8 APPROXIMATE METHODS
can now be written as
εφ
ί
άΩ = 0 i=l, 2, ... (1.23)
It is also common in Galerkin method to write the
w
function as Au,
i.e.
w = Au
where
Au = Αα
1
φ
ί
+ Ja
2
$2 + ^
a
-3$3+ · · · (1-24)
and
AoLi
= /?
f
.
This representation is sometimes preferred to indicate that
w
can be
identified with a variation or virtual quantity (such as virtual
displacements or velocities). Expression (1.24) also indicates that the
same functions are used for
u
as for
Au.
This.property is important as
it produces symmetry in the expressions as we will see in Example 1.3.
Example 1.3
Let us return to our original equation, defined in Example 1.1 and try
to solve it using Galerkin's. We can choose the same approximating
functions as in Example 1.2, i.e.
u = (χ
ι
φ
ι
+<χ
2
φ
2
= ajx(l -χ) +
α
2
χ
2
(1
-x) (a)
Note that as previously,
α
1?
α
2
are not nodal values of u but
unknown generalised coefficients.
The weighted residual statement is,
ί
ενν
dx = 0 (b)
c
which produces the following two equations:
I
εφ
ι
dx = 0 and
εφ
2
dx = 0 (c)
The ε function is the same as in Example 1.2. Hence,