How to Choose the Best Computational Physics Project as an Undergrad?

[Sojourner01]

It's time I started thinking about my final year project (it's next year, and I want to make a really good go of it)

The three main streams my department normally offer as project themes are:
i) Experimental
ii) Computational
iii) Scholarly Research

Theoretical physics topics tend to take the form of (iii) - hardly anybody does them as it's naturally extraordinarily difficult to come up with something original as an undergrad - you'd have to be an absolute genius.

Once you've picked this, you outline a few broad subject areas and you're then matched to a faculty member who has relevant research interests in order to give you something relevant and do-able with adequate support.

I'm leaning toward a computational project - I'm hopeless with labs, and research sounds dreadfully dull. Quite a few interesting projects seem to have come out of instrumentation and computational methods for particle physics in recent years, so I'm inclined toward that.

Does anyone have any suggestions for specific scenarios that I can base an undergrad computational project on? I know FORTRAN and am willing to learn C++ or C#.

 

[Dr Transport]

You need to find an advisor, they will then discuss topics that may interest you and then you can choose. Given some time I can come up with a project or two, but we need to know what intersests you.

 

[^_^physicist]

From the sounds of it you have already picked out the your choice of stream (computational).

I think the next step would be to find some material that interests you, and see how your advisor reacts to them. And then ask for suggestions (in my university if you didn't even try to come up with your own thought, it isn't very encouraging to the professors who would act as advisors).

 

[Sojourner01]

We're matched with our advisors/supervisors in the third term, just before finishing. This can be left as late as the first term of the next year, to finalise details. Persuading people to go faster than their own timetable in my department is rather tricky, so I'm looking around at other sources for ideas.

Something that seems to be fairly popular throughout the spectrum, from undergrad to postdoc, is data analysis for particle collisions. One thing my TA for Chaos theory is working on is a system that allows a logging computer to automatically rescale its sampling rate on the fly to whittle down the boring regions of a particle interaction, while ramping up the sampling to preserve detail when the interesting stuff occurs. Very clever.

 

[Sojourner01]

*thread res*

An idea occurred to me the other night - computer modelling of fluid flow around various curved surfaces. I would program a simulation of a 'fluid' of varying density and temperature passing over parabolae, hyperbolae, catenaries and so on in order to determine the drag forces and turbulent effects produced. The object would be to determine how important aerodynamic parameters vary in exotic media such as atmospheres of planets other than Earth, and so on, and whether there are any important differences in choosing aerodynamic surfaces for atmospheres other than our own.

Does this sound i) doable in terms of programming ii) doable in terms of flow dynamics and the maths involved and ii) scientifically valid?

 

[complexPHILOSOPHY]

How come you aren't able to do research in theoretical physics? Surely your advisor could help construct a minor problem for you to research?

Or does that never happen? I am ignorant towards the entire process.

 

[Sojourner01]

Well, it's not that you *couldn't* do a theoretical project, it just wouldn't be original and therefore I didn't categorise it as a research project in my own mind. It might well be new to me, but a purely mathematical project, someone will have done exactly that before me.

To be fair, flow analysis isn't groundbreaking, but such a project might still have a hint of originality in that nobody's ever presented that data in this exact way before - nobody has a program that looks like mine.

 

[J77]

I think it best if you talk to the people who could supervise you and get them to suggest possible projects.