CFD Thesis Topic
 

Hello,

I am currently pursuing a PhD degree in CFD. I would like to know what possible topics there are for writing PhD thesis.

Thanks,

~Tiger

Re: CFD Thesis Topic
 

I would imagine that there are many avenues to follow. As a PhD candidate, you will need to take the lead in identifying suitable topics & motivate Thesis Supervisors to join your program.

Best of luck - may you succeed in your quest... :)

diaw...

---------------------- Tiger wrote:

Hello,

I am currently pursuing a PhD degree in CFD. I would like to know what possible topics there are for writing PhD thesis.

Thanks,

~Tiger

Re: CFD Thesis Topic
 

Thanks for your response,

Maybe I wasn't specific enough. Do you know any specific CFD problems that exist? Any websites that discuss this?

Thanks, Tiger

Re: CFD Thesis Topic
 

Tiger, I would say that CFD-Online is the best forum at the moment with a confluence of experts.

In CFD, many problems do definitely exist - at least in my mind. Some spring to mind:

1. Why do solutions blow up?

2. Which 'discretisation method' is 'best' under 'what' circumstances - better as in 'accuracy', 'speed' etc.

3. What about low speed, compressible flow? How to deal with 'Second Viscosity'?

4. Why do we use the 'wrong' energy equation for incompressible flow - ie. derive energy equation as compressible & only once complete, apply incompressible assumtion. Seems we should have a 2 in the coefficient for the convection term - not 1.

5. Do real 'singularities' exist, or are they only a figment of N-S?

6. How to compute & understand turbulence - without using a turbulence model.

7. How to apply the 'Sampling Theorem' (frequency->temporal) to spatial framework. Is this where the critical cell Reynold's number actually originates? Methinks...

8. Are current 'discretisation schemes' actually discretising, or are they merely transforming the real space into a calculation fet?

9. If #7 correct, then where does the actual discretisation step take place?

The list can go on... :)

diaw...

Re: CFD Thesis Topic (errata)
 

8. Are current 'discretisation schemes' actually discretising, or are they merely transforming the real space into a calculation fet?

9. If #8 is correct, then where does the actual discretisation step take place?

diaw...

CFD Thesis Topic (errata #2)
 

8. Are current 'discretisation schemes' actually discretising, or are they merely transforming (squeezing) the real space into a redistributed calculation *net*?

(Sorry, finger trouble) :(

Re: CFD Thesis Topic (errata #2)
 

or, like me, you could combine CFD with a rsearch topic. I am currently pursuing my phd and hope to finish this year. My dissertation involves the thermal modeling of a matrix heat exchanger. I have used an actual laboratory experiment, Fluent, and have written a simple 2-D steady CFD program - all of which, hopefully, will have similar outputs.

Re: CFD Thesis Topic (errata #2)
 

One way is to look at a variety of numerical and theroretical issues, as diaw laid out. on the other hand, expanding on Andrew's suggestion, you could remember that the ultimate goal of CFD is to predict and understand fluid flow. Maybe you can get some ideas on possible applications of CFD from this (old but still good) link which I copied from a previous post by "M":

http://www7.nationalacademies.org/us..._Dynamics.html Try to follow current trends in fluid dynamics, and you will see what kinds of problems you will have to address with CFD presently and in the future. Look at relatively new fields such as microfluid flow (is not really that new, but you get the point). Especially take a look at interdisciplinary problems, like fluid-structure interaction, flow control, reacting flows... and combinations. There are a lot of numerical and practical issues to deal with, with probably an even wider room for improvement than in classical CFD, and it will widen your horizon to deal with several disciplines (and make you more marketable). You probably don't want to choose something that has been chewed on successfully for quite a while, making it very hard to come up with a significant contribution. But that doesn't mean you cannot take on something old. Turbulence is a prime example for an old yet unsolved problem.

In essence, what I am trying to suggest is: look at the directions in fluid dynamics and interdisciplinary research and you'll see where to go with CFD.