Validation Data for Hyperloop/Evacutated Tube/Train in Tunnel

spggodd · January 23, 2016, 07:47

Hi,

I'm currently working on a CFD study of the full scale Hyperloop (not the design competition that is currently happening) and I am finding it difficult to obtain any trusted data for validation purposes, probably due to lack of experiments etc..

But does anyone know of any studies that have been published that I could model and compared which relate to either the Hyperloop, Evacuated Tube Transport or even a train in a long tunnel.

Data source such as the following would be ideal:

ERCOFTAC
NASA
Flownet
Or reputable journals.

Many Thanks
Steve

Update: This is the closest case I have found, can I get an experienced opinion if this is good enough to act as a base from which I will modify to suit my own simulations?
http://link.springer.com/article/10.1007%2FBF03325776

Springroove · October 31, 2017, 08:43

Hi,

I was curious about your simulation as I am working on the same problem in trying to perform a CFD analysis for the Hyperloop. Can I ask a few questions if you're still on it?

Cheers

spggodd · October 31, 2017, 12:34

Hi,

The best data I could find at the time was a simplified train geometry in a tunnel by Zhang (2013). Is be happy to help by sharing my other sources if you like but I’m currently out and about and on my mobile.

I also carried out my own study specifically on the Hyperloop in both 2D, 3D steady and 3D transient. Coincidently I’ve just submitted a cut down version of my investigation report to the International Journal of Theoretical and Applied Multiscale Mechanics (IJTAMM) for review. I’ll let you know when it’s available!

In the meantime I can try and help with your questions.

Cheers
Steve

Springroove · October 31, 2017, 14:22

Thanks for the reply and congrats on the publication. I was also looking at Zhang and besides that, Ma (2013) and Kim (2011), but I guess you've figured those out as well.

I am currently performing simulations for the 2D case but can't seem to find the results that are comparable to the aforementioned papers, so I was mainly wondering on any specific conditions that you have utilized?

I'm running a (steady state) simulation is with an SST model. The vehicle is moving at 300 m/s so I let all the walls move with that speed in the flow (x) direction. I've imposed a velocity inlet with a total temperature and the outlet pressure at 100 Pa.

I was wondering if my results could be different due to the fact that I let my walls move instead of utilizing a moving mesh for instance? Have you by chance used comparable boundary conditions as well?

Many thanks, Xilef

spggodd · November 2, 2017, 07:58

I'm not familiar with Ma and Kim do you have a link to them?

Zhang only lists drag results in his paper for a number of configurations.
I took the 101 Pa case and compared over the range of velocities with 5 different incrementally finer meshes until my results matched within a reasonable accuracy.

Not to say that SST is wrong but I used a K-e model as it was more efficient computationally and seemed to be the model of choice on most of the literature I reviewed.

As I got to the higher speeds >200m/s the results became more inaccurate. Perhaps try your simulation at a slower speed (at least to start with). With my further simulations (where I added compressor and nozzle geometry) I had to limit speed to around 200m/s due to some issues with supersonic flow in small areas of the simulation which affected the computation.

The tube wall on all my simulations were moving at the same speed as the inlet flow. This should be correct as there should be minimal viscous flow on the tube walls.

--

I've queried about sharing my research but unfortunately I've been advised to hold fire until the paper is published - hopefully that won't be too long but I can still answer general queries etc..

Springroove · November 3, 2017, 09:48

Sure, I find them quite useful. Kim even compares different turbulence models.

Kim:

http://www.fem.unicamp.br/~phoenics/...be%20train.pdf

Ma (Actually Chen et. al):

https://link.springer.com/article/10.1007/BF03325788

Ah too bad; hope it's available soon too! I'm mostly struggling with the high-speed conditions as well. I don't suspect it to be mesh related as I am starting out with a quite coarse (unstructured) mesh. However, a mesh with only ~300k nodes already gives me 20+ hours computation time, so I suspect it has something to do with my boundary conditions.

Did you just impose a static (ambient) pressure at the outlet as bc as well? Cheers

January 23, 2016, 07:47	Validation Data for Hyperloop/Evacutated Tube/Train in Tunnel	#1
spggodd Member Steven Goddard Join Date: Mar 2015 Posts: 34 Rep Power: 11	Hi, I'm currently working on a CFD study of the full scale Hyperloop (not the design competition that is currently happening) and I am finding it difficult to obtain any trusted data for validation purposes, probably due to lack of experiments etc.. But does anyone know of any studies that have been published that I could model and compared which relate to either the Hyperloop, Evacuated Tube Transport or even a train in a long tunnel. Data source such as the following would be ideal: ERCOFTAC NASA Flownet Or reputable journals. Many Thanks Steve Update: This is the closest case I have found, can I get an experienced opinion if this is good enough to act as a base from which I will modify to suit my own simulations? http://link.springer.com/article/10.1007%2FBF03325776 Last edited by spggodd; January 23, 2016 at 16:09.

October 31, 2017, 12:34		#3
spggodd Member Steven Goddard Join Date: Mar 2015 Posts: 34 Rep Power: 11	Hi, The best data I could find at the time was a simplified train geometry in a tunnel by Zhang (2013). Is be happy to help by sharing my other sources if you like but I’m currently out and about and on my mobile. I also carried out my own study specifically on the Hyperloop in both 2D, 3D steady and 3D transient. Coincidently I’ve just submitted a cut down version of my investigation report to the International Journal of Theoretical and Applied Multiscale Mechanics (IJTAMM) for review. I’ll let you know when it’s available! In the meantime I can try and help with your questions. Cheers Steve Last edited by spggodd; November 2, 2017 at 07:45.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Experimental data validation and/or grid independence	svp	Main CFD Forum	5	June 6, 2014 04:24
CFX-Pre problem, pls help!!!	cth_yao	CFX	0	February 17, 2012 01:52
Validation Study: How do I plot wind tunnel data (*.dat) in Paraview	bigbang	OpenFOAM	4	August 5, 2011 09:00
CFD Validation Data	Alan Stewart	Main CFD Forum	1	November 9, 1998 07:53
Validation data for winglets	Bo Jensen	Main CFD Forum	1	October 28, 1998 11:44

October 31, 2017, 08:43		#2
Springroove New Member Xilef Smith Join Date: Apr 2017 Posts: 19 Rep Power: 9	Hi, I was curious about your simulation as I am working on the same problem in trying to perform a CFD analysis for the Hyperloop. Can I ask a few questions if you're still on it? Cheers

October 31, 2017, 14:22		#4
Springroove New Member Xilef Smith Join Date: Apr 2017 Posts: 19 Rep Power: 9	Thanks for the reply and congrats on the publication. I was also looking at Zhang and besides that, Ma (2013) and Kim (2011), but I guess you've figured those out as well. I am currently performing simulations for the 2D case but can't seem to find the results that are comparable to the aforementioned papers, so I was mainly wondering on any specific conditions that you have utilized? I'm running a (steady state) simulation is with an SST model. The vehicle is moving at 300 m/s so I let all the walls move with that speed in the flow (x) direction. I've imposed a velocity inlet with a total temperature and the outlet pressure at 100 Pa. I was wondering if my results could be different due to the fact that I let my walls move instead of utilizing a moving mesh for instance? Have you by chance used comparable boundary conditions as well? Many thanks, Xilef

November 2, 2017, 07:58		#5
spggodd Member Steven Goddard Join Date: Mar 2015 Posts: 34 Rep Power: 11	I'm not familiar with Ma and Kim do you have a link to them? Zhang only lists drag results in his paper for a number of configurations. I took the 101 Pa case and compared over the range of velocities with 5 different incrementally finer meshes until my results matched within a reasonable accuracy. Not to say that SST is wrong but I used a K-e model as it was more efficient computationally and seemed to be the model of choice on most of the literature I reviewed. As I got to the higher speeds >200m/s the results became more inaccurate. Perhaps try your simulation at a slower speed (at least to start with). With my further simulations (where I added compressor and nozzle geometry) I had to limit speed to around 200m/s due to some issues with supersonic flow in small areas of the simulation which affected the computation. The tube wall on all my simulations were moving at the same speed as the inlet flow. This should be correct as there should be minimal viscous flow on the tube walls. -- I've queried about sharing my research but unfortunately I've been advised to hold fire until the paper is published - hopefully that won't be too long but I can still answer general queries etc..

November 3, 2017, 09:48		#6
Springroove New Member Xilef Smith Join Date: Apr 2017 Posts: 19 Rep Power: 9	Sure, I find them quite useful. Kim even compares different turbulence models. Kim: http://www.fem.unicamp.br/~phoenics/...be%20train.pdf Ma (Actually Chen et. al): https://link.springer.com/article/10.1007/BF03325788 Ah too bad; hope it's available soon too! I'm mostly struggling with the high-speed conditions as well. I don't suspect it to be mesh related as I am starting out with a quite coarse (unstructured) mesh. However, a mesh with only ~300k nodes already gives me 20+ hours computation time, so I suspect it has something to do with my boundary conditions. Did you just impose a static (ambient) pressure at the outlet as bc as well? Cheers