# 2D transient analysis for heat conduction

I need a fem code either in matlab or fortran to refer to as I am stuck at some places. Please help.

## Answers and Replies

minger
Science Advisor
Simply giving you a code for your homework (probably a final project even) would be quite unethical...more importantly though it wouldn't be fair to those of us who did it by ourselves.

Perhaps you can shed some light on where you're stuck. No one is going to do your work fro you.

I love these threads that start out with a completely open ended question, without any indication as to what the problem actually involves. Then they ask for all the computer code and show no work at all.

The fact that gets me, is that these posters actually think this forum is intended for this type of help.

LOL

Matt

minger
Science Advisor
I wish I had like a 2000 line code with several modules, dozens of subroutines and such that in the end just:
Code:
Hello World!

Hell, I'd zip it up and make sure it was always hosted.

Hey, enough of criticism. Its not that if you give a code, it will be terrible. I know that I need to do my own code, which I am doing. Even if you give me a code, first I need to understand your code to implement it, which for sure wont be easy. Everybody has his/her own way of coding. Neways , my bad. I will explain where I am stuck.
When you solve a transient problem of the format.

If you have M du/dt+K u=F, we can break it into Mv(1)+K(u1)-F(1)=0 at t=t1 and
Mv(0)+K(u0)-F(0)=0. Using the trapezoidal rule, it asks to guess the V vector at t1 even though its a linear problem. I am very confused with the approach.

Why did you ask for a FEM (finite element code) when you are told to use the trapezoidal rule?

The trapezoidal rule is nothing more than a numerical integration algorithm. Albeit a very robust one.

Hey, enough of criticism.

If your going to post a silly question then get ready for the criticism. When you post questions here act as if you are asking your supervisor the question. Then ask yourself the following questions.

1. Have I specified enough information so that someone can actually help me?
2. Is there any additional information that I can post that will help the person(s) trying to help me?
3. Is the spelling/grammer correct to the best of my knowledge?
4. Have I searched the forum for a similar question?
5. Is there a graph/picture that I can attach to help with the visualization of the problem?

Remember, we are not in the class with you, so make it easy for us to help you.

Thanks
Matt

Last edited:
Thanks. It is a numerical technique to match in time.right? Lets say I have formed my mass matrix and my element matrix. If I know the initial boundary conditions, lets say primary as well as secondary boundary conditions. When we solve MV(0)+KU(0)-F(0)=0 for V(0), where V(0) is the velocity at time t=0 and U(0) being the displacement. Do we impose the secondary boundary conditions like stresses in our F(0) vector and also the given displacements in our U vector. I am getting confused as if the primary value i.e U is known at certain nodes, V should be zero. About the secondary boundary conditions(du/dx or du/dy)...do I impose them too? If I am calculating V for time t=0 using that equation, my question is if I am calculating V for all nodes? or the interior nodes? or the nodes where only the primary value is given?

Ugh.

Ugh.
@Brian, y was your response like this? Was my question so annoying?

Nobody has any idea what you're talking about. The title of the thread mentions heat transfer, but it looks like you're solving some kind of structural dynamics problem.

Nobody has any idea what you're talking about. The title of the thread mentions heat transfer, but it looks like you're solving some kind of structural dynamics problem.

Let me rephrase the question in terms of temperature.
When we solve MV(0)+KU(0)-F(0)=0 for V(0), where V(0) is the dT/dt at time t=0 and U(0) being the temperature. Do we impose the secondary boundary conditions like flux in our F(0) vector and also the given temperatures in our U vector. I am getting confused as if the primary value i.e U is known at certain nodes, V should be zero. About the secondary boundary conditions(du/dx or du/dy)...do I impose them too? If I am calculating V for time t=0 using that equation, my question is if I am calculating V for all nodes? or the interior nodes? or the nodes where only the primary value is given?
Hopefully, this might be clearer.

Also if on two boundaries ,adjacent ones, fluxes are prescribed on both of them, which one takes precedence? for example , one boundary has flux given 10 and another 100. The node common to both these boundaries will take what value of flux?