1
Finite Difference, Y U NO WORK?!
Posted by
Akano Toa of Electricity
,
in
Math/Physics
Dec 09 2012
·
51 views
E&M FORTRAN Programming
So, I'm working on a computer project for my Electrodynamics course. I'm using a computer method called the Relaxation or Finite Difference Method. It basically takes a physical scenario, divides the space of interest into a grid, and assigns voltages for each grid intersection. Then, using a computer language of choice (I'm using FORTRAN, like a boss), I make a program that essentially takes a weighted average of all the points whose voltages aren't fixed until the program doesn't change those voltages anymore. This gives a surprisingly good approximation for a physical system.
I'm basically modeling a system with two conducting cylindrical shells of equal radius separated by some height and which are at voltages +V0 and -V0. The problem is that my output graphs do not look physical; the voltage just drops to near zero rapidly for points outside and between the cylinders, whereas I expect that the graph should gradually drop.
Curse you, technology!
I'm basically modeling a system with two conducting cylindrical shells of equal radius separated by some height and which are at voltages +V0 and -V0. The problem is that my output graphs do not look physical; the voltage just drops to near zero rapidly for points outside and between the cylinders, whereas I expect that the graph should gradually drop.
Curse you, technology!
