Multidimensional Reactor Systems in Diffusion Theory PDF
Multidimensional Reactor Systems in Diffusion Theory PDF
Multidimensional Reactor Systems in Diffusion Theory PDF
DIFFUSION THEORY
© M. Ragheb
12/9/2015
1. INTRODUCTION
The treatment of multidimensional systems, other than the spherical geometry, is
an important topic since reactor systems normally have a finite cylinder geometry. Two
approaches are possible: numerical methods, and in the simplest case, the separation of
variables method can be used for homogeneous systems. We start with a mathematical
introduction on orthogonal and orthonormal functions. The method of separation of
variables is then used to study the criticality and the flux distribution for the
parallelepiped reactor geometry; of which the cube is a special case, and the finite
cylinder reactor core which is the geometrical configuration of most existing nuclear
power reactors. The treatment will cover the case of multiplying media. The minimum
volume for a critical assembly and the peak to average flux ratio will be derived.
Two functions m (x ) and n (x) are said to be orthogonal over an interval [a,b] if
the integral of the product m n over that interval vanishes:
a
m ( x) n ( x)dx 0
In a more general sense, the functions m (x) and n (x) are said to be orthogonal
with respect to a weighting function r(x), over an interval [a, b], if:
r ( x)
a
m ( x) n ( x)dx 0
d2X
2
2 X
dx
with the boundary conditions: X(0) = X(a) = 0, has the eigenvalues or characteristic
values:
n
2
n2
a
n x
X n ( x ) sin( )
a
mx nx
a a
X m ( x) X n ( x)dx sin(
0 0
a
) sin(
a
)dx 0 , for m n
b
Cn r ( x ) n ( x ) dx
2
nx
a a
a
C n X dx sin 2 (
2
n )
0 0
a 2
nx
Thus in order to normalize the functions sin( ) over the interval [0, a], we
a
would divide them by the normalizing factor:
a
.
2
2 nx
X n ( x) sin( )
a a
is thus an orthonormal set in the interval [0,a].
2 ( x, y, z ) 2 ( x, y, z ) 2 ( x, y, z )
Bg2 ( x, y, z ) 0 (1)
x 2
y 2
z 2
We can use the separation of variables method to solve the partial differential
equation, by assuming:
( x, y, z ) X ( x)Y ( y ) Z ( z ) (3)
2 X 2Y 2Z
YZ XZ XY Bg2 XYZ 0 (4)
x 2
y 2
z 2
1 d 2 X 1 d 2Y 1 d 2 Z
2
2
2
Bg2 (5)
X dx Y dy Z dz
Each term must be separately equal to a constant if it is to hold for all allowed
values of x, y, z. This results in three ordinary rather than partial differential equations:
1 d2X
2
X dx 2
1 d 2Y
2
2 (6)
Y dy
1 d 2Z
2
2
Z dz
d 2 X ( x)
2
2 X ( x ) (8)
dx
X ( x ) A cos x C sin x
since we are not interested in the trivial solution A = C = 0, this can be satisfied if:
n
C = 0; a' for all n odd
2
or by:
n
A = 0; a' for all n even
2
Thus:
n
2
2
(9)
2a '
The even and odd solutions can be taken as members of the normalized set:
1 nx
X n ( x) cos( ) , for all n odd
a' 2a '
1 nx
sin( ) , for all n even (10)
a' 2a '
1 py
Y p ( y) cos( ) , for all p odd
b' 2b'
1 py
sin( ) , for all p even (11)
b' 2b'
Thus:
p
2
2 (12)
2b'
1 qz
Z q ( z) cos(
) , for all q odd
c' 2c'
1 qy
sin( ) , for all q even (13)
c' 2c'
Thus:
q
2
2
(14)
2c'
The geometrical buckling is the sum of the B g2(npq ) given by Eqns. 7, 9, 12 and 14:
n 2 p q
Bg2( npq ) ( ) ( )2 ( )2 (15)
2a ' 2b' 2c'
npq ( x, y, z ) X n ( x )Y p ( y ) Z q ( z ) (16)
The only choice of n, p, and q which gives a nonnegative flux over the whole core
is:
n p q 1
Bg 2 Bg2(1,1,1) ( )2 ( )2 ( )2 (17)
2a ' 2b ' 2c '
1 x y z
( x, y, z ) A'111 ( x) A' cos( ) cos( ) cos( ) (18)
a ' b' c ' 2a ' 2b' 2c'
V 8.abc (19)
1
2B 1 1 2
a ' a d ( g )2 2 2 (20)
b' c'
V V
0
b c
1
V 2 B 1 1 2 8cb 1
8c ( )
g 2
2
d 2
b (b d ) ( c d ) (b d ) 2
2 3
3
2 Bg 2 1 1
2
2
0
(b d ) ( c d )
2
1
V 2 Bg 2 1 1 2 8cb 1
8b ( ) 2
d 2
c (b d ) ( c d ) ( c d ) 2
2 3
3
2 Bg 2 1 1
2
2
0
(b d ) ( c d )
2
b
a (a d ) 3
(b d ) 3
c
a (a d ) 3 (22)
(c d ) 3
b c
(23)
(b d ) 3
(c d ) 3
abc
Thus the critical parallelepiped with minimum volume is found to be a cube with:
a ' b' c' 3 (24)
2 Bg
A
c b a
x y z
V c b a
cos( ) cos( ) cos( )dxdydz
2a ' 2b' 2c'
x y z
a b c
2 a ' b' c '
A ( )3 sin( ) sin( ) sin( )
8abc 2a ' a 2b' b 2c' c
max ( x y z 0) A (26)
Thus:
1
max abc a b c
( )3 sin( ) sin( ) sin( ) (27)
2 a ' b' c ' 2a ' 2b' 2c'
1 ( r, z ) 2 ( r, z )
(r ) Bg2 ( r, z ) 0 (29)
r r r z 2
(r , z ) R(r ) Z ( z ) (30)
We get:
1 1 R 1 2Z
(r ) Bg2 0 (31)
R r r r Z z 2
1 1 d dR
(r ) 2 (32)
R r dr dr
1 d 2Z
2
2 (33)
Z dz
2 2 B g2 (34)
Z ( z ) A cos(z ) (37)
and:
h' n
, for all n = 1, 3, 5…
2 2
or:
n 2
2 ( ) for all n odd (38)
h'
d 2 R 1 dR
2
2 R (39)
dr r dr
d2y dy
x2 2
x (x2 n2 ) y 0
dx dx
where n is a constant.
This has a general solution:
y ( x) EJ n ( x) FYn ( x)
d2y dy
x2 2
x (x2 n2 ) y 0
dx dx
y ( x) E ' I n ( x) F ' K n ( x)
The four Bessel functions of zero order are shown in Fig. 3, and are compared to
the cos (x) function.
It can be noticed that both J 0 ( x ) and Y0 ( x) are oscillatory. The distance between
the roots, or the values at which the functions have a value of zero when they cross the x-
axis, become larger and approach the value of as x increases. The amplitudes of these
two functions decrease as x increases, and they are bounded and not infinite everywhere
except for Y0 ( x) at x = 0, which reaches - .
It is of interest to note that the first root or zero of the function J0 occurs at x =
2.405.
The two functions, I 0 ( x) and K 0 ( x ) are non-oscillatory and unbounded, the
former going to at x = , and the latter at x = 0.
Figure 3. Bessel functions of zero order compared with the cosine function.
The general solution of Eqn. 39 in terms of the Bessel functions of the first and
second kind of zero-th order is:
For the flux to vanish at the extrapolated boundary, EJ 0 (a ' ) 0 , we must have:
p a ' jo , p , for all p = 1, 2, 3, …
2.405
a'
and:
2.405r
R(r ) EJ 0 ( ) (42)
a'
2.405r z
(r , z ) max J 0 ( ) cos( ) (43)
a' h'
2.405 2
B 2 12 12 ( ) 2 ( ) (44)
h' a'
max a 2
z 2.405r
2
cos( ) J o ( ) 2rdrdz
a h 0 h h' a'
2
where: rJ 0 (r )dr rJ 1 (r )
Thus:
Table 1. Geometrical Buckling and flux distribution in different nuclear reactor core
geometries.
Reactor core
Geometric buckling Flux distribution
shape
r
2
Sphere
Radius: R B
2
sin
(r) A
g
R R
r
R
x y z
2 2 2
Rectangular
parallelepiped B
2
( x, y, z ) A cos cos c os
g
a b c a b c
Side lengths: a,
b, c
x y z
2
Cube
Side length: a B 3
2
( x, y, z ) A cos cos cos
g
a a a a
2.405 2.405r z
2 2
Finite height
cylinder B 2
( r, z ) AJ 0 cos
g
R H R H
Radius: R,
height: H
2.405r
2
Semi infinite 2.405
cylinder B
2
( r ) AJ 0
g
R R
Radius: R
x
2
Semi infinite
slab B
2
( x ) A cos
g
a a
Thickness: a
We calculate the parameters that are characteristic of a PWR core using the one-
group diffusion theory model.
From the table, the summation of the total macroscopic cross sections for the
whole homogenized core are;
tr 0.03618 cm 1
a 0.1532 cm 1
f 0.157 n.cm 1
The diffusion coefficient is:
1 1 1 1
D tr 9.213 cm
3 3 tr 3 0.03618
f aF f 0.157
k pf .1.1. 1.0248
aF a a 0.1532
where we have adopted the values:
p 1
D 9.213
L2 60.137cm 2
a 0.1532
L 60.137 7.754cm
1 0.71
d 0.71tr 0.71 19.62cm
tr 0.03618
The axial geometrical buckling is thus:
2 2
2
B
2
H e H 2d 370 (2 19.62)
z
2 2
3 2
(7.68 10 )
370 39.24 409.24
5 2
5.9 10 cm
We can thus deduce the value of the critical extrapolated radius as:
2
2.405 4 2.405 240.5
B
2
r 3.534 10 Re 127.9cm
Re 3.534 104 1.88
One can also estimate the neutrons leakage fraction from the critical core as:
1 1
PL 1 1 1
1 L B
2 2
1 60.137 4.124 x104
1 1
1 1 1 0.976
1 0.0248 1.0248
0.024 2.4 percent
REFERENCES
1. M. Ragheb, “Lecture Notes on Fission Reactors Design Theory,” FSL-33, University
of Illinois, 1982.
2. J. R. Lamarsh, “Introduction to Nuclear Engineering,” Addison-Wesley Publishing
Company, 1983.
EXERCISES
1. Consider a cubical bare reactor of edge dimensions a = b = c, as a special case of the
rectangular parallelepiped reactor core.
Use separation of variables to solve from first principles for the neutron flux in the
reactor as a function of position, and derive an expression for the geometric buckling.
2. Consider a cylindrical bare reactor of unit height to diameter ratio: H = 2R, where H is
its height and R is its radius, as a special case of the finite cylindrical reactor core.
Use separation of variables to solve from first principles for the neutron flux in the
reactor as a function of position, and derive an expression for the geometric buckling.
3. Choose a different critical core configuration with a unity height to diameter ratio and
with the same material compositions, and recalculate the reactor parameters for the
typical PWR core. Discuss your results.
4. For:
a) A spherical reactor core with R = 20 cm,
b) A cubical reactor core with a = 40 cm, and
c) A cylindrical reactor core with H = 2R = 40 cm,
containing a mixture of U235 as fuel and graphite as a moderator, compare the moderator
to fuel ratios : S = Ng / Nu that will achieve criticality for each configuration.
Use: ρ(graphite)=1.6 g/cm3, microscopic absorption cross-section of graphite σa = 3.4x10-
3
b, microscopic absorption cross-section of U235=681 b, ν = 2.07, D = 0.85 cm.
5. Compare the critical masses of fast reactors composed of U235 in the following
geometrical shapes:
a) A spherical reactor core.
b) A cubical reactor core.
c) A cylindrical reactor core with H = 2R.
Use:
microscopic transport cross section = 8.246 [barns]
microscopic absorption cross section = 2.844 [barns]
density = 18.75 [gm/cm3]
product of average number of neutrons released in fission () and the microscopic fission
cross section = 5.297 [neutrons.barn].
6. Using the one group steady state neutron diffusion equation and ignoring the
extrapolation lengths, derive the expression for the flux distribution in a finite height
cylindrical reactor core of radius R and height H as shown in Fig. 1.
Apply the appropriate boundary conditions and derive the expression for the geometrical
buckling for such a reactor core.
By equating the geometrical buckling to the material buckling, derive the one group
criticality equation for the finite height cylindrical core.
Generalize the one group criticality equation to a two group formulation including a fast
neutrons group with Fermi age and a thermal group with diffusion area L2.
For a large reactor deduce the modified one group criticality equation in terms of the
migration area M2 = + L2.
H/2
y
H/2
R
x
15.24 cm
R
Figure 2. Lattice configuration and the unit cell of the natural uranium rods in the D2O
moderator.
Consider a D2O cooled and moderated reactor containing 2.54 cm diameter vertical rods
of natural uranium arranged in a square lattice configuration with pitch (spacing) of 15.24
cm, and suspended in the D2O as shown in Fig. 2.
The height to diameter ratio (H/2R) of the cylindrical core is 1.2.
The infinite medium multiplication factor k∞ = 1.28.
The diffusion area for thermal neutrons is L2 = 175 cm2, and the Fermi age for fast
neutrons is = 120 cm2.
Calculate the following reactor parameters:
a. Material buckling.
b. Critical radius, critical height and critical volume.
c. Fast neutrons non-leakage probability Pf and thermal neutrons non-leakage probability
Pth
d. Estimate of the number of natural uranium fuel rods that would fit in this core.
e. Calculate the weight of natural uranium and of the heavy water to be procured to make
this reactor just critical and the total weight of the whole reactor core in metric tonnes.
The density of the natural uranium metal can be taken as 19 gm/cm3, and the density of
heavy water as 1.1 gm/cm3.