13
$\begingroup$

I wonder how the helicopter's rotors became damaged during landing. Yeah, I know, tips hit the sand. BUT...I would think a fairly detailed simulation would have been created for a full understanding of how to be better in future. How much slope can helicopter tolerate? How much rocking of craft happens as approaching sloped surface. How much nonsymmetrical ground effect is tolerable? Would legs 2cm longer have saved Ingenuity?

$\endgroup$
1

1 Answer 1

26
$\begingroup$

NSF/Intrepid Museum's interview with Travis Brown, PhD, Ingenuity's Chief Engineer contains some discussion on post-crash conclusions. I don't know if that actually falls under the umbrella of what you want to know or not; an interview certainly isn't as good as if there were a published analysis.

Starting around the 1hr20min mark, Travis starts talking about the lead-up to the crash and the crash day. I've added a couple links to the quotes to roughly the right timestamps in the video.

Sand dunes are our bread-and-butter because we can land in them. One of the things we were afraid of more than anything is landing on a rock. Sand dunes don't have a lot of rocks, so it's a great place for us. It turned out this sand dune just happened to be a little bit more rock-free than every other sand dune we've ever flown over. And so on Flight 71, we got a "Land Now," triggered an emergency landing, because the state estimation system wasn't able to get enough features to sort of get a sense for where it is.

... Our feature-tracker in particular is looking for rocks and pebbles. ... That actually caused Flight 71 to fail. We landed on the side of a sand dune, we came down pretty hard; we think we probably came down at about 1 m/s lateral velocity, possibly more. If the state estimator gets confused, when it's coming down it thinks it's coming down straight but it can be coming in sort-of sideways. ... To be fair, this thing was never designed for this. We never had the budget or the time or the engineering effort to be able to design a helicopter that could operate on every surface on Mars.

... So we got an emergency landing, we came down, and that spooked the team quite a bit. We spent three weeks, I think, looking at it and trying to rule out every other possible cause and eventually we're like "okay well it's pretty obviously just a nav issue with these sand ripples. We need to get out of here as quickly as we can, because this is dangerous terrain."

... To plan a safe flight, you have to know exactly where you landed, and you have to know exactly which direction you're pointed, because that's the starting point for the flight. To do that, we have to take off, and we have to take a picture of where we are to localize. And so Flight 72 was a very modest flight. It was take off, spend the bare minimum amount of time in the air, and come back down, send the images down, so we can plan our way out of this mess.

... So we were popping up, we were coming back down, and a couple meters above the surface, once again the nav system got confused from lack of features and--there's some speculation here--a little bit of speculation--that we may have locked on to our shadow. ... If your feature count gets low enough, such that the shadow now looks like the only reliable thing in the screen, now you don't just have No Data you have Wrong Data. And so we think that's what happened.

... We did what we could, we came down hard, harder on Flight 72 than Flight 71, and I think most people would be familiar with the result. ... It actually took us a while, we were so focused on the blade tips that we missed something pretty critical here--for a little while. We caught it eventually. ... It wasn't until a week later or so, when we were finally able to get some support from the rover and they were able to take some shots with Z-Cam... and we were able to confirm yes: not only are all of the blades damaged, but one is actually entirely missing. screenshot of the NSF stream showing Ingenuity's photo of its shadow showing damaged rotor blades, with video of the interviewers and interviewee to the right

A couple portions of the interview touch on specific things in your question:

Was there a tip strike?

Emphasis mine.

We had different theories about the sequence of events during the crash. Originally, we had assumed it was a tip-strike; that's usually what happens when helicopters have issues, and we saw the broken tips and so we're like "okay... the slope here is sloped up about 20 degrees, we had some nav error so we just came in a little steep and just went FWIT FWIT FWIT and clipped the tips." That was the original hypothesis. There's no tip strike impact here. screenshot from the NSF live stream showing Perseverance Supercam stills of the Ingenuity crash site, showing a shed blade and the sand dune region where Ingenuity is stranded

Either from our imagery or the rover's, there's no sign of a tip strike. And so the likely hypothesis now is that we just came down really hard, and the gyroscopic forces on the blades were more than it could handle. ... if you suddenly touch down and you have a very big change in roll, that exerts a huge gyroscopic force on the blade. ... So the leading hypothesis is we snapped the blade--chucked the blade way off to the west! You can see the original touchdown prints sort of to the left of the helicopter, and then once we have a blade missing this thing is horrendously unbalanced and just washing-machine'd itself down the dunes.

They go on in this section to talk a little bit about images showing what people supposed to have been tip strike impacts which are more likely to be gouges from the helicopter's feet skipping down the dune face.

Would legs 2cm longer have saved Ingenuity?

Likely not--both from "if you suddenly touch down and you have a very big change in roll, that exerts a huge gyroscopic force on the blade" and some following quotes:

It's actually kind of a miracle that we stayed upright and I would say it's another one of those "110% sure it's going to work"-type things, where they--our collaborators at AeroVironment who did a lot of mechanical design and testing, they tested the heck out of the landing gear. They had an entire landing gear testbed where they had a mass and landing legs and they just sat there dropping the thing from different heights and different slopes and all sorts of stuff because we weren't 100% sure we could trust the nav system to put us down softly, so we needed to be able to take a hard landing. So I think that's just kind of a testament to that engineering is that even after all of that: after losing a blade on a 20-degree slope, we still somehow managed to stay upright.

So the current hypothesis seems to be that the impact, as a combination of horizontal deceleration and sudden tilt onto the dune face, was too harsh for any amount of additional leg cushion to have saved the helicopter. It's not clear to me when in the crash sequence the team supposes the tip damage occurred but it seems that they don't think it's from actual blade strikes with the dune.

$\endgroup$
15
  • 5
    $\begingroup$ I did not know how to transcribe the sound Travis made for the tip-strike hypothesis, so if anyone has suggestions for that I'd appreciate it. Otherwise I think I did a pretty good job on the transcription. $\endgroup$
    – Erin Anne
    Mar 24 at 22:20
  • 5
    $\begingroup$ With one blade missing and the other three spinning at 2600 rpm, the vibrations must have been enormous. I have no problem imagining that the tips simply shook loose. I just finished watching the show and I think you did a very good job both with the transcription and the summary. $\endgroup$ Mar 24 at 22:57
  • 5
    $\begingroup$ This is great stuff. $\endgroup$ Mar 25 at 0:11
  • $\begingroup$ I had just finished watching the entire program when I came across your answer, I think your transcription is great. Two things to possibly consider adding, although I realize you had to work to get the length down and there is a limit. First is "may have locked onto our shadow, because frequently the shadow is in the field of view and it's not a problem because we have sophisticated outlier rejection. So all the features are moving together except the shadow, so you're like, that's garbage we don't care about the shadow we only care about the other features. But if your feature count… $\endgroup$ Mar 25 at 2:26
  • 1
    $\begingroup$ The other one was "so we think that's what happened. And we did do things to try to help, for instance flight 72 was flown an entire hour earlier than flight 71 to try to get a little bit more sun angle and more relief on the terrain bring out more texture. But ultimately it just wasn't enough. We did what we could..." $\endgroup$ Mar 25 at 2:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.