cooling panel design

[thunderspeed]

Hi guys ,
I wonder if it would be that hard to write my own code according to the numerical and mathematical modeling equations of heat and mass transfer , where I know each piece of code detail.
Or is it too hard and time consuming and I'd better just go for a commercial package like openfoam or ansys , where I wont know how each detail is calculated and yet it would need validation for my model .
I'd be more clear about what I want exactly :
1- My goal is making a reliable mathematical describing behavior of panel where input and output relation are more precise (there isn't an exact model for that
up till now ) . I'm not sure yet what that would be like but i think it would
be something like a state space mathematical model .
2-From the model I construct , I want to come up with a way to increase cooling capacity (performance) of the panel , maybe would be through changing some dimensions or angles of the panel or proposing some changes in design and reconstructing a new model .
3-Verify results that theoretical improvements in performance really apply experimentally.
4-Understand maximum amount of details of this design process to be able to defend my thesis.

Please anyone having experience with the above stated points share your opinions , Thanks in advance .

 

[EnergyMix]

I have some nominal experience and my reply is ... it depends.

How good a programmer are? (I'm not, it takes me a long time, with a lot of trial, but I occasionally used to like to do it.)
How much time do you have? (I have very little free time any more. I used to have some and I'd play around with things like this.)
How thoroughly have you explored the free-ware sites already available (ok, for what I was doing, those were "good enough." Plus I got my programs to match fairly well, so I didn't need them to go any further.)
How much money do you or your company have? (If you have access to a computer program, then it can save your hair from coming out in chunks.)

Now maybe for you, programming is much easier, you have much more time, and much less money and you want to do it yourself ... go to it.
Or maybe you have only a little free time and need to brush up on programming skills first -- I recommend looking into the available freeware.
You, like the rest of us, only have 24 hours in a day. How do you want to spend it?


Want to know the do's and don'ts of Eng-Tips? Read FAQ731-376.
English not your native language? Looking for some help in getting your question across to others or understanding their answers? Go to forum1529.

 

[IRstuff]

You can, but that just seems like a busywork kind of thing. IF you have $X of budget, do you really want to spend $Y on writing code that's essentially the same as the code that already exists in some FEA program, result in $X-$Y left for doing the actual engineering, or do you spend it all on the actual problem you are trying to solve?

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

Of course I can. I can do anything. I can do absolutely anything. I'm an expert!

 

[Tunalover]

Sounds like an unnecessary programming exercise. If money is a concern then there are freeware (or cheapware) nodal analysis, boundary element, finite element, and finite difference codes out there. If you're uncomfortable putting blind trust in a program, work a few simple problems having handbook solutions and compare the solutions. Another way to build confidence in the tool is to run tests and compare the results with the test results. If you can afford ANSYS then there are things to learn by reading the theory in the manual.


Tunalover

 

[thunderspeed]

Thank you guys for your precious point of views
By the way I'm not working in a company its my Masters Thesis topic ,,, I have about six months of time , zero budget and the most important thing for me is to be as powerful in the subject as I could to be able to defend my thesis in front of the testing committee
I've spent time since posting this question yesterday skimming thorough heat and mass transfer and CFD books and I found that its a whole new world , Allot of Differential equations . It's really nothing but pure differential equations math . Yet its deadly essential for my mechanical engineering thesis .
So form your answers I conclude that its better to use the available software like openfoam for example . Rather than reinventing the wheel and giving more time on other engineering work in the project . The problem here is validation . I can draw my parts on cad . pass flow and run simulations . But you never know you're on the right path with your results .I thought of sending new design and results to one of companies that construct this component for them to validate the outcomes . Is that a possible option ?
I'll go for openfoam and couple it with my python programs of the system and keep you updated guys .
thanks again all for your advice

 

[Tunalover]

thunderspeed,
As a student you can get robust commercial software for very cheap if not free. Check with your local reseller of ANSYS, FLUENT, CFDesign, ICEPAK, MSC/NASTRAN, COSMOS,...Also, check with your adviser. As a faculty member who has advised other grad students he should know how to get the free (or cheap) software. You may even be able to use an educational version of a code on-line from your school. Check with your adviser. The CFD codes (CFDesign, ICEPAK, FLUENT) are advantageous as you do not need to know convection coefficients to solve problems. Good luck.



Tunalover

 

[thunderspeed]

Tunalover
I already have ansys , I was curious weather it's a good idea to construct and work with my own cfd code . Of coarse working with your own code will be superb since you know each and every minute detail . But from what people say here I'd rather use fluent or cfx right away and spend this extra time and effort on something else .

 

[IRstuff]

That gets into the details of your thesis; how much of it depends on the modeling algorithm. If it's the design approach, then the modeling is only incidental to the concept. Would there be potential issues arising from your choice of modeling software, and its accuracy at modeling odd structures, or whatever?

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

Of course I can. I can do anything. I can do absolutely anything. I'm an expert!

 

[Tunalover]

thunderspeed-
Unless you have a highly reliable method of determining convection coefficients (to use as an ANSYS FEA boundary condition) you're best bet is to go with a CFD code. Most CFD codes work on the finite difference method but some (CFDesign) use the FE method. Finite difference codes are much simpler to write than FEA codes so if you choose to write your own code that's the way to go.




Tunalover