Thread by @DrPhiltill

The UAP debates reminded me of a personal story about handling evidence and judging likelihood. I think this is interesting & amusing.

🧵:

On Shuttle mission STS-128 the rocket exhaust blew out 3,500 tiles from the side of the flame trench. /1

2/ The bricks were smashing into each other as they blew, fragmenting into millions of pieces of all sizes. The results were devastating to the perimeter of the pad. This is what the security fence looked like about a kilometer away. See the brick fragments? 🤯

3/ My research group collected an unbiased sample of fragments so NASA Marshall could use a realistic range of fragment sizes in their computer modeling of the event. This is @Ryan_N_Watkins and John Lane measuring and weighing fragments. (I’m holding the camera.)

4/ Ordinarily, nobody would have worried too much about the safety of the Shuttle itself since the brick release was beneath the mobile launch platform and the flame trench did its job ducting everything off to the side.

5/ However, we had an experimental infrared camera mounted on the top of the VAB watching the launch, and it detected a single piece of debris flying UPWARD near the Shuttle. This raised the question, could this or other fragments hit the Shuttle?

6/ This was a big deal because we lost Columbia when a piece of foam slammed into the leading edge of a wing and broke a hole in it, so during reentry hot plasma flowed through the hole and the metal structure failed. Loss of vehicle and crew.😢(image: ) https://t.co/2N8InQiJ0cspaceflightnow.com/shuttle/sts107…

7/ When we saw a piece of debris flying near the vehicle in liftoff of STS-124, we went into emergency mode. We were given 4 days to decide if it would be safe to land the vehicle, or should it stay in orbit until a rescue mission can be launched to save the crew? Four days!

8/ I had the idea that we could start by identifying the object by measuring its ballistics. My colleague John Lane had written software to interpolate between the frames of film cameras so we could combine multiple launch pad views to make the measurement.

9/ The problem was that the launch pad cameras were film cameras and the shutters were not synchronized between any two cameras, so each frame of the video was staggered relative to any other camera. John’s software interpolated the frames to synchronize them post facto.

10/ John got the blueprints of the launch pad and also went out to the pad to verify camera locations so he could do the math combining the two views.

11/ From this, John calculated the three-dimensional coordinates of the debris versus time as it flew upward next to the launching vehicle.

12/ John also measured the diameter, elliptical elongation, and rotation rate of the debris as it flew upward. (You could see it spinning.)

13/ My part of the effort was to write software that did the ballistics calculations and performed “simulated annealing” to determine the best fit parameters including origin, starting velocity, and material density of the object.

14/ The material density that was the best fit was exactly the density of a type of foam used to fill the throats of the solid rocket boosters before launch — to keep birds and insects out of the rocket so they don’t accidentally light the propellant by heating it somehow!

15/ So we showed conclusively that the high flying debris was just “throat plug foam”, not a flame trench brick fragment. And since the shuttle was still going slow at that time, a foam impact could not hurt the vehicle. Whew! 😮‍💨 Well, what happened next?… 😁

16/ We got that done in two days, so we we got scheduled to present our results the next day to the program managers to say it was safe to land. We went to the launch control center and gave our presentation with one day to spare.

But then,…

17/ …some folks from another NASA center had come to the meeting and they stood up and said that our analysis was WRONG. They had proof. They had checked our math, and they said we did the trigonometry wrong when we combined the two cameras. 🤔

18/ so they said the two cameras were not pointing exactly as we thought, and therefore the two videos showing debris flying up were actually showing TWO DIFFERENT PIECES OF DEBRIS.

That’s what they said.

19/ Well I didn’t say this in the meeting (because the managers took our recommendation anyways and proceeded toward landing), but I thought this to myself…

20/ Okayyy, so we know somebody made a math error. Either my guy Dr. John Lane who is one of the best applied mathematicians in the world with works-class experience in photogrammetry, or your guy. One of them made a math error.

If your guy is the one who is right, then…

21/…that means there were 2 pieces of debris instead of just one. And both pieces of debris had to have exactly the same size, same elongation, same rotation rate, same rotation phase, same starting velocity, same altitude in each timestep, same density & same deceleration, AND

22/…our math error had to be the *precise* error that made the 2 pieces exactly overlap in the horizontal dimensions so it made them look like the same piece. And on top of this, the infrared camera only saw one piece, so one was hot and one was stone cold despite the flames.😜

23/ On the other hand, if it was your guy who made the math error, then it was just a math error.

When you consider all that, you don’t even need to see the detailed math to know which one of them made the math error.🙃

24/ We published our results in Acta Astronautica. The preprint is available on the arXiv, here:

I was surprised one day shortly after these events when we had a staff meeting and an astronaut walking into the room...arxiv.org/pdf/0910.4357.…

25/ …and he presented the Silver Snoopy award for solving a problem that made human spaceflight safer. I felt deeply grateful for being noticed by the flight crew. Still one of the highlights of the NASA career. 🥲

But my point for now…

26/ …is that sometimes you can tell if math is wrong without even looking at the details of the math. I still laugh every time when I think about the extreme coincidences we were being asked to believe if our math had been wrong.😆 /end🧵