Why InternalHeatSource function doesn't accept variable internal heat expressions in terms of state.u?

Greetings

I am trying to create a internal heat source that is a function of Temperature for a steady state simulation. However it seems that InternalHeatSource function doesn't accept functions that declares the internal heat source value that are function of state.u, but when I write, in the same expression, location.x in place of state.u, the code does works. Does anyone know the reason why?

[Ravi Kumar]

Hi Diego,

Can you post a simplified reproduction code and which version of MATLAB are you using.

Regards,

Ravi

[Diego]

Hi Ravi,

Thank you so much for your help! I have been struggling with this problem for days.

I'm sending the simplified version of my code. If you need a shorter version please let me know and I'll cut more the code.

I am having a problem to write the variable internal heat source function, the q function on the code.

My Matlab version is the Student Version 2019a.

Again, thanks for your help!

Diego

[Ravi Kumar]

Hi Diego,

I am getting a syntax error, which I tried to fix based on the expression you have written in the file. But I could not get nonlinear sover to converge even after my attempt to fix the expression. Can you try your code in newer version of MATLAB (R2020a) if you have access.

Regards,

Ravi

[Diego]

Hi Ravi,

Thank you for taking some time to try to help me solving that problem. I realy appreciate!

I don't have access to the R2020 through my University but if this upgrade version is going to solve my problem I can purchase it. But, anyways, I'll tring to solve this problem using the c, f, d, m, a matrix and solve the problem with the general solver. However, please, if possible, let me know if the newer version of Matlab will be able to deal with my previous code.

Thaks againg

Diego

[Ravi Kumar]

Hi Diago,

Take a look at the attached version, where you will get the same nonlinear sovler initial condition issue. The change I made is to specify one of the BCs as a function handle to trigget the nonlinear solver and fix a syntax error in the function 'q' . You should be able to identify issue with IC based on your knowlege of the problem.

I would suggest staying with R2019a and thermal workflow and see if the porblem is feasible before you move to newer version.

Regards,

Ravi

[Diego]

Hi Ravi,

Thanks for solving the syntax error in the q function! I have converted the simulation to a transiente model and applied the IC, but it seems that the system is not "seeing" the internal heat generation. The midle section, where the curing process occurs still don't increases its temperature due heat released by the resing curing modeled by q. The convecntion and conduction seems to be working, the temperature profiles close to the constant temperature boundaries looks good. It seems the only problems is the internal heat generation. Would you have a idea what is it possible to fix it? I'm attaching the Code with what I did.

Thank you so much for spending your time helping me with this problem!

Best Regards

Diego

[Ravi Kumar]

Hi Diego,

If you insert a breakpoint in the q function and set a condition (numel(state.u)>1), then you will observe that solver does call the q function with various values of solution. So it does see new values of q at different solution values. I think it might be just plotting issue. You are plotting the initial step if you don't specify right column ID in the T matrix (note T is a matrix now). If I change the plot command to:

pdeplot(model,'XYData',T(:,end),'Contour','on','ColorMap','hot');

then I get the following plot:

Hope this helps resolve your issue.

Regards,

Ravi

[Diego]

Thank you so much Ravi,for spending your time to help me! It completly solves the issue!

The routine gives a maching but whith a better resolution than the finite difference version that I made previously. I'm realy happy with the results since this code will allow me to analize the problem in a 3D version.

Again, thanks for helpping me solving this problem!

Best Regards

Diego