Discussion Forum

Hello,

in general I want to solve following kind of time dependent differential equation in my model of an electrical machine:

i(t)= a(t) + b(t) + d/(dt)*i(t) +d/(dt)*c(t)

in the ODE table this equation would like:

i-a-b-it-ct=0

Comsol does now think, that "ct" is a new Variable. So I tried to do the derivation "d/(dt)*c(t)" in the normal definitions.

I picked a more simple example and disabled my whole model, just to test if the ODE works.
Follwing was defined in the "mod1-Definitions":

u = t
b = 2*d(u,t)

In the ODE Physics I defined:

i-u-b=0

the result is a graph:

i=2+t

which is right ofc.
When I change the time- operator in b to "TIME" like:

b=2*d(u,TIME),

the result is wrong because the derivation is zero. So I thought I have to use the "t"- operator in my model, if I want to couple equations of the "mod1-definitions" with the ODE- Physics.

But when I use my complete model with the electrical machines physics, I get the following error message:

"Time derivative only allowed with respect to mesh time.
- Expression: d(mod1.Az,t)"

So I tried to change "t" to "TIME" and check if the error still occurs. But it doesn't. The problem is now, that the "TIME"- operator didn't work properly in the easy example. So it wouldn't work in the big model either...


Is ODE maybe the wrong physic for this? Should I maybe define something in the "global Definitions"?

I'd be glad if you can help me.

Thanks in advance.