Information Technology Reference
In-Depth Information
Thus, the unknowns
u
are obtained by solving
k
u
u
=
p
u
(7.58)
i
v
N
u
In the case of a uniform beam,
R
=
(
EI
w
)
−
p
z
and
w
=
w
,
so that Galerkin's relations
would be
L
L
EI
N
i
u
)
N
u
(
N
u
dx
w
−
p
z
dx
=
0
0
0
(7.59a)
−
=
k
u
w
p
u
0
or
k
u
w
=
p
u
(7.59b)
Remember that it is necessary that the basis functions
N
u
satisfy
all
of the boundary condi-
tions. They must also be sufficiently differentiable.
Galerkin's method can also be viewed from the standpoint of a variational technique.
R
ecall from Chapter 1 or 2 that the solid continuum relations of Eq. (7.5a),
D
T
E
Du
+
D
T
ED
p
V
=
0
,
is an equilibrium expression. Thus, the residual (Eq. 7.56)
R
=
u
+
p
V
is
an “out-of-balance” force. A reasonable variational integral would seem to be
u
T
R
dV
V
δ
=
0
(7.60)
since this is of the form of work. Thus,
u
T
D
T
ED
V
δ
(
u
+
p
V
)
dV
=
0
(7.61)
N
u
With
u
=
u
,
Eq. (7.61) becomes
u
T
N
u
(
D
T
ED N
u
V
δ
+
p
V
)
=
u
dV
0
(7.62a)
or
u
T
dV
dV
δ
N
u
(
D
T
ED N
u
)
N
u
p
V
u
+
=
0
(7.62b)
V
V
k
u
−
p
u
so that
u
can be obtained from the familiar linear relationship
k
u
u
=
p
u
(7.63)
These are the same expressions given in Eq. (7.57), which were based on Galerkin's
weighted residual method.
Galerkin's relations of Eq. (7.59) for a uniform beam are readily derived from a variational
integral. Begin with
L
0
(
i
v
EI
w
−
p
z
)δw
dx
=
0
(7.64)
which, with
w
=
N
u
w
,
becomes
w
T
L
0
N
u
EI
N
i
u
dx
N
u
p
z
dx
L
δ
w
−
=
0
(7.65a)
0
Search WWH ::
Custom Search