Chandrayan-2 Mission

[dinesha]

The speed at which it was travelling didn't give moon lander a chance: Expert
https://economictimes.indiatimes.com/ne ... 234359.cms

[ramana]

The speed at which it was travelling didn't give moon lander a chance: Expert
https://economictimes.indiatimes.com/ne ... 234359.cms

Lot of bolivating bokwas.
No real admirer of ISRO will talk like that.

[ramana]

They will have summary report.
No need for the back up or else folks like that Chicago person will talk more rot.

[Kashi]

Lot of bolivating bokwas.
No real admirer of ISRO eill talk like that.

Especially This bit...

Expecting for divine intervention is very much cultural phenomenon in India.

[nrshah]

Since this is a science mission all information should be made public even failure analysis report. There is nothing called space race ... any triumph is human triumph.

I hope ISRO will not dissappoint us. In fact the detailed failure analysis report should be included in the curriculum of undergrad and postgrad engineering courses by AICTE/UGC for relevant engineering streams be it aerospace, mechanical, chemical, computer science or even inter-disciplinary safety critical systems.

Tell that to NASA or ESA or Chinese and Russian.
We Indian have not taken contract of moral science solely.

[UlanBatori]

As shown by the sad example of the Expert who claimed I was mad because I said the lander would be designed to take "somewhere under" 10m/s and is (or should be) now dining on kauvva-kichdi, and other Experts who shall remain unmentionable, it is better to avoid speculation on whether UBCN is "serious" and just focus on whether one's casual assumptions are valid. Like by doing a Google Search .

Speaking of focus:

Pls see this paper from a few years ago. I very much doubt if that is the state of the art in places where they don't tell you what is the state of the art.GEO is 36000+km from Earth surface (OK, an order of magnitude in distance) and the imaging is through atmospheric clutter. I suppose one could do better, say, with an array of GEOsats aimed at the Moon.

When something is as stationary as the V-Kramji has been (except while being eaten by Godzillaji), many things are possible. The authors here speak of spatially separated elements, but same can be done with temporally separated elements if the object is stationary, or its motion (i.e., motion of the lunar surface w.r.t. imager) is known. So by their logic, it is possible to combine imaging from a number of different viewing elements, taken at different times, to synthesize one picture. If you just want a few pictures of ONE thing at ONE site, it is I believe possible to get that.

Now as for "synthetic color" vs. "Real Color", what is that, anyway? Everything you see on TV is synthesized from at least 3 imaging streams with primary colors and then mixed and manipulated. I bet the human eye also has RGB microlenses. Another of those "assumptions" which need to be corrected: there is no such thing as "Real Color". All is Maya onlee.
So nothing wrong with using, say. 0.3micron (UV?), or for that matter, 0.03 micron imaging, which would get that order of magnitude.
Human-built machine pieces like Vikram (pieces?) are easy to detect against older lunar debris.

[UlanBatori]

More on this: Hubble resolution if aimed at the Moon would be about 43 meters. James Webb may be better in some spectral windows. But my point is that if Hubble were to image the same area hajaar times from different perspectives (just as an example) you might be able to achieve the effect of a hajaar Hubbles spread out over a huge aperture. Each becomes only one imaging element of a huge lens. Sin- Thetic aperture, Ulan Bator style. Might even be able to reconstruct a 3D terrain map of the locality and see the legs of V-Kram sticking out. Point is, lighting is not in short supply.

I don't know how the infrared scatter from Vikram which is made of metal and plastic (with gold foil over some parts), would be different from that of its surroundings. Probably different: they can test that by looking at any NextOfKin still at ISRO labs. The solar panels would be highly absorbent, nearly black, unlike anything on the lunar surface since they have no known carbon there.

[SriKumar]

The speed at which it was travelling didn't give moon lander a chance: Expert
https://economictimes.indiatimes.com/ne ... 234359.cms

Forget his political comments, his technical comments are all over the place, no specific focus or train of thought. His comments on Vikram border on the comical in some spots. Projectile velocity, horizontal motion and vertical motion are three different topics ...apart from mechanics. . And he' asking if we know the coefficient of restitution of moon to predicts its Vikram's landing place after falling from 2.1 km. . Where does one begin in trying to educate him on mission design and spacecraft design. And the craft was never in freefall (for as long as the craft had telemetry). Its engines were running.

[UlanBatori]

Its the *&^%$#@* interviewer's fault. He was just wandering and wondering.

[Vayutuvan]

Since there was a bit of discussion on stability of the system which is related to the stability of the control system control, let me first link the Wikipedia page on OPtimal Control below as a starting point for my couple of posts on the topic.

Optimal Control

Important points to note from the above page -

1.

It is also noted that the optimal control problem as stated above may have multiple solutions (i.e., the solution may not be unique). Thus, it is most often the case that any solution {\displaystyle [{\textbf {x}}^{*}(t^{*}),{\textbf {u}}^{*}(t^{*}),t^{*}]} [{\textbf {x}}^{*}(t^{*}),{\textbf {u}}^{*}(t^{*}),t^{*}] to the optimal control problem is locally minimizing.

2.

Depending upon the type of direct method employed, the size of the nonlinear optimization problem can be quite small (e.g., as in a direct shooting or quasilinearization method), moderate (e.g. pseudospectral optimal control[7]) or may be quite large (e.g., a direct collocation method[8]). In the latter case (i.e., a collocation method), the nonlinear optimization problem may be literally thousands to tens of thousands of variables and constraints. Given the size of many NLPs arising from a direct method, it may appear somewhat counter-intuitive that solving the nonlinear optimization problem is easier than solving the boundary-value problem. It is, however, the fact that the NLP is easier to solve than the boundary-value problem. The reason for the relative ease of computation, particularly of a direct collocation method, is that the NLP is sparse and many well-known software programs exist (e.g., SNOPT[9]) to solve large sparse NLPs.

(My note: NLP is an acronym for Non-linear Programming problem, i.e. it is neither linear programming problem nor a convex programming problem both of which have (pseudo) polynomial-time algorithms to solve - for the former the most famous practical algorithm being the Interior Point Algorithm, which belongs to the class of algorithms called Ellispoid methods, by Narendra Karmarkar, with a sqrt(n) improvement by Pravin Vaidya, and the latter class of Programs can be solved by an algorithm - actually a family of algorithms - by Pravin Vaidya which guarantees an O(N) calls to a subroutine that answers a query about a point in an implicitly/explicitly defined polytope which is related to the feasibility set of the original Convex Programming problem)

So what does this all got to do with the System Stability of the Vikram Lander?

Multiple fixed points (or minima) is the key to understand what had happened.

Original optimal control problem (consisting of several feedback control loops) would have been assumed certain bounds for controlled variables and the mechanism would have had been designed those ranges (which appear as constraints). The importance of the range assumptions is that there an NLP becomes a small convex programming problem. These kinds of problems with a few hundred controlled variables can be solved by even an inefficient Conjugate Gradient or the weaker Steepest Descent method quite easily since convergence to the fixed point is guaranteed, i.e. there is only minimum.

So what has gone wrong?

(continued tomorrow)

[sudarshan]

...

Sir ji, my post was in response to your statement that "if you send a beam with x mm wavelength, you will get x mm resolution." This statement is plain wrong, and gives the wrong impression to any non-physics guys who come here to get gyan.

The other part of my post was simply to address your question of "how do they get images from hajaar million LY away?"

None of your google searches address the above incorrect statement. The techniques you're pulling up from google (spreading the light collection over space for instance) are well known, and not some deep gyan you have attained after "surpassing the textbook-mugging nerds." They are very much part of the textbooks themselves, and very much known even to lowly nerds.

As far as I know, you also can't improve optical resolution by doing that ensemble averaging you're talking about (100,000 scans and average to "remove noise"). I'm not 100% sure of this (this is not my field) but I don't think that the signal processing theory you're talking about applies to this problem, which is one of resolution limits imposed by diffraction.

If you want to take this post as yet another personal affront and respond with withering sarcasm in an attempt to make the offender eat "kauva khichidi," please be my guest.

[UlanBatori]

^^ So, since this is all so well known etc, what exactly is wrong with using several space-based telescopes, or the same telescope from different points, to achieve the same effect as a synthetic aperture antenna, please? The object is not going anywhere, and the general location is precisely known. The images acquired at every point can be stored and synthesized.

IIRC, this is how the Sony Kinect/ Playstation system forms 3D pictures of the players.

In general, I believe the resolution has something to do with the wavelength, the distance and also the aperture. A GEO-based telescope, used from many points to look at the same point, is an 88,000 km effective aperture diameter, hain? A bit larger than the biggest telescopes. 88000 km aperture is pretty impressive when looking at an object less than 400,000 km away.

FYI I really have no axe to grind here: I post ideas. Some may work, others may not work. When people come in all puffed-up and sneer, well... that can also lead to enjoyable outcomes If they get madder and get even more puffed up like in the previous post... that is their choice.

BTW, reading that paper might be illuminating. I don't believe that people have grasped what they were conveying.
Point is that 2 weeks after contact was lost, not a single image has appeared of the lost lander. And yet we know it is right up there, and we are looking straight at it when the Moon is full. Very frustrating.

[sudarshan]

^ Hmm, so when somebody starts a post with "UB ji" and then asks out of curiosity whether your post was serious or not, and then tries to explain the issue from first principles, that is "puffed-up and sneer." Could you please point out what part of my first post seemed like "sneering" to you? And when they point out the unnecessary sarcasm and injured tone in your response (part of which you have since edited out), that is "get madder and even more puffed up." Wokay!

Anyway, yes, you can do that spatial spread. I think it works best when there is no time lag between the light gathering at various points (i.e., with different telescopes) but the phase-matching might trip you up. If there is also a time separation, I don't think the recombination algorithm is that trivial anymore.

As you yourself point out, it is very frustrating that we have no images when we are looking straight at the lander when the moon is full, so that should tell you that maybe it's not that trivial to get an image. If you just want one picture of one site, it is of course much more doable than wanting the full moon image at that resolution.

But then - ISRO had years to prepare for CY-2, so they could have imaged the landing site in great detail themselves using the technique you point out, then Vikram wouldn't have had to gather its own images and do its own AI on them. So if they could have done that, don't you think they would have?

EDIT: I also indicated how to calculate angular resolution in my earlier post - it is related to wavelength and aperture (not just to wavelength, like you had said). Ensemble averaging might help to get the noise down (due to atmospheric effects etc.) and get the resolution down to the *best possible theoretical value,* but I don't think (again, not my field) that it is going to improve the resolution beyond the theoretical value.

[ramana]

I think we should look at the simplest explanation that fits the signature of the lander anomaly.
What's the signature?
1) The lander reverse thrusters experienced delta Thrust
2) Lander tumbled in pitch axis leading to the reverse thrusters adding delta V towards moon surface
3) Lander lost communications

So any scenario that fits the signature is the most likely cause.
To get to is cause we need direct evidence.

[Ashokk]

Experts find Sivan’s ‘98% success’ remark laughable

CHENNAI/NEW DELHI: Soon after Indian Space Research Organisation (Isro) chairman K Sivan reiterated on Saturday that Chandrayaan-2 was 98% success (citing the orbiter), some senior scientists have challenged the claim, one of them posting on social media his thoughts on leadership and rocket science.
A senior space scientist said such claims “without any deep introspection make us a laughing stock in front of the world”. Isro sources said Vikram, the Chandrayaan-2 lander, probably crashed at great speed and was lost forever.
Moon landing was the stated highpoint of the mission.

The talking point on Sunday was a social media post by Tapan Misra, adviser to the Isro chairman, who took a dig at Sivan’s leadership without naming him. “Leaders inspire, they do not manage,” wrote Misra, who was moved out of the post of director of Space Applications Centre, Ahmedabad, after Sivan took over as Isro chairman.

“When you see a sudden spurt in emphasis on adhering to rules, sudden increase in paperwork, frequent meetings, winding discussions, you surely know leadership is becoming a rare material in your institution. Institutions do not evolve with time as they stop innovating. Ultimately, they become living fossils, footnotes in history,” read the post.
A space scientist with expertise in moon missions told TOI on condition of anonymity there were technical mistakes in the mission. “Had Isro gone with a single thruster rather than five thrusters, the technology would have been simpler and easier to handle. This is a conventional configuration being followed across the world.”

[Pratyush]

What is wrong with so many Indian experts.

The lander was ment to drive for what 500 meters from the landing site??

The orbiter was ment to be in the orbit for atleast 1 year.

So how is it a massive failure?

How is the 98% success not a credible claim??

These are questions I have for the experts.

[suryag]

Folks no sniping please warnings are on the way if you indulge further

[ramana]

Misfortunately Mr. Tapan Mishra has not posted anything meaningful to the lander anomaly.

If single engine was so common why did ISRO use the four engine solution?
Because they wanted more precise control and attitude system.
Everyone can do easy things.
The landing site chosen was also difficult pockmarked with craters.


I still think that we need to wait for the FAB to issue their report of what went wrong.

Rest is time pass.

UB normally during an FAB, things won't get released.
Doesn't mean they don't have pictures, evidence etc.

[Vayutuvan]

What's the signature?
1) The lander reverse thrusters experienced delta Thrust
2) Lander tumbled in pitch axis leading to the reverse thrusters adding delta V towards moon surface
3) Lander lost communications

Let me continue my previous post with this as the jumping point.

When lander reverse thrusters experienced delta thurst, it excited modes which were not supposed to be there, especially so if the overperformance is pretty large and shortlived, which is more like an impulse function. That made some of the the feedback loops settle into another closeby local minimum, i.e. subspace, which makes the whole system unstable. If that happens stability is not recovered. Remember dynamics is governed by a coupled set of nonlinear diffrential equations.

My money is on either some structural flaw or non-homogeneity of the liquid fuel. In the latter case, richer (denser?) fuel might have gone into the overperforming thruster which set off the whole chain events.

[Vayutuvan]

If single engine was so common why did ISRO use the four engine solution?
Because they wanted more precise control and attitude system.
Everyone can do easy things.

The geometric stability of multiple engines reduces the feedback mechanism complexity. On the other hand, reliability goes down somewhat. One has to pick one's poison.

[ramana]

If people want to use my post please hypothesis a failure mechanism and show who the three signatures are met.
Lets to root cause analysis(RCA) right.
Format:
Hypothesis: xxxx
Evidence to support
Evidence to refute:

Analysis:

Evidence needs to support all three signatures.

Conclusion: Only four choices.
1) IS cause
2) Most Likely cause
3) Most unlikely Cause
4) Not Cause

[SSSalvi]

What's the signature?
1) The lander reverse thrusters experienced delta Thrust
2) Lander tumbled in pitch axis leading to the reverse thrusters adding delta V towards moon surface
3) Lander lost communications

----
My money is on either some structural flaw or non-homogeneity of the liquid fuel. In the latter case, richer (denser?) fuel might have gone into the overperforming thruster which set off the whole chain events.

Add this ..
One of the PSLV flight was doomed because a nozzle of engine was 1 or 2mm more in dia than the design value resulting in less torque.
Root cause .. Material incoming QC failure.

[chetonzz]

I think vikram lander has suffered the same type of crash (approximate velocity and angle) as "artistically" represented during opening scenes of this movie...I am sure we/LRO can track the marks on the lunar surface easily next month...

https://www.youtube.com/watch?v=bKXE93kgCcw

[Vayutuvan]

SSS, only problem the "manufacturing defect(s)" theory has is that it doesn't explain why the communication was lost (3rd point in the signature Ramana posted).

If there is a (micro)meteorite hit or a high radiation anomaly coupled with QC of the hardened electronics failing, that will fit the signature. in that case, we will not know unless the lander is somewhat intact, brought back, and analyzed.

A programming error, i.e. bug in the software, also fits the signature.

My hypothesis, most probably, is NOT correct

[SSSalvi]

^
Communication loss is the result of total disorientation of TM antenna aboard Lander due to tumbling of the platform ( Lander itself ).

The Landing videos in ISRO publicity materials show that the lander orientation changes from legs in front to legs at bottom. ( PITCH rotation ).During this transition 4 engine to 2 engine torquers also take place. A small mismatch in co-ordination of the whole system in this operation is sufficient to send the craft in PITCH tumble ( A very common malfunction seen in almost every country's first satellite experiment ) .

Of course we can't say that ISRO is novice in this maneuver ... it has been used time and again. That isthe reason that one can't think that it is a algorithmic failure... it has to be a hardware malfunction / Hardware has deviated ( at manufacture stage ) from its intended design parameters.

It can also happen if one of the opposite pairs of torquers is shutoff while the other one has has not shut of simultaneously.

[SaiK]

Is it going from 5 engines to 2 engines or from 4 engines to 2 engines for fine-braking?

q: why shut off 2 or three engines? why can't just adjust the throttle of all 5 firing !? [I read all 5 of them are throttleable - of course the minimum thrust to weight must have been known already.]

[neeraj]

Where are the orbiter pictures

[SaiK]

^the only oped reported one is from ISRO. NASA failed to capture it due to bad light! [no umpires required ]

[ramana]

somdev,
For each of your four scenarios tell us how will they match the 4 signatures of the anomaly?

[Amber G.]

Amber ji, is that video a simulation of the actual path Vikram took?

That was, as mentioned, animation, from data (photographs etc) from LRO database, and (known - or best guess) of CY2's ephemerides.

LRO Data, is being examined, pixel by pixel, but it is hard as the following picture shows:
(Snap from the LRO quickmap - which shows why it was difficult for LRO to take the picture of Vikram Lander. The sunlight! Shadows are already covering the blue dot where its assumed to be lying on lunar surface)



Sharing some more pictures: (Source NYtime, NASA)


^^^ Yellow rectangle: Suspected location of Vikram - LRO images taken in details.



^^^ Dotted line, orbit of LRO (on last Tuesday), White Square - South Pole, - Red square - Vikram landing site .. One can see Manzinus crater just above the red-square.

^^^

[Vayutuvan]

^
Communication loss is the result of total disorientation of TM antenna aboard Lander due to tumbling of the platform ( Lander itself ).

That explains the loss of the communications link. Either disorientation or some cable snapped off.

[kit]

a vintage pic to cool minds remember what Konstantin Tsiolkovsky said ( big fan here )

"First, inevitably, the idea, the fantasy, the fairy tale. Then, scientific calculation. Ultimately, fulfillment crowns the dream."

[Haridas]

^
Communication loss is the result of total disorientation of TM antenna aboard Lander due to tumbling of the platform ( Lander itself ).

That explains the loss of the communications link. Either disorientation or some cable snapped off.

I also mentioned the same, since High gain antenna needs pointing, the actuator motor has finite slew rate that is much lower compared that what a tumble can present.

[UlanBatori]

^^ Kudos.

[Vayutuvan]

^somdev,

how strong is the field at 2km height? I think milspec hardening will take care of this kind of radiation. that said, there could be a bad key component or an exposed connector.

reliability figures for these components will be needed to assign a probability of the kind ramana is asking, data like MTTFF for each component, subsubsystem, subsystem.

when doing system tests, ISRO would have exposed VL to intense radiation. they also,would have stress (added later) tested the engines, the antenna assembly structural integrity for shocks and vibrations.

[UlanBatori]

This means that they need to send a low-ballistic craft ("jaspreet") to probe this layer. Maybe send down 6 of them. Single thrust chamber valve-connected to 6 nozzles distributed around the ball. Gravity sensor to tell up from down. Don't depend on external e-mag signals.
Finally something other than regolith and iron - manganese deposits to study on the Moon.
But note: now the discussion has shifted to a scenario exactly like Godzillaji reaching up and grabbing V-Kramji.

[hnair]

UB-saar, fourth cousin had chai and:
- higher up (till now) is against release of pics because, well, with what they have, there might be none of shakinaw effect that is nowadays demanded of missions
- orbiter took pics allright. Visual spectrum might have glare/albedo/something causing lack of resolution (probably needs multiple passes under various light conditions to stitch up a half decent passport-photo of the lander)
- infra-red also was suboptimal: by the time orbiter flew past, the lifeless lander has cooled down to ambient temperature and melded into the backdrop
- visual spectrum however seem to show shadows of lander legs doing a Jean-Claude van Damme style sidekick

Added later: the earlier chai, just after the thud was that at last point of contact, the flame was pointing upwards. That flip causing anomaly is probably where the entire investigation is going on about.

[Vayutuvan]

1. Spacecraft surface charging leading to massive electrostatic discharge
2. Total ionising dose exceeds permissible levels
3. Displacement damage causing displacement of atoms from their normal lattice positions
4. Single event upset, latch-up or burnout
....
Like everyone else I am trying to think of what could have gone wrong!

Sure, I have no problem with that. I am just trying to qualitatively understand the radiation levels.

We also need to look at solar flare data and sun spot data in an interval straddling the time at which communications were lost.

[UlanBatori]

Added later: the earlier chai, just after the thud was that at last point of contact, the flame was pointing upwards. That flip causing anomaly is probably where the entire investigation is going on about.

Maybe someone programmed rotation by 270 deg. instead of 90 deg. JUST IDLY SPECULATING!!! Fortran in some versions could not distinguish between 1st and 4th quadrant when doing InBharj Tanghent unless u put ATAN2, and a pi/2 is a pi/2, + or minus. But if the flame was upwards at 2.1 km, it is a miracle that anything is left of anything to point upwards. So it kept on rotating maybe? And the legs are there, but rest of lander is in Godzillaji's tummy?

Too many unknowns. They don't know the gravity field too well. They don't know the dust field. They don't know the plasma field, They don't know the variations in surface magnetic field - and they live dangerously by zooming along at 2000 m. - or was it 300m and do they know?

Maybe Vikram landed smack dead center in the deepest ice lake on the Moon. Like the Valujet flight that dived at transonic/supersonic speed vertically into the Florida Everglades. Disappeared without a trace until months later.

They need to send the Jaspreets.

Landing in any orientation is a success then, and it can hop around wherever it wants. Take a mobile Phone foto each time it hits the ground and each time it reaches zenith. Maybe I should get the Ulan Bator Yak Madarssa to design this. If there is interest in the chaikadai.

[Vayutuvan]

Just to give a sense of the energy produced when something similar to VL crashes into the moon -

https://en.wikipedia.org/wiki/Beresheet

The final telemetry reading indicated that at an altitude of 150 meters the craft was still traveling over 500 km/hour (139 m/s), resulting in a total loss on impact with the lunar surface.[39][40]

I read somewhere else that this produced energy equivalent of 2 tons of TNT. I posted the before and after animation from Bersheet crash site from LRO.

If VL had crashed, it would have produced a before-after signature similar to the Bersheet. If anything, the altitude was ~2x and VL has higher almost 4 times the mass of the Bersheet lander. Of course, it was traveling slower than the Bersheet.

The autonomous system might have recovered but communication circuitry got fried, may be.