Computational Skills of a Nuclear Engineer

[Vnt666Skr]

This Fall I will be joining my university for MS Nuclear Engineering. My undergraduate major is Mechanical Engineering. I have a few questions.

1. How important is computational physics in Nuclear Engineering?
2. How to go about learning and practicing it in the context of nuclear engineering applications?

I have some experience in basic programming using Python 2.7 and would like to take it to the next level.

3. How does python fare in comparison to C,FORTRAN etc in scientific computing?Thank you

 

[kinkmode]

I can answer 3 to a certain extant. For quick calculations, maybe, but for any real scientific computing, you should probably be using Fortran or C. That being said, python is quite nice for data analysis, particularly with scipy and numpy. I used python in that manner for about 10 years, producing publication quality plots quite easily.

 

[Vnt666Skr]

Thank you very much kinkmode ! Waiting for more responses.

 

[Smed]

computational skills are not necessary in nuclear engineering, but they will certainly place you in a valuable niche if it's something you enjoy. Python is an excellent language that can be integrated with much of the software used in the industry, especially for data analysis as mentioned above. With that said, much of the industry software is written in Fortran. This is mostly because much of the industry software is very old and resistant to change. In my personal opinion, Fortran does not have much of a future, but perhaps I'll be proven wrong.

 

[x86]

I do not have much experience in nuclear engineering, however, I am a novice programmer and I would like to state that python is a great language for computations and it has fast developing time and many libraries. I wouldn't bother with fortran, because it is dying out.

I can't think of any scenario where you would want to use C over Python for calculations, with the exception of boolean algebra or if you have to directly access some memory addresses or do some OS internals. And even then, Python can even do boolean algebra

 

[Thermalne]

This Fall I will be joining my university for MS Nuclear Engineering. My undergraduate major is Mechanical Engineering. I have a few questions.

1. How important is computational physics in Nuclear Engineering?
2. How to go about learning and practicing it in the context of nuclear engineering applications?

I have some experience in basic programming using Python 2.7 and would like to take it to the next level.

3. How does python fare in comparison to C,FORTRAN etc in scientific computing?Thank you

1. It depends on what type of applications of nuclear engineering you end up working on. There is a lot of subfields in nuclear engineering and the starter question is what interests you the most?

2. Introduction to Nuclear Engineering by Lamash is a great starter texts for learning the basics of nuclear engineering. https://www.amazon.com/dp/0201824981/?tag=pfamazon01-20. That will give you some insight into where computational physics is utilized. (See Neutron Transport Equation and diffusion equation)

3. I've never seen python used outside of simple data analysis. In my experience, Fortran is almost exclusively used for any major programming efforts in nuclear engineering. It's not great, but that's what a lot of people still use in the industry. I have seen a lot of noise about using C or C++ in nuclear detection applications though.

Good luck!

 

[Vnt666Skr]

Thanks for your insight Smed,x86 and Thermalne.
I enjoyed programming in Python. FORTRAN and C although more popular, are not as easy to use and learn as Python. Yes, Python tends to be slower than the other two. But now that I am learning and liking it, I don't think it's wise to leave it midway and pursue something else. Better to be good at one thing than dabble in two, isn't it?

 

[Astronuc]

Most scientific programming is done in Fortran or C/C++, with some support (e.g., script files) doen in python.

A lot of legacy code (core simulators, CFD/TH, reactor/plant analysis, . . . ) is written in Fortran. Modern methods (INL's MOOSE system and herd) are more likely to be written in C++, especially for large scale analytical software requiring massive parallelization.