Comments on: Rockets Shake And Rattle, So SpaceX Rolls Homegrown CFD https://www.nextplatform.com/2015/03/27/rockets-shake-and-rattle-so-spacex-rolls-homegrown-cfd/ In-depth coverage of high-end computing at large enterprises, supercomputing centers, hyperscale data centers, and public clouds. Mon, 11 Feb 2019 14:56:27 +0000 hourly 1 https://wordpress.org/?v=6.5.5 By: Andrew https://www.nextplatform.com/2015/03/27/rockets-shake-and-rattle-so-spacex-rolls-homegrown-cfd/#comment-112270 Mon, 11 Feb 2019 14:56:27 +0000 http://www.nextplatform.com/?p=491#comment-112270 In reply to Adriaan.

It seems with the rapid increase in output by the Raptor engine over the last week, which has been record breaking with the speed of development of a groundbreaking rocket engine.. it seems to suggest the CFD method has worked fantastically. To go from a 50% test to about 80% in a few days with their first ever full scale model of the engine, with still component testing a few years ago is unbelievably fast progress.

]]> By: 4i4o https://www.nextplatform.com/2015/03/27/rockets-shake-and-rattle-so-spacex-rolls-homegrown-cfd/#comment-3155 Wed, 03 Jun 2015 20:58:30 +0000 http://www.nextplatform.com/?p=491#comment-3155 In reply to Fred Willett.

most of this 24 minutes they talk about notorious facts, adaptive orthogonal meshing, multiresolution etc, etc. what actually they reinvent?

]]> By: 4i4o https://www.nextplatform.com/2015/03/27/rockets-shake-and-rattle-so-spacex-rolls-homegrown-cfd/#comment-3153 Wed, 03 Jun 2015 20:40:58 +0000 http://www.nextplatform.com/?p=491#comment-3153 “radical improvement” of efficiency of rocket engines is simply impossible because best of them ( Space Shuttle main engine, RD 180) already are very close to theoretical thermodynamic limit. Improvement from from 98% to 99% for example couldn’t be described as “radical”. Falcon 9 Merlin engines may be improved because of its modest performances but again not radical.

The future methane engine development may be benefited by more suitable CFD software replacing test burns with computational simulation

Ironically, any new CFD software must be tested and verified on physical model e.g. rocket engine.

]]> By: Adriaan https://www.nextplatform.com/2015/03/27/rockets-shake-and-rattle-so-spacex-rolls-homegrown-cfd/#comment-940 Thu, 09 Apr 2015 20:24:09 +0000 http://www.nextplatform.com/?p=491#comment-940 Just read an interesting article by guys from CRNS -http://wavelets.ens.fr/PUBLICATIONS/ARTICLES/PDF/123.pdf. Notwithstanding the fact that they have been working in using wavelets to model turbulence – for 10 years – the authors claim that the wave local fractal compression compares poorly against classical spectral methods for CFD. the conclusion is a classic!

In terms of this article I applaud the effort by SpaceX but lets give credit where credit is due – it is not their idea and neither did they perfect it.

Furthermore, I watched the Youtube video and I am quite skeptical about the results from this CFD method – where are the validation cases? Anyway, lets wait and see.

]]> By: Adriaan https://www.nextplatform.com/2015/03/27/rockets-shake-and-rattle-so-spacex-rolls-homegrown-cfd/#comment-939 Thu, 09 Apr 2015 19:56:10 +0000 http://www.nextplatform.com/?p=491#comment-939 Well actually wavelets local fractal compression for CFD was invented by two British researchers at Rolls Royce plc and patented in 2003. Just saying.

]]> By: Fred Willett https://www.nextplatform.com/2015/03/27/rockets-shake-and-rattle-so-spacex-rolls-homegrown-cfd/#comment-530 Fri, 03 Apr 2015 10:28:03 +0000 http://www.nextplatform.com/?p=491#comment-530 In reply to Mark.

Watch this presentation by SpaceX computer guys.
https://www.youtube.com/watch?v=txk-VO1hzBY
at the 24 minute mark they show a simuation of the flow around a dragon capsule and they superimpose the grid of the simulation. You can see the number of grid points dynamically changing to map the changing pressure fronts. Areas that have little change remain at the lower resolution, Areas of interest are mapped at higher and higher detail.
Watch the whole video. It’s very instructive.

]]> By: Alec Melvin https://www.nextplatform.com/2015/03/27/rockets-shake-and-rattle-so-spacex-rolls-homegrown-cfd/#comment-392 Tue, 31 Mar 2015 06:29:46 +0000 http://www.nextplatform.com/?p=491#comment-392 If the reactions are as fast as this, the problem is to solve sets of hyperbolic pde’s with equilibrium chemistry. This puts the turbulence model you choose as the prime factor. It must be DNS they are using, but like all turbulence models there is no guarantee that it will give realistic results.

]]> By: Terry https://www.nextplatform.com/2015/03/27/rockets-shake-and-rattle-so-spacex-rolls-homegrown-cfd/#comment-387 Tue, 31 Mar 2015 04:26:49 +0000 http://www.nextplatform.com/?p=491#comment-387 In reply to Ron.

And Musk is on the board of Solar City, which is building the first of several large advanced solar cell factories in Buffalo NY. Also, yesterday SpaceX bought $90m in Solar City bonds.

]]> By: Alec https://www.nextplatform.com/2015/03/27/rockets-shake-and-rattle-so-spacex-rolls-homegrown-cfd/#comment-386 Tue, 31 Mar 2015 04:23:11 +0000 http://www.nextplatform.com/?p=491#comment-386 In reply to Mark.

I mean, how do you resolve the kolmogorov scales without using DNS? And, if you’re resolving those scales, you certainly can’t be modeling those scales…

Good question.

]]> By: Mark https://www.nextplatform.com/2015/03/27/rockets-shake-and-rattle-so-spacex-rolls-homegrown-cfd/#comment-363 Mon, 30 Mar 2015 19:09:01 +0000 http://www.nextplatform.com/?p=491#comment-363 It appears that they’re using an advanced turbulence model such as ILES, but there’s no mention of this. All we have to go on is that they’re resolving Kolmogorov scale, surely they’re not using DNS?

]]>