Neutron Diffusion Theory: One Velocity Model: 22.05 Reactor Physics - Part Ten
Neutron Diffusion Theory: One Velocity Model: 22.05 Reactor Physics - Part Ten
Neutron Diffusion Theory: One Velocity Model: 22.05 Reactor Physics - Part Ten
We now have sufficient tools to begin a study of the second method for the
determination of neutron flux as a function of position and energy. This method is,
diffusion theory. Both diffusion theory and Monte Carlo methods are currently in
use as reactor design tools. Both are actually quite similar in terms of set up
because, for each, a detailed model of the reactor is essential and the accuracy of
the result is largely a function of the quality of the model. Once the model is
developed, the two differ in the method of solution.
The model that is needed for diffusion theory has the following inputs:
• A list of the mathematical constraints on the model. Is the core symmetric
and hence only half is modeled with the other half being taken as a mirror
image? What boundary conditions are to be imposed? What geometry (x,
y, z; or r, θ z,) for example, is to be used?
• A list of the materials that will be present. These typically include: light
water, U-235, U-238, Zircalloy, steel, boron, etc.
• For each material, the cross-section is specified. These are usually a
function of neutron energy.
• A geometric description of the reactor. Where is the fuel, the moderator,
the control devices?
• For each geometric location, what material composition is present?
In the introductory lecture, we noted that the goal of reactor physics analysis is to
calculate the neutron population at any given time and at every point r as a
function of energy and angle of travel. Thus, we could attempt to use diffusion
theory to solve for φ(r, E,Ω, t) . Such an approach is feasible but it is rather
daunting. Accordingly, we will first apply diffusion theory for the solution of
φ(r, t) or perhaps only φ(r) . Such an approach is called “one velocity theory”
because we are treating all neutrons as having one speed or energy.
The one velocity model is formally obtained by integrating over both E and Ω.
Thus,
Is it of value? We saw from the life cycle analysis that energy dependence is
essential. The various factors each pertained to a particular energy range. So, we
do not expect a one velocity model to provide useful results for situations where
changes of neutron energy are important. For example, it could not describe a
thermal reactor because, in that case, neutrons are born at fast energies and slow
down to be absorbed by U-235. But, the one velocity model could describe a fast
reactor where neutron slowing down is not as important. In fact, it has been used
for this purpose. In particular, it is of use for bare (unreflected) cores.
3. Equation of Continuity:
The object is to write a relation that will allow the calculation of φ(r, t ) . This is
sometimes called a neutron balance equation or a neutron conservation equation.
The starting point is the word definition of the core multiplication factor:
or
Let n(r,t) be the number of neutrons at r and t per unit volume. Thus,
∂
[n(r,t)dV] = S(r,t)dV - A(r,t)dV - L(r,t)dV
∂t
2
Where S, A, and L are the source, absorption, and leakage terms, respectively.
The above may be rewritten as:
∂
n(r, t) = S(r, t) − A(r, t) − L(r, t)
∂t
The definition of neutron flux is that φ =nv. Thus, the term on the left may be
written as:
∂ 1 ∂
n(r, t) = (φ(r, t) )
∂t v ∂t
The source term, S, is simply the product of the fission reaction rate and the
number of neutrons per fission. Thus,
A(r, t) = Σa φ(r, t)
This leaves the leakage term to be evaluated. To repeat from the previous section:
let J be the neutron current density vector on the surface of V and let n in a unit
normal pointing outward from the surface. Thus, the leakage of neutrons from
diffusion through the surface A of volume V is:
= ∫v ∇ • J (r, t)dV
1 ∂φ(r, t)
= νΣFφ(r, t) − Σa φ(r, t) − ∇ • J(r, t)
v ∂t
If the flux is not a function of time (reactor exactly critical), then, we obtain
4. Fick’s Law:
“As might be expected, the best that can be done to relate J to Φ without actually
finding a solution for N(r, Ω,E) is an approximation. Fortunately there turns out to
be a rather accurate approximation that is quite satisfactory for many reactor-
design calculations. It is called Fick’s Law, and it relates J(r,E) to ∇φ(r, E), by the
equation
This relationship between the net current and the gradient of the scalar flux seems
plausible if we recall that φ(r, E) is v(E) n(r, E), where n(r,E) dE is the number of
neutrons in dE per unit volume. Thus Fick’s Law states that the net current across
a surface of neutrons with energies in dE is proportional to the rate of decrease of
the density of neutrons in dE across that surface. Moreover the direction of the net
current is the direction in which n(r, E) is decreasing at its maximum rate. (Recall
that, mathematically, the gradient of a function is a vector pointing in the direction
of the maximum rate of increase of that function; thus − ∇n(r, E) points in the
direction of the maximum rate of decrease of n( r ,E).) The neutron population
thus tends to drift from a region of high concentration to one of low concentration,
like gas diffusing through a porous plug, and this drift gives rise to a net current.”
Fick’s law is valid provided that there are roughly equal number of neutrons
moving in all directions at all energies. Consider two material compositions.
Fick’s Law can model the first situation but not the second. In general, Fick’s law
is not valid whenever there are abrupt changes in the scatter and absorption cross-
sections such that the net flow of neutrons is in one direction only. Thus, it is
expected to:
a) Be valid whenever more than a few mean free paths from a surface.
b) within three mean free paths of either a neutron source or the surface of a
material; and
λ 1
D = tr =
3 3Σ tr
and
1 1
λ tr = 1/ Σ tr = ≅
[Σ t − μ0Σs ] Σs (1 − μ0 )
Σ t = Σ a + Σs
2
μ0 =
3A
Where A is the atomic mass of the medium through which the diffusion occurs.
Other symbols are:
λ tr is the transport mean free path (cm),
μ0 is the average value of the cosine of the angle at which neutrons are
scattered (Lab System).
Note: The above expression for D and the theoretical basis for Fick’s Law can be
obtained from the study of the transport equation.
7. Neutron Diffusion Equation: The leakage term in the continuity equation can
now be written as:
or,
∇ • J(r) = −D∇ 2φ(r)
1 ∂φ(r, t)
= νΣFφ(r, t) − Σa φ(r, t) + D∇ 2φ(r, t)
v ∂t
a) Limits on Flux:
φA = φB
(J A )n = (J B )n
Where φA and φB are the neutron fluxes in A and B respectively and (JA)n
and (JB)n are the normal components of the neutron currents both evaluated
at the interface. Note that the gradient of the flux (dφ / dx ) is NOT
continuous across the boundary. That is:
J = −Ddφ / dx
c) E
xtrapolation Distance: Fick’s Law is not valid at surfaces and hence
neither is the diffusion equation. However, transport theory is valid.
Comparisons of transport (exact) and diffusion (valid only 3 mean paths
from surface) calculations show that if the flux, as calculated from
diffusion theory, is assumed to go to zero at some distance d beyond the
surface, then the diffusion theory prediction in the interior of the medium
will be accurate. The following figure illustrates the idea.
Diffusion
Theory
Surface
Transport
Theory
Diffusing Medium
The value of the extrapolation distance d is found to be 0.71 λtr and because
the diffusion constant equals λtr/3.
d = 2.13D
Most values of d are a few cm or less. For large reactors where d < <
actual size, d can be set to zero.
The MIT Research Reactor (MITR) is a hexagonal core that contains three rings of
fuel. The inner or A-Ring has three elements, the middle or B-Ring has nine, and
the outermost or C-Ring has fifteen. (See following figure.) Light water, which
serves as both coolant and moderator, flows through the fuel elements that form
each ring. In addition, there are plenums of light water both above and below the
core. As one moves radially outward from the C-Ring, one encounters a region of
light water, the control blades (borated stainless steel), the aluminum core tank,
and the heavy water reflector.
The MITR was designed using a diffusion theory code. Experiments were then
done to verify the accuracy and to develop flux correlation factors for regions
where the diffusion theory prediction was inaccurate. One would expect diffusion
theory to fail at the fuel water interfaces which are the plenums that are above and
below the core and the region of light water that surrounds the C-Ring.
Measured Value
Correction Factor =
Pr edicted Value
The attached figure shows the axial correction factors for the A and C-Rings. As
expected, the value of the factor is near unity (excellent agreement) for the interior
portion of the A-Ring. It deviates from unity (poor agreement) at both the lower
and upper regions because of the presence of the water plenums. Fast neutrons are
traveling into those plenums; thermal neutrons are returning. So, Fick’s Law does
not hold. The agreement is marginal for all of the C-Ring which forms a surface
with the control blades (strong absorbers) or the light water moderator. So, we
conclude that our understanding of the capabilities and limitations of diffusion
theory is correct.