Experimental Fluid Mechanics Vs. Computational Fluid Mechanics

[AJSayad]

Hi everyone,

I just started graduate school and I've been working in an experimental fluid mechanics lab. We have a shock tube and run tests on studying supersonic and hypersonic airflow under different conditions. We use Schileren imaging and transducers to take measurements during tests. We then use a MATLAB code to process the data and generate graphs and then interpret the data.

I was wondering if anyone has any experience working in a lab that does more of the computational and numerical analysis side of fluid mechanics (i.e. CFD, algorithims/codes to process data, etc.) and is willing to share their experience to help me compare and contrast the two lab settings and workflows. I've been thinking about switching to a computational lab since I really enjoy the mathematical/numerical/data processing side of the field as well as having the independence of working at my own pace rather than relying on the entire lab to do tests but before switching labs and changing advisors, I want to get a better idea of what the change might entail.

Any insight is greatly appreciated and is a huge help.
Thanks.

 

[berkeman]

Paging @boneh3ad

 

[boneh3ad]

Why are we paging me? I don't know jack about life as a CFD jockey. I'm a wind tunnel guy!

I think the best way to sum up the differences are as follows:
Experimentalists have wind tunnels (or similar facilities) that "solve" the Navier-Stokes equations in real time over the full scale of a model and down to the resolution of a Planck length. Unfortunately, it is very difficult to control all extraneous variables and difficult to extract information from that "solver."

Computationalists have to run codes that take various liberties with the physics and it generally takes a considerable amount of time to solve a flow field corresponding to a few seconds (or even fractions of a second) in real time. However, they can extract essentially any flow variable at ease and have almost perfect control over all variables going into their simulations.

Ultimately, the two need to work together to get anything done.

Sorry, I can't tell you what life is like running simulations for a living. Both types of people above use computers extensively but CFD folks definitely use them more. However, you can get a ton of data processing time that is heavily based on some pretty hairy mathematics as an experimentalist if you really want to dig into the data. Actually, because data are so hard to extract from a wind tunnel and they are often riddled with noise, I would argue that the data/signals processing involved with experiments is usually far more difficult than what CFD folks do. I'd also argue that the best experimentalists are well-versed in theory so that they can interpret their results more readily.

 

[AJSayad]

Hey it's all good thanks for your input! Any insight into at all is further along than where I am! I appreciate the help!