PDEs: Diffusion Equation Change of Variables

1 reply · 3K views · last reply
Master1022
Messages
590
Reaction score
116
TL;DR
Question about change of variables for the Diffusion PDE
Hi,

I understand the underlying concept of changing variables in PDEs (so that we can reduce it to a simpler form), however, I am just not completely clear on the mathematics of it so I have a quick question about it.

For example, if we have the transmission line equation [tex]\frac{\partial V}{\partial t} = \frac{L^2}{RC} \frac{\partial ^2 V}{\partial x^2}[/tex]
with initial condition: [itex]V(0,x) = 0[/itex] and boundary conditions: [itex]V(t,0) = V_0[/itex] and [itex]V(t,L) = 0[/itex]. Now we want to change variables to reduce the problem into a simpler form without any free parameters.

The Method:
1) Rescale the length variable so that it ranges over one

I let [itex]\xi = \frac{x}{L}[/itex]

2) Rescale the other independent variable to remove free parameters from the general expression
[tex]\frac{\partial V}{\partial t} = \frac{L^2}{RC} \frac{\partial ^2 V}{\partial x^2}[/tex] [tex]\frac{\partial V}{\partial t} = \frac{L^2}{RC} \frac{\partial ^2 V}{\partial \left( L \xi \right)^2}[/tex] [tex]RC \frac{\partial V}{\partial t} = \frac{\partial ^2 V}{\partial \xi^2}[/tex]
then I let [itex]\tau = \frac{t}{RC}[/itex], therefore [tex]\frac{\partial V}{\partial \tau} = \frac{\partial ^2 V}{\partial \xi^2}[/tex]

So far, I am happy with this. Now this is where I start to have some questions

3) Redefine the independent variable to non-dimensionalize and simplify B/ICs
We want to simplify the boundary condition [itex]V(t, 0) = V_0[/itex] and we therefore let [itex]U = \frac{V(t, 0)}{V_0}[/itex]. So now that boundary condition is = 1.

However, does making this change not introduce a new free parameter back into the geral expression?
[tex]\frac{\partial \left(U V_0 \right)}{\partial \tau} = \frac{\partial ^2 \left(U V_0 \right)}{\partial \xi^2}[/tex]
The source I am using seems to suggest that step 3) does not change the GE in the way I suggested, but instead changes V to U directly, but I cannot see why.

I see that we are non-dimensionalising it, but we did the same thing with both of the other variables and we subsituted their new values into the GE.

Thank you, any help is appreciated.
 
I have solved my misunderstanding, when we apply the partial derivative twice on the RHS, we are not squaring the numerator and hence there will only be one [itex]V_0[/itex]. Therefore, the [itex]V_0[/itex] (to the power 1) in the numerator of both the LHS and RHS of the equation will cancel out.
 
  • Like
Likes   Reactions: BvU