Agriculture Reference
In-Depth Information
set of N equations having N unknown concentrations. The form of the N
equations is
+
+
=
aC
b
Cu
C
e
++
(3.71)
ij
,
i
-1,
j
+
1
ij
,
ij
,
+
1
ij
,
i
1,
j
1
ij
,
where N is the number of incremental distances in the soil ( N = L x ). By
including the appropriate initial and boundary conditions in their finite-
difference forms, Equation 3.33 can be written in matrix-vector notation as:
n
+
1 =
(3.72)
Bh
(
)
w
where B is a tridiagonal real matrix and h and w denote the associated real
column vectors (the arrows indicate vectors). The matrix B may be written as:
n
n
b
u
1
1
n
n
n
a
b
u
2
2
2
a
n
b
n
u
n
3
3
3
_
(3.73)
n
n
n
a
b
u
N
1
N
1
N
1
n
n
a
b
N
N
The coefficients of the main diagonal of the matrix B, in absolute values, are
greater than the raw sum of the off-diagonal coefficients. Hence, the matrix
B is strictly diagonally dominant. Therefore, the matrix is nonsingular, and
there exists a solution n + 1 for the matrix vector equation (Equation 3.71) that
is unique. The tridiagonal system of equations was solved by an adaptation
of the Gaussian algorithm. Therefore, for any time step, n + 1, where all vari-
ables are assumed to be known at time step n, Equation 3.71 can be solved
sequentially until a desired time t is reached.
The coefficients a , b , u , and e are the associated set of equation parameters.
The above N equations were solved simultaneously, for each time step,
using the Gaussian elimination method (Carnahan, Luther, and Wilkes,
1969) in order to obtain the concentration C at all nodal points ( i ) along the
Search WWH ::




Custom Search