. (And by 'failure' I mean unexpected system behavior only; the Akatsuki mission is by no means a failure, and by no means is it over...not even close!)

There is another piece of news in today's Mainichi newspaper, in Japanese.

I am now very pessimistic about 6 years from now on. This article says;

Akatsuki's failure is due to closure of reverse flow valve, JAXA said.

JAXA said on 27th Dec that the failure was due to the closure of the valve within the fuel supply piping. The valve cannot be operated from Earth. JAXA will do ground testing to see if firing in 6 years is possible. The valve is inside the piping and operates with a coil spring.

JAXA think that for some unknown reason this valve remains closed, leading to insufficient fuel supply and the wrong mixture ratio, leading to abnormal burning.

This view will explain all other abnormal records, it says.

Pandaneko

I do not think Akatsuki will make it..., sad

Thats odd, having a one time operation valve, not reversable by a command.

there are more pdfs released today by JAXA
http://www.jaxa.jp/press/2010/12/20101227_...akatsuki_j.html

Pandaneko, clarification re the valve: Do they mean that it is stuck in its current position & no longer responding to commands?

Thanks, Pandanenko, this is very useful information.

The valve in question is CV-F on the schematic. 'Check Valves' or 'non-return valves' are the fluidics
equivalent of diodes - let the fluid flow one way, not the other (thus you wouldnt expect to have to
command them). Usually a ball is held on an aperture by a spring - pressure one way just holds the ball
onto the gap more strongly and no flow occurs; pressure the other (desired) way causes the ball to push
against the spring and opens a gap allowing flow to occur (as long as the pressure difference is big
enough to overcome the spring).

It sounds like the ball somehow got stuck. Quick tap with a hammer might unstick it.

If the valve cannot be fixed (maybe it would just fix itself, or temperature cycling, or something
distressingly nondeterministic might make it happen - there are 6 years after all) then the only thing I
can think of is as follows, and this only works if there is an awful lot of fuel margin. (For those with a
copy of 'Space Systems Failures' to hand, this is reminiscent of the 2002 recovery of TDRS 9, which
also had a pressurant valve failure.)

Do lots of short burns to use up all but the minimum fuel needed for VOI and some minimal operations
(these burns will have to be arranged to not provide a nondesirable deltaV, and the burns will have to be
short enough to not drop the tank pressure too much, and separated by long enough in time to allow the
slow leak of helium past the check valve to restore the pressure to the regulator value between burns).

As the fuel gets used up, the space in the tank above the fuel ('ullage') increases, and so the pressure
drop per unit fuel use will decline. (i.e. if at the burn the tank is 90% full, then using 10% of the tank
capacity doubles the ullage volume and drops the pressure by 50% ; but if the tank is only 50% full,
using 10% of the tank capacity drops the pressure by only 20%.. a charitable interpretation of the
VOI telemetry is that things didnt go badly wrong until the fuel pressure fell by 30%... but that was with a
burn that only lasted a fraction of the time it needed to). Some combination of setting up
the main burn this way, and perhaps towards the end switching to the monoprop thrusters which are not
as sensitive to the fuel pressure, might squeak us into orbit (of course we still have to be spinning
in order to average out the asymmetric main nozzle torque).

The viability of this approach really depends on the margins in the fuel load.

That sounds like a rather lengthy process. Do you think they can get the Δv out fast enough to make it into orbit?

2.2 FTA

Here on this page there are 6 columns. There is only one box at the leftmost column. I will start with this box. This boc contains the tree top event.

Raw 1: Box 1 (Column 1): Brun stopped when attitude anomally detected. (This connects to Box 2 on the same raw)
Box 2 (Column 2): Anomally in propulsiion
Box 3 (Column 3): OME stops at 152 s and torque generated (and this box connects to Box 1 on Raw 2 in column 3)

Raw 2: Box 1 (Column 3): (coming from the torque box just above it) Attachment anomally at 152 s
>>> Box 2 (Column 6): negative becauselaunch environment was within expectation, also attitude histroy ndicates
no large enough force leading to attachment deformation

Raw 3: Box 1 (Column 3): (this box is just below raw 2 box 1, attachment anomally) Thrust gas direction anomally at 152 s
Box 2 (Column 4) (coming from thrust gas box) Thrust gas chamber deformation
Box 3 (Column 5) Thruster nozzle throat break (This particular box connects to a small box which contains A)

Raw 4: Box 1 (Column 5): Burn chamber break
>>> Box 2 (Column 6): Negative, because we obtained near constant deccerlation just before VOI-s stop, and propulsion coeeficient estimated from deccerlation is about 1.3. Thus, burn chamber is not damaged.

Raw 5: Box 1 (Column 4, just below thruster gas channell) Thrust gas seperation/peeling
Box 2 (Column 5) Inner nozzle surface anomally
>>> Box 3 (Column 6) negative, because test manouver was completed normally and there are no other factors which will change the status.

Raw 6: Box 1 (Column 5, just below inner nozzle surface box) Throat rear burn
>>> Box 2 (Column 6): possible, because burn was made in abnormal condition

Raw 7:

Here, I give up for this evening. Apologies. I jotted down box contents on 3 pieces of paper and I am no longer sure how each raw and cloumn are related. I will start all over again, perhaps with a different scheme. I do not have an easy access to a printer.

Please note that therefore, above information may not be correct...

Pandaneko

Yes, this is from my non-specialist knowledge. However, I can at least confirm the meaning of irreversible, or non-reversible, I mean and they mean that the valve in question is a one-way mechanical valve with no control.

I have no idea how much ajar the valve is at the moment. It should not be closed compeletely, I hope, and my only hope is that by slow and long burning at the calculated rate through this small gap they might be able to place the probe into the right orbit.

Pandaneko

There may be ways to free up the valve through thermal management.

Probably what they want to do is to try to force thermal contraction of the ball and/or spring, given that the spacecraft is already in a heat-positive environment. That could conceivably be accomplished by shading the valve & associated plumbing from the Sun, but the systemic consequences must be carefully considered before doing so. The main issues to consider would be whether the necessary attitude to shade & cool the valve would 1] still allow the arrays to generate enough power, 2] still allow communication with Earth (antenna geometry), and 3] be even possible within the dependent constraints of the spacecraft's thermal control system. which presumably is designed to disperse excess heat to the shaded side of the vehicle.

Such situations really provide the acid test for the 'robustness' of a given system's design. And, yeah, it's pretty difficult to devise a pre-flight test for these sorts of things...

Good luck, Akatsuki team; we're all rooting for you!!!

Paolo, you are driving me nuts! That is a joke, of course. I continue to be amazed at your ability to find new source files. I quickly looked at the first files, the first on is a Q&A document from SAC members in response to JAXA 17 December report. The second one is a much more detailed report by JAXA dated 27 December. All these are very informative.

For instance, they talk a lot about this one way valve. One SAC member in A&A even suggested that FTA tree top should be shifted to something else and I think they did that, too. What I will do from now on is to finish off with the rest of their 17 December report, without translating the bits I ignored this time and start translating these new files.

I was, before the demise of Akatsuuki, naively thinking that I will be watching Venus images over the Christmas period with fried potatoes and a glass of wine. That joy is now long gone.

About the only joy I can get now is to chew up on the demise of Akatsuki in close details with specialists' occasional comments. It is another way of spending my evening times, I think.

Pandaneko