Hi all,

I was just wondering how the "CFD Community" views graduates of Applied Math vs Mech Eng vs CS etc at the graduate level, as in...what is more desirable, Ph.D. applied math, Ph.D. MechEng or Ph.D. Computer Science? Assume all had dissertations that were somehow related to CFD...

I imagine all three have their "perks": MechEng: High Fluids Knowledge Math: High Math Knowledge CS: Good at parallel computing, HPC, programming

If my goal was working at a software company (ex Fluent etc) to write simulations (or perhaps national labs..), what would be the best academic path to plan out?

Thanks, Dave

Dear friend, I feel its better to make a choice between applied mathematics or Mech Engg. These two are the natural subjects which can take you to the depth of the nature. Computers are being artificial, you cant enjoy them much.

As a CFD researcher and teacher, I find that the applied mathematicians understand the numerical analysis concepts (FEM,FVM,etc) better than the Mechanical Engineerinng students. Sometimes, the ME students lack certain math skills and the mathematicians know more about the theory and therefore have an easier time with applications. I guessed I'm biased since I do have an Applied math degree, but I also teach Mechanical engineering classes so I have firsthand knowledge. Also, it seems that Mathematicians have very interesting ways to look at problems. If you are interested in working for a company doing the same "old" thing, then go to mechanical engineering, however if you are more interested in doing research/teaching and discovery then go to applied math. Who else can have fund deriving the Navier-Stokes equations? Only mathematicians.... Good Look with your future and may the force be with you. ~Tiger

Mathematicians usually have a hard time relating math to physical problems.

In my experience mathematicians are better suited to software development and accademic type problems, this way they can spend hours poring over lines of code and situations that do not require an engineering brain.

Engineers are better suited to getting solutions to real world problems and are more adept at problem visualisation and have the skills required to inovate, adapt and think outside of the box (as the americans would say).

In maths you (generally speaking) have to stick to fairly rigid rules where as in engineering you have to be flexible in every way.

So in answer to your original question, it depends what you want to do and what you feel you are better suited to, being a good engineer is something you cant learn or be taught the best and most gifted engineers simply have a certain type of brain and a certain way of looking at things and problem solving, if you don't have this already you never will.

(wouldnt guess im an engineer would you!)

Suppose someone had equal skill and inclination in both areas (engineering and maths) but had to make a decision between them in selecting the school at which to do his PhD. On the basis of eventual market value/employability , which way should he/she go?

Personally I would say engineering is the way to go, with CFD being a growing field there is going to be more need in the future for end user engineers than developers and academics.

So as a mechanical engineering Ph.D., the jobs are more "end-user" type rather than development? What would be typical jobs for each type of Ph.D.?

Who are the people that write industrial CFD codes?

Thanks, Dave

Obviously there is no hard and fast rule as to what type of job you go into. But in my opinion the breakdown would be something along the lines of:

Engineering based : Consultants, support engineers, analysis engineers

Maths based: Development (numerics mainly but possibly general), Accademic fields

Physics based: Developent (turbulence models, multiphase etc), R&D engineering more likelyhood of crossover into the engineering fields, accademia

Although I may be wrong of course, thats just my opinion from my experience.

I work in an office where poeple "write industrial CFD codes". We are a wierd mix of mechanical and other engineers by schooling, plus a mix of physics and maths people. Generally speaking, there are two types here:

1) Those with a PhD in CFD.

2) Engineers who just happened to like writing code and "graduated" into their software jobs.

I am one of the second type. I liked writing code to help me in my job and got bored when I couldn't. So I changed it.

Your degree may land you your first job, but after that it's your skill and experience.

The only problem with this breakdown is it ignores the rather important fact that, in the UK at least, an applied mathematician will know a great deal more about fluid mechanics than a mechanical engineer while a physicist will know next to nothing - fluid mechanics is generally taught in mathematics departments and not in physics.

Its worth noting that most of the great fluid dynamicists were applied mathematicians; e.g. Lamb, Lighthill, Batchelor, Stewartson, Goldstein, Rayleigh, Kelvin, etc.

An applied mathematician may know more about the mathematical basis of fluid mechanics but less about using it to get real world engineering solutions, indeed more and more engineering degrees are incorporating CFD into the syllabus as it's importance in a modern engineering industry increases year on year.

It is true that in the early days of CFD when computer power wasn't up to solving many real world problems, mathematicians and academics held sway but now computer resource is cheap and readily available engineers are taking over and, although not running at the same levels of accuracy as the mathematicians and academics, are making CFD into a viable tool that can allow innovation in areas from biomeds to tractors.

Us engineers rely on you purists to do the hard yards and leg work getting tools such as CFD working so we can create things that are useful to the rest of the world! Would Navier and Stokes be as famous as they are if it wasnt for Orville and Wilbur?

(sorry that sounds like some sort of advert!)

Just a simple advice:

Go to the path that your skills points to. You can't be a good professional forcing what you have to learn. First of all, you must have pleasure about what you're learning/working

and good luck...

Renato.

Very good point, after all the discussion about who is good at what the best advice is do what you are good at!

"Its worth noting that most of the great fluid dynamicists were applied mathematicians; e.g. Lamb, Lighthill, Batchelor, Stewartson, Goldstein, Rayleigh, Kelvin, etc."

Oh, this I love. So if we continue this discussion and ask ourselves what gender I shuld have if I am to be a great fluid dynamicist.......

I am toast since all great ones before me were males. Back to kitchen school again :P

Think outside the box?

What people on this website tend to forget is that CFD is a tool to solve problems in Fluid Mechanics. What you have to decide is do you want to be a "toolmaker", or do you want to be the "mechanic"?

If you want to be the toolmaker, I would say that getting a degree in computer science, but taking a few courses in Fluid Mechanics would be the best way to go. You will then be great at writing code, and have a rudimentary knowledge of fluid mechanics which is all a code developer at a comercial CFD company really requires, you can then pick up more fluid mechanics once you get there ... the basics are not so hard. The core solver of a CFD code is a small part of the whole package nowadays, it is becoming increasingly more important to have good pre and post processors and coding these has little to do with fluid mechanics. The modelling aspects are still very important, but most models are not developed by the CFD companies but come from academia. You can code these models without understanding all of the details.

If you want to be the "mechanic", you need to fully understand fluid mechanics and all of the nuances of this field. You must have a good feel for the PHYSICS of what a fluid will do, you must think like a fluid The best way to get this feeling is to do some experimental and computational fluid mechanics and the best way to get this experience is to do a PhD. in Mechanical or Chemical Engineering.

I hope this helps, I bet there are people out there that disagree with my assessment. I have a lot of experience as both the toolmaker and the mechanic and did my PhD. with a combination of experiments and CFD. In my job, I also use both methods to solve fluid dynamic problems.

If you want to make a lot of money in your career, steer well clear of this field and get into financial modelling!

Think outside the box?

I have appologised for that once and I do so again, I am not american!

Great answer!!!

Also, a PhD is not an expert in a particular area of study but someone who has spent considerable time in that area, with the skill to work independently, focused, who loves being challenged and does not give up easily..

If you can finish a Ph.D., you should be able to tackle any problem in related areas to your undergraduate background. Some areas may require more initial work based on accumulated experience, that's all.

Good luck with whichever choice you make, any choice will be fine as long as you like it.

Opaque

Forget Engineering and do the Math. Then go to Wall Street and make some real money. Seriously.

If I had the opportunity to do it again, this is the route I would choose.

Father Jack PhD Engineering.

During the time of Kelvin at least, engineering as a field of study was much different than it is now. Kelvin's brother was actually an engineer.

Also the question this guy asks isn't really reasonable anyway..unless he is applying to undergrad and is thinking long term his undergraduate degree will more than likely determine what dept he gets his degree in..it won't determine his research since who he chooses for an advisor will more than likely determine that.