@Mariusz Iwaniuk 

Thanks a lot for your answers. Obviously I know how to solve the equation by hand, but I wanted to do it automatically. The solution by @Mariusz Iwaniuk  is not what I was looking for, but it could work out.

Thanks again,

Javier

@Thomas Richard  

Thanks for your answer however "pdsolve, boundaryconditions" only solves single partial differential equations. 

I have not been able to find any example in Maple2017 neither :(

@tomleslie Thanks for your answer but this is not a system with boundary conditions.

Actually I have tried to add some random boundary conditions to your example and Maple does not solve it :(

@Thomas Richard 

Dear Thomas:

Thank you very much for your answer. Unfortunately all the examples there contain only a single partial differential equation, there isn't any example with a system of partial differential equations.

Please, let me know if I'm wrong and I'm checking a different help page. I'm looking at http://www.maplesoft.com/support/help/Maple/view.aspx?path=examples/pdsolve_boundaryconditions

Javier

@vv  Thanks for your help, but that's not what I want.
I think it was clear before. Since t=t1/km + t2/km^2 whenever I have diff(f(t),t) I would like to have

I understand maybe I shouldn't use the symbol f for both the one and the two variables functions but that is what I need for my code. Is there any way to achieve it?

Thanks again.

I'm afraid I wasn't clear enough. I would like to find a command to apply the chain rule following

when  t=t1/km + t2/km^2.

OK, I see,  dchange only works if you have the same number of transformation equations as the number of new variables. It's a pity because it would be useful for cases like this one.
Any other idea how could I apply the chain rule without redefining the variable t?

Thanks again for your interest.

@vv 

I would expect dchange applies the chain rule to get:

 Thanks for your answer.

Javier

@vv 

Thank you very much for your answer.

I think the second one is an intereseting way to obtain it.

Maybe we  could improve the precision extending the derivative of the temperature to more points inside the domain. I guess the maximum precision is controlled by the spatial step employed in pdsolve. Am I right?

Thanks again

@Kitonum 

Thanks to everybody for your answers.

The curve fitting to a list of random numbers using splines could be a useful way if we were able to get an analytical continous expression.

I think pdesolve only accept a continous functions for the initial condition and does not accept a list of numbers.

Thats again for all your help.

Javier

@Kitonum 

Thank you very much for your answers. Unfortunately it is not a good solution.
If you plot that function from x=0 up to x=50 you can see it is far to be a random collection of numbers :(

Thanks for your answers. The expression is more involved. I did not want to writte it here, but subs is not working, it has to be algsubs. 

I solved it using a for loop, but I guessed it should be a more elegant solution,

Hi,

Thanks but Im not using the ODE Analyzer Assistant but the "odeplot" command in a worksheet. Any idea how to do it in this way?

Thanks,

Javier

Hi,

Thanks but Im not using the ODE Analyzer Assistant but the "odeplot" command in a worksheet. Any idea how to do it in this way?

Thanks,

Javier

Hi,

Thanks for your answer. Unfortunately I have Maple 13 so this is not working. Is there any way to do it for older Maple versions?

Thanks,

Javier