Digital Signal Processing Reference
In-Depth Information
The solution to this partial differential equation can be found in terms of two
ordinary differential equations, assuming that the potential can be represented by
the product of a function for each coordinate [Jackson, 1999]:
(x,y)
=
X(x)Y(y)
(3-49)
If equation (3-49) is substituted back into (3-48) and then divided by
, the
result is
d
2
XY
dx
2
d
2
XY
dy
2
d
2
X
dx
2
d
2
Y
dy
2
1
XY
1
XY
1
X
1
Y
+
=
+
=
0
(3-50)
where the partial derivatives are replaced by total derivatives because each term
involves only one variable. This allows us to write two separate ordinary differ-
ential equations:
d
2
X
dx
2
1
X
β
2
=−
(3-51)
d
2
Y
dy
2
1
Y
=
β
2
(3-52)
The constant
β
must satisfy the conditions
β
2
+
(
−
β
2
)
=
0 to satisfy Laplace's
equation. The differential equations in (3-51) and (3-52) are solved by finding
the roots of the characteristic equations:
X(x)
=
C
1
n
e
jβx
+
C
2
n
e
−
jβx
=
C
1
n
(
cos
βx
+
j
sin
βx)
+
C
2
n
(
cos
βx
−
j
sin
βx)
(3-53)
Y(y)
=
C
1
n
e
βy
+
C
2
n
e
−
βy
=
C
1
n
(
cosh
βy
+
sinh
βy)
+
C
2
n
(
cosh
βy
−
sinh
βy)
(3-54)
The potential is then calculated by substituting (3-53) and (3-54) into (3-49):
(x,y)
=
X(x)Y(y)
=
[
C
1
n
(
cos
βx
+
j
sin
βx)
+
C
2
n
(
cos
βx
−
j
sin
βx)
]
[
C
1
n
(
cosh
βy
+
sinh
βy)
+
C
2
n
(
cosh
βy
−
·
sinh
βy)
]
(3-55)
Note that since the solution is periodic,
β
must be an integer. The boundary
conditions that must be applied to solve (3-55) are
d
2
(x, y)
=
0
when
x
=±
(3-56)
which is the potential at the sidewalls and
(x, y)
=
0
when
y
=
0
,
∞
(3-57)
Search WWH ::
Custom Search