Thread by @DrPhiltill

I could write a 50 page paper answering this :)

A few points in outline form only:

1) The rocket exhaust is expanding into vacuum, so viscosity breaks down, so the gas does not obey the Navier-Stokes equation, which is the basis of CFD (computational fluid dynamics) models. /1

https://twitter.com/johnfkross2/status/1761623775873794259

2/ When I was at NASA, one of the things I was doing was writing solicitations to industry to write physics-based code to do CFD without Navier-Stokes. There are many ways to treat the fundamental physics (the Boltzmann Transport Equation) and they all work for different…

3/…approximations, but it is really hard to write a code that will handle the full range of conditions from dense gas inside the rocket nozzle all the way to rarefied gas on the Moon far from the rocket.

2) We don’t understand turbulence when the gas becomes rarefied.

4/ As the path lengths of the gas molecules becomes longer between collisions of the gas molecules, it takes larger distances for them to close a loop of motion, so turbulent eddies have to become larger in diameter. (Pic source: ) researchgate.net/publication/32…

5/ So the spectrum of turbulent kinetic energy becomes truncated at larger and larger sized eddies. When I first started writing research solicitations to industry 20 years ago there were exactly zero published papers on this topic. So there are ZERO available turbulence models

6/ …that you can use in a CFD model of gas flow in these conditions. Turbulence matters. It affects transfer of momentum from the rocket exhaust to the soil.

3) CFD models are generally crude and inaccurate for how they handle particulates mixed in the gas, especially for…

7/…very dense particulates, like when sand starts as a solid surface then erodes into lower-density, dispersed in the gas. We do not know how energy and momentum from the gas couple to the individual sand particles at that scale, so there are ZERO equations available for it.

8/ I have a paper in peer review that develops an equation down to the scale of an average sized sand grain, telling how energy flow from the gas down to that roughness height of the sand governs erosion rate, BUT… (pic: ) magnifiedsand.com

9/…even down to that size scale there is a parameter (“erosion efficiency”) that we cannot predict from physics yet, because it describes what the gas does at even smaller size scales between the sand grains, so we can only measure its value during lunar missions.

10/ That physics gets us to the rate that gas lifts the dust, but what about after the dust is lifted?

4) The existing CFD models all produce results that disagree with each other by a factor of TEN on how fast the dust is blown. That is crazy inaccurate! Why?

11/ Nobody has ever yet published a study to figure out why. It has only been the last few years that multiple studies have tried to predict the speed of the gas, so we only just found out that the predictions are all disagreeing. (I suspect less than 5 people in the world…

12/…have even paid enough attention to realize the studies are all disagreeing with each other!)

I suspect the problem is that CFD modelers are gridding the volume of space too coarsely in the region close to the soil, so when the code integrates the speeds of the dust…

13/…it is doing it in steps that are too coarse and thus are not good approximations of nature. But it is computationally expensive to grid more finely. These codes already require supercomputers as it is!

If we can’t predict how fast the dust blows, then we can’t predict…

14/ …how much they spread out as they travel away from the rocket, so we can’t predict how much the dust will scatter light. And then, when we try to measure the blowing dust during a mission, we have no way to relate the measurement back to calibrate the erosion physics.

15/ I had better stop there because this really could be a 50-page review of the gaps in the physics. I could go on about lift and drag force correlations on dense assemblages of particles, on the collisions transferring
momentum between the different sized dust grains,…

16/…on the light scattering effects of non-spherical dust particles of random mineral and glass compositions, of the amount and type of damage these particles cause when impacting various spacecraft materials at six times the speed of a bullet, etc.

17/ The basic problem is that here on Earth we do not deal with these regimes in physics. Dust cannot blow six times faster than a bullet because the Earth’s air stops it, so we never study the damage of hardware getting hit by dust at that speed. Fluid physics is hard, and it…

18/ …gets a lot harder in the extremes. Landing a rocket with supersonic gas on a dusty surface in vacuum incorporates many extremes, so putting them together is beyond anything we have solved before.

We can’t even do the right experiments here on Earth:

19/ We cannot fire a large-scale rocket inside a vacuum chamber while maintaining vacuum, with abrasive dust that destroys vacuum pumps, with the entire experiment inside a reduced gravity aircraft. (They won’t even let us fire tiny rockets inside airplanes 😄)

20/ Ultimately we have to get data from landing rockets on the Moon in order to guide the physics so we can develop models that start to become more reliable and predictive. For example, the IM-1 mission carried the SCALPSS camera system to measure blowing dust.

21/ With the Nova-C lander tipped over they do not have the high gain antenna pointed back at Earth, and with a weaker radio signal you cannot send data at a high rate. I hope the IM team is able to get enough data back to Earth to see results from SCALPSS. But if not…

22/ …then I hope we can get data from the next few missions. We have done almost all we can with the Apollo landing data. We need new lunar data!

/end