Civil Engineering Reference
In-Depth Information
[f1,i
1
(
x
1
,
x
2
,
...
,
x
n
)
=
0
[f1,i
2
(
x
1
,
x
2
,
...
,
x
n
)
=
0
[f1,i
3
(
x
1
,
x
2
,
...
,
x
n
)
=
0
(6.66)
For each equation, a multivariable first-order Taylor series expansion is written.
For example, for the
k
th equation,
∂
[f1,i
k
,
[f1,i
∂
x
1
+
∂
[f1,i
k
,
[f1,i
∂
x
2
[f1,i
k
,
[f1,i
+
1
=
[f1,i
k
,
[f1,i
+
x
1,
[f1,i
+
1
−
x
1,
[f1,i
x
2,
[f1,i
+
1
−
x
2,
[f1,i
(6.67)
x
n
,
[f1,i
+
1
−
x
n
,
[f1,i
∂
[f1,i
k
,
[f1,i
∂
x
n
+···+
where the first subscript (
k
) represents the equation of unknown and the second
subscript denotes whether the value or function is at the present value (
[f1,i
) or at the
next value (
[f1,i
+
1).
By setting
[f1,i
k
,
[f1,i
+
1
to zero, this means that we are looking for the roots of the sys-
tem of equations. This then gives
−
[f1,i
k
,
[f1,i
+
x
1,
[f1,i
∂
[f1,i
k
,
[f1,i
∂
x
1
+
x
2,
[f1,i
∂
[f1,i
k
,
[f1,i
∂
x
2
+···+
x
n
,
[f1,i
∂
[f1,i
k
,
[f1,i
∂
x
n
(6.68)
=
x
1,
[f1,i
+
1
∂
[f1,i
k
,
[f1,i
∂
x
1
+
x
2,
[f1,i
+
1
∂
[f1,i
k
,
[f1,i
∂
x
2
+···+
x
n
,
[f1,i
+
1
∂
[f1,i
k
,
[f1,i
∂
x
n
By examining Eq. (
6.68
), the only unknowns are the
x
k
,
[f1,i
+
1
terms on the right-hand
side of the equation as all other quantities are known at the present value (
[f1,i
). This
now provides a system of linear equations that can be solved. Matrix notation is
used to simplify the expression. The partial derivatives can be expressed by
∂
[f1,i
1,
[f1,i
d
x
1
∂
[f1,i
1,
[f1,i
d
x
n
···
.
.
.
.
.
[
J
]
=
(6.69)
∂
[f1,i
n
,
[f1,i
d
x
1
∂
[f1,i
n
,
[f1,i
d
x
n
···
where
J
is commonly called the Jacobian matrix.
The initial and final values in vector form are as follows:
{
X
[f1,i
}
T
=
x
1,
[f1,i
...
x
n
,
[f1,i
(6.70)
and
{
X
[f1,i
+
1
}
T
=
x
1,
[f1,i
+
1
...
x
n
,
[f1,i
+
1
(6.71)
The function values in vector form are as follows:
{
F
[f1,i
}
T
=
[f1,i
1,
[f1,i
...
[f1,i
n
,
[f1,i
(6.72)