Environmental Engineering Reference
In-Depth Information
Let M and S be the point and the boundary of one-, two- or three-dimensional
domain. When the dependent variable u is given on the S , the boundary condition
reads
u
(
M
,
t
) | S = ϕ (
M
,
t
)
or
u
(
M
) | S = ϕ (
M
) ,
(1.79)
where
are known functions. Such boundary conditions are called
the boundary conditions of the first kind .
ϕ (
M
,
t
)
and
ϕ (
M
)
Boundary Conditions of the Second Kind
In many applications, boundary conditions are not given by values of the depen-
dent variable on the boundary, but by its directional derivative along the boundary
normal, the outward-drawn normal in particular. For one-dimensional cases, we can
express a directional derivative by a partial derivative.
Consider the vibration of a string of length l . If the string can slide freely along
the y -direction at x
=
0 without any force, a force balance at x
=
0 yields
x = 0 =
T
u
0or u x (
0
,
t
)=
0
.
x
This is the boundary condition at x
=
0. If the string slides along the y -direction at
x
=
l with the action of a force f
(
t
)
, the boundary condition at x
=
l is
x = l =
T
u
f
(
t
)
or
u x (
l
,
t
)=
f
(
t
) /
T
,
x
where T is the tension of the vibrating string.
Consider heat conduction in a body. If the heat flux density on the boundary S
is given, the normal derivative of temperature on S is known by the Fourier law of
heat-conduction, say
ϕ (
M
,
t
)
. The boundary condition is thus
S = ϕ (
u
M
,
t
) .
n
If the boundary is well insulated such that the heat flux vanishes on the boundary,
the boundary condition reduces to
S =
u
0
.
n
If the two ends are well insulated for heat conduction in a solid rod of length l ,the
boundary conditions are
(
,
)=
(
,
)=
.
u x
0
t
u x
l
t
0
 
Search WWH ::




Custom Search