Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Dan Shields
December 15, 2011
Introduction
It is important in nuclear science and engineering to understand heat flow in and around
fuel in reactors. This is still an open area of research, although it is well understood for most
commonly used systems, solutions can be drastically changed for only subtle changes in parameters. This paper attempts to take a very
simple approach to the fist steps in building up
this theory of reactor fuel. I will apply the
concepts of reactor theory from E. Lewis text
on the subject[1], along with A. Fetters and J.
Waleckas text on classical mechanics [2].
First I will give the background necessary for
understanding the results. This will include a
brief look at the basic solution for a nuclear fuel
rods energy production distribution, a review of
thermal physics and extrapolation of a solution
to the temperature distribution of the fuel over
time.
As the equations governing this process for
nontrivial energy production inside our the fuel
is highly complicated and almost impossible to
solve for by hand analytically, I will only derive
the first principals that then could be used to
generate a numerical solution for the process.
This is exactly what modern models are built
on, so it is very useful to understand where they
come from.
(H
~ D (~
~ r) + f (~r)(~r) a (~r)(~r) = 0 (1)
2 (~r) + B(~r) = 0 ; B =
k /k 1
L2
Z
f dV =
P =
V
H
2
f dzdr
(5)
H
2
(2)
r (r, z) + 2 (r, z) + B 2 (r, z) = 0 (3) power density (P ) at any point:
r r
z
This can be solved by separation of variables
so that (r, z) (r)(z) as the rod is cylindrically symmetric. The derivation is spelled out
in [1], but I will simply state the result 2 :
(r, z) = CJ0 (
1
2.405 r
z
) cos (
)
R
H
V
R
H
(7)
This approximates the scattering processes that neutrons go through as a diffusion process that is directly
analogous to something like salt diffusing in water. See
Ficks Law.
2
Note: The boundary conditions are set to 0 at the
H
. This gives good approxiextrapolation lengths R,
mation as to the actual at the true boundaries R, H.
This is described in [1]. Also the number in our equation
comes from J0 (2.405...) = 0.
variable volume. For solids, this can be misleading and harder to solve, and one can restate
We aim to get to how heat is conducted in
this as just the opposite. This retains the same
this rod now, but first we must go over some of
meaning (p v = W ). So to make our equathe basic concepts from thermodynamics. The
tion more functional we introduce a new equamost important concepts being:
tion from a Legendre transformation of eq.8 to
find Enthalpy (H) :
The Laws of Thermodynamics:
(9)
dH = d(E + pv) = T
dS} + v dp
1st : The internal energy of a system is
| {z
|{z}
, dQ
, dW
equal to the amount of heat supplied
Now to further simplify we define a parameter
to the system, minus the amount of
work performed by the system on its for our system in terms of constant pressure: the
surroundings. (A statement of energy heat capacity:
conservation)
Q
S
H
2nd : This is an expression of the tendency
Cp = (
)p = T ( )p = (
)p
(10)
T
T
T
that over time, differences in temperature, pressure, and chemical potential
This implies that we can integrate H over a
equilibrate in an isolated physical sys- range of temperature of find the change in H:
tem. (Entropy (S) is always increasing
for spontaneous processes)
Z
T
dT 0 Cp (T 0 , p)
H(T, p) = H(T0 , p) +
(11)
T0
, dW
(12)
Cpo = Cp (T0 )
T
(~
r
,
t)
+
(r)
(17)
~
~
= dA jh + q
(14)
t
cp
cp
T (~r, t)
~ j~h
=
t
T(r, s) =
dt est T (r, s)
(19)
0
q
T (~r, t)
= 2 T (~r, t) +
(15)
t
cp
kth
Where : =
; kth is the thermal diffusivity
cp
s) = 2 T(r,
t) +
sT(r,
000
P0000
(r)
cp s
(20)
1 2 s
r
T T=0
r2 r r
(21)
s) =
T(r,
P0000
sinh s 2 (0 )
(26)
1
1
4cp s 2
sinh 0 s 2
This gives us an inclination as to go about
solving the problem by integrating over space
Now we need to convert back to time to undo
around the point source of heat to get a bound- our Legendre transformation in t, thus:
relaary condition at the origin for our new T
tion:
Z
1
P0000
ds st sinh s 2 (0 )
e
T (r, t) =
1
1
4cp s 2 C 2i
sinh 0 s 2
Z
(27)
s) =
d3 r T(r,
(22)
s
This is not a trivial solution to find. This
Z r<
000 Z
becomes much easier if one uses complex contour
t) + P0
d3 r (r)
d3 r 2 T(r,
c
s
p
r<
r<
integration about the poles that appear for our
contour C (as seen in the above integral). We
We can then use a similar integration about
have a countably infinite set of these, so we gen
the equation describing the conservation of heat
an infinite sum (that does prove to be convergent
flow descried in the previous section. This will
for all t > 0):
enable us to find a true boundary condition:
Z
000
s) + P0 ]
lim[ = dA T(r,
0 r
cp s
1 1 X 2
nr n2 22 t
sin
e a ]
T (r, t) =
1 [( )
a
4cp s 2 r a n=1 nr
(28)
This then a superposition of sinc like functions
from 0 until their first zero. They all decrease
exponentially in time, meaning that over time
our temperature moves towards 0. This makes
some sense: we will see a function that starts
as something like an erfc(r) that has a zero at
the boundary and as a maximum temperature
at r = 0.
P0000
(23)
lim 4
0
P 000
T (r, s)
+ 0 =0
r
cp s
(24)
s)
dT(r,
P0000
)=0 =
1
d
4cp s 2
(25)
r r
R
For our solutions from the first section, we can
2
z
2
use a basic analytic method to try and look at
T
(z)
+
k
T
(z)(r,
z)
=
C
cos
(
)
z
z
z 2
H
the fundamentals of the solution.
Solution for a Fuel Rod
V cp
R
H
Taking for the solution from my notebook, we
This seems to be a prime target for separation find a solution that seems reasonable for the Tz
of variables as our P 000 function is already sepa- term:
rated in r and z and has no t dependence. Thus
we say:
z
Tz (z) = [(C H Cos
H
2 + (Hk
)(Hk
+ )T0
+ CH
2
Cos[kz]Sec[Hk])]
(33)
+ ))
(Hk + )(Hk
R
ODE. This is unfortunate, but has not stopped
z
2
2
( + k )Tz (z) = C cos (
)
us from building reactors. This is due to the fact
H
that very accurate solutions can be estimated by
numerical methods to solve for this exact type
P
) ; k=
C = (3.63
of problem.
V cp
Conclusions
So from our analysis we have build up the ma- References
chinery needed to solve any general heat prob[1] E.E. Lewis. Fundamentals of Nuclear Reaclem. This becomes very difficult and not in the
tor Physics. NY: Academic Press, 2008, pp.
least trivial for most realistic energy inputs like
139-166.
our neutronics solution. Although the solution
for P 000 can be easy to solve and interpret, heat [2] A.L. Fetter, J.D. Walecka. Theoretical Meconducts very differently, and is highly dynamic.
chanics of Particles and Continua. NY:
We must now use the governing equations in
Dover Publications Inc., 1980, pp. 406-433.
a numerical solution to try and find a close fit to
actual heat conduction. This has been shown to
be possible, and there are many ways in which to
approach it. For the interested reader, they only
Attached: Mathematica pertinent to solution of Tz , Tr
need to search nuclear papers heat conduction of
fuel rods.