NUEN 304: Nuclear Reactor Analysis

Lecture 4 – 1-Group Diffusion Applications

Sean Simonian
Graduate Research Assistant
Department of Nuclear Engineering
Thermal Hydraulics Verification and Validation Laboratory
 Learning Strategy: Feynman Technique
• You are not here to get a NUEN degree, you are here to become prepared with
the knowledge and skills that will be expected of you because of that degree.
• This class builds on NUEN 301 and 302, knowing this material is essential.
• For any quantity in this class (e.g. 𝜎, Σ, 𝜙, 𝜒,…) you should be able to provide:
-Its name, physical meaning (a definition), its units, and an equation if possible
• We should also be familiar with fundamental concepts and their definitions, for
example, there is not really an equation for doppler broadening, but we should
understand it, be able to explain it, and know its importance in NUEN.
• When you want to learn anything new, the Feynman Technique is worth trying:
1) Write the concept down, 2) Explain it in your own words as simply as possible,
3) Interrogate your explanation, 4) Explain it to others, 5) Revise in simpler terms…
• Do this after every lecture and start each HW as soon as you can (PRACTICE!)
 Lecture 4 Preview
Last week, we derived the “one-speed” or “one-group” diffusion equation (1GD):
Net Leakage Rate Density Production Rate Density

Rate of change of 1 𝜕𝜙 Absorption

− 𝐷∇2 𝜙 = 𝜈𝑓 Σ𝑓 𝜙 − Σ𝑎 𝜙
the neutron density 𝑣 𝜕𝑡 Rate Density
• We need to apply boundary, initial, continuity, and physical conditions to solve.
• For multi-region systems, we will also need to apply interface conditions.
• Each term in the 1GD equation is a physically meaningful rate density.
• In L3, we solved time-dependent and steady-state forms of 1GD.
• Today, we will apply our knowledge of 1GD theory to new problems:
→ Learn how to compute critical masses, volumes, and power peaking factors
→ Study multi-region systems, their interface, and criticality conditions
 Review: 1GD Time-Dependence
When we solved the time-dependent 1GD equation we found:
Ԧ 𝑡 = 𝜙0 𝑅1 𝑟Ԧ T0 exp(− 𝜆1 𝑡) = 𝜙0 𝑅1 𝑟Ԧ T0 exp(− 𝑣𝐷 𝐵𝑔2 − 𝐵𝑚
𝜙 𝑟, 2
𝑡)
Since criticality implies 𝜙 does not change with time, we have now established
the following criticality conditions (you must know and be able to state these):
2 > 𝐵2
𝑘𝑒𝑓𝑓 > 1 ⇒ 𝐵𝑚 𝑔 ⇒ 𝜆1 < 0 ⇒ 𝑆𝑢𝑝𝑒𝑟𝑐𝑟𝑖𝑡𝑖𝑐𝑎𝑙
2
𝑘𝑒𝑓𝑓 = 1 ⇒ 𝐵𝑚 = 𝐵𝑔2 ⇒ 𝜆1 = 0 ⇒ 𝐶𝑟𝑖𝑡𝑖𝑐𝑎𝑙
2 < 𝐵2
𝑘𝑒𝑓𝑓 < 1 ⇒ 𝐵𝑚 𝑔 ⇒ 𝜆1 > 0 ⇒ 𝑆𝑢𝑏𝑐𝑟𝑖𝑡𝑖𝑐𝑎𝑙
To design a bare critical reactor, you, the nuclear engineer, must find the perfect
2 and 𝐵 2 ).
marriage between the material and geometric buckling's (𝐵𝑚 𝑔

2
𝜈𝑓 Σ𝑓 − Σ𝑎 𝑘∞ − 1 𝐷
𝐵𝑚 = = 2
, where: 𝐿 ≡ ≡ the diffusion length
𝐷 𝐿 Σ𝑎
 Review: The Steady-State Problem
• We have seen that the steady-1GD equation is an eigenvalue problem.
• We know that only 1 eigenfunction can satisfy the conditions of our problem.
• We gave this eigenfunction a name: the fundamental mode = 𝑅1 𝑟Ԧ
• Its corresponding eigenvalue is called: the geometric buckling = 𝐵𝑔2
• There is a different flux shape and geometric buckling for each geometry:

Using the definitions herein, the

Fundamental-Mode Scalar Flux is:

𝜙 𝑟Ԧ = 𝐶𝜙 𝑅1 𝑟Ԧ ,

and here, 𝐶𝜙 is determined from the

reactor power (𝑃𝑓 ).
 Review: Grad., Div., Lap. Operators
Recall from multivariable calculus:
• The Gradient operator (∇) operates on a scalar function, outputs a vector
• The Divergence operator (∇ ∙) operates on a vector function, outputs a scalar
• The Laplacian operator (∇2 ) operates on a scalar function, returns a scalar
• Try to get familiar with forms of the operators for each coordinate system:
Table 1: Gradient, Divergence, and Laplacian operators in various coordinate systems
 Review: Power Density, Power, and 𝑪𝝓
The fission rate density is:
fissions
𝐹𝑅𝐷 𝑟Ԧ = Σ𝑓 𝜙 𝑟Ԧ units:
cm3 ⋅ s
And we know the average fission energy release is: 𝐸𝑓 =200 MeV/fission
Thus, multiplying 𝐹𝑅𝐷 by 𝐸𝑓 will convert the quantity into a power density, 𝑞’’’:
𝑞′′′ 𝑟Ԧ = 𝐸𝑓 Σ𝑓 𝜙(𝑟)
Ԧ Note: Dr. Tsvetkov uses:
𝑝𝑓 = power density, and
𝑃𝑜𝑤𝑒𝑟 𝑀𝑒𝑉 𝑓𝑖𝑠𝑠. 𝑛−𝑐𝑚 𝑃𝑓 = reactor power
=
𝑐𝑚3 𝑓𝑖𝑠𝑠. 𝑛−𝑐𝑚 𝑐𝑚3 −𝑠
′′′
To get the reactor power, 𝑃𝑓 , integrate 𝑞 𝑟Ԧ over the power producing region:

𝑃𝑓 = න 𝐸𝑓 Σ𝑓 𝜙(𝑟)
Ԧ 𝑑∀ = න 𝐸𝑓 Σ𝑓 𝐶𝜙 𝑅1 (𝑟)
Ԧ 𝑑∀
∀core ∀core
Applying this equation to each geometry yields the 𝐶𝜙 values in the prior table.
 A Large Reactor’s Fission Rate
Consider the AP-1000 reactor (𝑃𝑒𝑙 = 1000 MW, 𝑃𝑡ℎ = 3400 MW)
- First advanced reactor to be licensed by the NRC
MeV
- Recall the energy released per fission: 𝐸𝑓 = 200
fission
We can approximate the reactor’s fission rate as:
𝑃𝑡ℎ
𝐹𝑅 ≈
𝐸𝑓
Since M = 106, 1 W = 1J/sec, and 1 MeV = 1.602x10-13 J:
Joules
3400 ∗ 106 fissions
𝐹𝑅 ≈ sec ≈ 1020
−13 Joules sec
200 ∗ 1.602 ∗ 10
fission
Or: 100,000,000,000,000,000,000 fissions / second! Model cross-section of an
AP-1000 containment
 The Power Peaking Factor
With some sense of the astronomical fission rate within a large reactor, consider:
- All of this energy is generated within the reactor’s fuel in the form of heat
- A critical reactor has a flux (and power density) that varies with position
- The reactor flux and therefore the fission density peaks in the reactor’s center
- Thus, at the reactor’s center is where we have the highest temperature
• The main thermal limitation of a nuclear reactor is that the fuel centerline
temperature does not exceed the fuel melting point.
• An important quantity in reactor design is the power peaking factor which relates
the extreme (maximum) conditions in the core to the average core behavior.
The power peaking factor is defined as:
Max. Power Density at Hot Spot Location q′′′𝑚𝑎𝑥 1
𝐹𝑃𝑁 = = 𝑞′′′ = න 𝑞′′′ 𝑟 𝑑∀
Average Power Density in the Core 𝑞′′′ ∀𝑐𝑜𝑟𝑒
∀𝑐𝑜𝑟𝑒
 A Useful Table of Peaking Factors
The following table (from Dr. Tsvetkov’s notes) contains power peaking factors
(right-hand column) for bare, homogeneous, critical reactors of various geometries:

You should be familiar with how to derive these (HW and possible exam Q)
 Connection With 𝑘𝑒𝑓𝑓
Recall that we can express the steady 1GD equation as a k-eigenvalue problem:
2
1
−𝐷∇ 𝜙 + Σ𝑎 𝜙 = 𝜈𝑓 Σ𝑓 𝜙
𝑘
We can manipulate the expression to match the form of the Helmholtz Equation:
1 𝜈𝑓 Σ𝑓
−1
𝑘 Σ
∇2 𝑅𝑛 (𝑟) + 𝐵𝑛2 𝑅𝑛 (𝑟) = 0, 𝑛 = 1,2,3, … where: 𝐵𝑛2 = 𝑛 𝑎2
𝐿
The solution to this equation is a 𝐵𝑛2 , 𝑅𝑛 pair, but we also have a relationship between
the 𝐵-eigenvalues and the 𝑘-eigenvalues (from above). Solve it for 𝑘𝑛 :
𝜈𝑓 Σ𝑓
Σ𝑎 𝑘∞ 𝜈Σ𝑓 𝑘∞ 𝜈Σ𝑓
𝑘𝑛 = 2 = 2 = ∴ k eff = 2 =
1 + 𝐿 𝐵𝑛 1 + 𝐿 𝐵𝑛 Σ𝑎 + 𝐷𝐵𝑛2
2 2 1 + 𝐿 𝐵𝑔 Σ𝑎 + 𝐷𝐵𝑔2
2

The 𝑘𝑛 associated with the fundamental mode is the effective multiplication factor, 𝑘𝑒𝑓𝑓
 Size and Composition Effects on 𝑘𝑒𝑓𝑓
Results from our bare, homogeneous reactor analysis can be used to perform
preliminary analytical estimations of the core geometry and composition that will
yield a critical reactor. We know for a critical reactor that the buckling terms:
2 = 𝐵2
𝐵𝑚 𝑔

• 𝐵𝑔2 is typically known from dimensions which are based on thermal limits
2 is determined based on fuel composition and anticipated operation time
• 𝐵𝑚
• 𝐵𝑔2 is uniquely determined from dimensions, but the revere is not always true
• With spheres and cubes, there is only 1 length (𝑅 or 𝑎) that satisfies:
𝜋 2 𝜋 2
Sphere: 𝐵𝑔2 = and Cube: 𝐵𝑔2 = 3
𝑅෨ 𝑎෤
Which makes it easy to determine their critical dimensions (𝑅 or 𝑎) and volumes.
 Size and Composition Effects on 𝑘𝑒𝑓𝑓
The minimum critical volume for a bare reactor of a given composition depends
on leakage and hence the surface-to-volume ratio:
Geometry Sphere Finite Cylinder Cube
Minimum 129.88 148.31 161.11
∀𝑚𝑖𝑛 ≈ 3 ∀𝑚𝑖𝑛 ≈ 3 ∀𝑚𝑖𝑛 ≈ 3
Critical Volume 𝐵𝑚 𝐵𝑚 𝐵𝑚
• Spheres have the smallest surface-to-volume ratio of all the considered
geometries; hence the smallest ∀𝑐𝑟𝑖𝑡 for a reactor of a given composition:
𝑠𝑝ℎ𝑒𝑟𝑒 𝑐𝑦𝑙
∀𝑚𝑖𝑛 < ∀𝑚𝑖𝑛 < ∀𝑐𝑢𝑏𝑒
𝑚𝑖𝑛
• Reactor composition is important too, allowing for even smaller critical masses:
 Optimization for a Finite Cylinder
Now, consider a bare cylindrical reactor with extrapolated dimensions 𝑅෨ and 𝐻: ෩
2 2
2.405 𝜋
𝐵𝑔2 = +
𝑅෨ ෩
𝐻
• There are infinite combinations of 𝑅, ෨ 𝐻 ෩ that can yield the required 𝐵𝑔2 value
• Thus, we can do optimization to figure out what combination minimizes volume
2 = 𝐵2
1) Write the criticality condition: 𝐵𝑚 𝑔
2) Use the definition of 𝐵𝑔2 to write one dimension in terms of the other and 𝐵𝑔2
3) Plug the result from (2) into the volume formula for a cylinder ∀= 𝜋𝑅෨ 2 𝐻 so
that the volume is expressed in terms of 1 independent variable (𝑅෨ or 𝐻) ෩
4) First-derivative test on ∀ to determine what critical dimension minimizes ∀
5) Plug this dimension back into the result from (2) to find the minimum volume
6) You can also use Lagrange multipliers (covered in Dr. Tsvetkov’s notes)
 One-Dimensional, Two-Region Reactor
• So far we have focused on bare, homogeneous reactors.
• Next, we consider reactors with spatially varying properties, and reflectors.
• All real reactors are multi-region systems
• For example, a reactor core will contain fuel, cladding,
structural materials (e.g. steel), control rods, shielding,
coolant, electronics (e.g. in-core detectors), and more!
• We can safely assume that within each region:
Material properties are uniform
• But we still allow them to vary between regions
Extremely simplified diagram of a
Then, the approach to treating a multi-region system is to: nuclear reactor core; illustrating a
1) Solve the 1GD equation in each region multi-region system.

2) Match the solutions in each region using interface conditions

 Multi-Region Reactor Model
Within each region ∀(1) and ∀(2) we will
have region-specific material properties,
denoted with superscripts:
(1) (1) (1)
Region 1: Σ𝑓 , Σ𝑎 , 𝐷(1) , 𝐿(1) , 𝜈𝑓 , …
(2) (2) (2) (2) (2)
Region 2: Σ𝑓 , Σ𝑎 , 𝐷 , 𝐿 , 𝜈𝑓 , …
• Each region will have its own boundary
conditions (vacuum or Marshak).
• 𝜕∀𝑒𝑥𝑡 is the extrapolated boundary.
• There is now an interface surface 𝑆𝑖𝑛𝑡
between each region.
• The interface location is given by 𝑟Ԧ𝑖𝑛𝑡 .
• 𝑒Ƹ𝑛 is the unit normal at the interface.
 Diffusion Theory Conditions
The time-dependent 1GD equation can be expressed for the region (i) as:
1 𝜕𝜙 (𝑖) (𝑟,
Ԧ 𝑡)
− 𝐷∇2 𝜙 𝑖 𝑟,
Ԧ 𝑡 + Σ𝑎 𝜙 𝑖 𝑟,
Ԧ 𝑡 = 𝜈𝑓 Σ𝑓 𝜙 𝑖 𝑟,Ԧ𝑡 , ∀𝑟Ԧ ∈ ∀(𝑖)
𝑣 𝜕𝑡
Boundary Conditions: 𝜙 𝑟Ԧ𝑒𝑥𝑡 , 𝑡 = 0 vacuum
Initial Conditions: 𝜙 𝑟Ԧ , 0 = 𝜙0 𝑟Ԧ
Interface Conditions – to match solutions at the interface we must ensure:
1. Continuity of the scalar flux: 𝜙 1 𝑟Ԧ𝑖𝑛𝑡 , 𝑡 = 𝜙 (2) 𝑟Ԧ𝑖𝑛𝑡 , 𝑡
2. Continuity of the normal component of current:
1 2
𝑒Ƹ𝑛 𝑟Ԧ𝑖𝑛𝑡 ∙ 𝐽 𝑟Ԧ𝑖𝑛𝑡 = 𝑒Ƹ𝑛 𝑟Ԧ𝑖𝑛𝑡 ∙ 𝐽 𝑟Ԧ𝑖𝑛𝑡
Using Ficks Law: −𝐷 1 [𝑒Ƹ𝑛 𝑟Ԧ𝑖𝑛𝑡 ⋅ ∇𝜙 (1) (𝑟Ԧ𝑖𝑛𝑡 , 𝑡)] = −𝐷 2 [𝑒Ƹ𝑛 𝑟Ԧ𝑖𝑛𝑡 ⋅ ∇𝜙 (2) (𝑟Ԧ𝑖𝑛𝑡 , 𝑡)]
Physical Realizability: 0 ≤ 𝜙 𝑟,
Ԧ 𝑡 <∞
 Example: 1GD for a 2-Region Reactor
Consider the concentric spherical reactor: “Infinite” Core
Reflector (1)
1 core region (1) and an infinite reflector (2).
(2)
Radius of the central region = 𝑅𝑖𝑛𝑡
Let’s develop the steady 1GD problem.
𝑂
Start with the Core Region (1):
𝑅𝑖𝑛𝑡
1 1 1
𝜈𝑓 Σ𝑓 −Σ𝑎
∇2 𝜙 (1) 𝑟 + 𝜙 1
𝑟 = 0, 0 < 𝑟 < 𝑅𝑖𝑛𝑡
𝐷1
2 (1)
∇2 𝜙 (1) 𝑟 + 𝐵𝑚 𝜙1 𝑟 =0
(2)
−Σ𝑎
For the Reflector (2): ∇2 𝜙 (2) 𝑟 + 𝜙 2
𝑟 = 0, 𝑅𝑖𝑛𝑡 < 𝑟 < ∞
𝐷2
2 (2)
∇2 𝜙 (2) 𝑟 − 𝐵𝑒𝑓𝑓 𝜙 2
𝑟 =0 (we use “𝐵𝑒𝑓𝑓 " in non-producing regions)
Q’s: What does the gradient operator look like in spherical coordinates? Which way does e_n point in this problem?

1GD, 2-Region Problem Setup

We only have 1 interface (at the fuel-reflector
Core (1)
boundary), here we must have:
1 2
Infinite
𝜙 𝑅𝑖𝑛𝑡 = 𝜙 𝑅𝑖𝑛𝑡 Reflector (2)
1
𝑑𝜙 1 2
𝑑𝜙 2
−𝐷 ቚ = −𝐷 ቚ
𝑑𝑟 𝑟=𝑅𝑖𝑛𝑡 𝑑𝑟 𝑟=𝑅𝑖𝑛𝑡 𝑂
𝑅𝑖𝑛𝑡
We have no “boundary” condition for an
infinite reflector, but our physical condition
will confine the flux throughout the domain:
0 ≤ 𝜙 𝑟 < ∞, ∀𝑟 ≥ 0
We have now given the full problem statement (all governing equations and
required conditions). Our next step is to solve each equation separately.
 How Do We Solve This System?
We are trying to solve the following system of ODE’s:
2 (1) 2 (1) 1 (1)
∇ 𝜙 𝑟 + 𝐵𝑚 𝜙 𝑟 = 0, 0 < r < R int
2 2 1
2 (2)
∇ 𝜙 𝑟 − 𝐵𝑒𝑓𝑓 𝜙 2
𝑟 = 0, 𝑅𝑖𝑛𝑡 < 𝑟 < 𝑅෨
2 1 𝑑 1 𝑑
For spherical geometries, the Laplacian is: ∇ = 2
𝑟 𝑑𝑟 𝑟 2 𝑑𝑟
So our equations can be re-written as:
1 𝑑 1 𝑑 (1) 2 (1) 𝜙 1
𝜙 𝑟 + 𝐵𝑚 𝑟 = 0, 0 < 𝑟 < 𝑅𝑖𝑛𝑡
𝑟 2 𝑑𝑟 𝑟 2 𝑑𝑟
1 𝑑 1 𝑑 (2) 2 2 2
2 2
𝜙 𝑟 − 𝐵𝑒𝑓𝑓 𝜙 𝑟 = 0, R int < 𝑟 < ∞
𝑟 𝑑𝑟 𝑟 𝑑𝑟
Do we recognize the general solutions to these differential equations?
 General Forms of Common ODEs
• In engineering, we run across lots of ODE’s, many of them belonging to the same “family”.
• If an ODE is in a form that can’t be “easily” solved (e.g. separation of variables / integrating
factors) the approach is to usually just look up the solution form.
• But it is good to be familiar with the general forms of some common ODE’s:
Geometry ODE Form: 𝛻 2 𝑓 ± 𝐵2 𝑓 = 0 General Solution Form
𝑑2 𝑓 2𝑓 = 0
𝑓 + = 𝐶1 sin 𝐵𝑥 + 𝐶2 cos 𝐵𝑥
1D Slab ± 𝐵
𝑑𝑥 2 𝑓 − = 𝐶1 ex𝑝 𝐵𝑥 + 𝐶2 exp(−𝐵𝑥)
+
𝐶1 𝐶2
1 𝑑 𝑑𝑓 𝑓 = sin 𝐵𝑟 + cos −𝐵𝑥
1D Spherical 𝑟 2 ± 𝐵 2𝑓 = 0 𝑟 𝑟
𝑟 2 𝑑𝑟 𝑑𝑟 −
𝐶1 𝐶2
𝑓 = ex𝑝 𝐵𝑟 + exp(−𝐵𝑟)
𝑟 𝑟
1 𝑑 𝑑𝑓 𝑓 + = 𝐶1 𝐽0 𝐵𝑟 + 𝐶2 𝑌0 𝐵𝑟
1D Cylindrical 𝑟 ± 𝐵2 𝑓 = 0
𝑟 𝑑𝑟 𝑑𝑟 𝑓 − = 𝐶1 𝐾0 𝐵𝑟 + 𝐶2 𝐼0 (𝐵𝑟)
Where the solution form (𝑓 ± ) depends on the sign of 𝐵2 ; 𝐽0 and 𝑌0 are ordinary Bessel Functions of the 1st and
2nd kinds; 𝐾0 and 𝐼0 are the modified Bessel functions; these functions have a known form, please look em’ up!
 General Solutions to Our Problem
Now we can (hopefully) see that the general solutions to regions (1) and (2) are:
1
𝐶1 1 𝐶2 2
𝜙 𝑟 = sin 𝐵𝑚 𝑟 + cos 𝐵𝑚 𝑟 , 0 0 ≤ 𝑟 ≤ 𝑅𝑖𝑛𝑡
𝑟 𝑟
2
𝐶3 2 𝐶4 2
𝜙 𝑟 = exp 𝐵𝑒𝑓𝑓 𝑟 + exp −𝐵𝑒𝑓𝑓 𝑟 , 𝑅𝑖𝑛𝑡 ≤ 𝑟 < ∞
𝑟 0 𝑟
cos 𝑟
Core: Recall that as 𝑟 → 0, → ∞, so for physical realizability: 𝐶2 = 0
𝑟
Reflector: As 𝑟 → ∞, the flux cannot → ∞, so we must have : 𝐶3 = 0
• We have 2 remaining constants (𝐶1 and 𝐶4 )
• We have 2 conditions left to apply (our interface conditions) to find 𝐶1 and 𝐶4 .
• Generally, for each interface, you will solve for 2 constants using 2 conditions.
• So a 3-region system would have 2 interfaces, therefore 4 interface conditions,
and therefore you would need to solve for 4 𝐶’s.
 Applying Interface Conditions
We have simplified our solutions for each region and found:
1
𝐶1 1 2
𝐶4 2
𝜙 𝑟 = sin 𝐵𝑚 𝑟 and 𝜙 𝑟 = exp −𝐵𝑒𝑓𝑓 𝑟
𝑟 𝑟
Next, we apply interface conditions to force these solutions to agree at 𝑟 = 𝑅𝑖𝑛𝑡
Continuity of the Scalar Flux:
1 2
𝜙 𝑅𝑖𝑛𝑡 = 𝜙 𝑅𝑖𝑛𝑡
𝐶1 1 𝐶4 2
sin 𝐵𝑚 𝑅𝑖𝑛𝑡 = exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡
𝑅𝑖𝑛𝑡 𝑅𝑖𝑛𝑡
1 2
C1 sin 𝐵𝑚 𝑅𝑖𝑛𝑡 − 𝐶4 exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 = 0 (𝐸𝑞. 1)
 Applying Interface Conditions
1
𝐶1 1 2
𝐶4 2
𝜙 𝑟 = sin 𝐵𝑚 𝑟 and 𝜙 𝑟 = exp −𝐵𝑒𝑓𝑓 𝑟
𝑟 𝑟
We have one last interface condition - continuity of the normal component of
1 2
1 𝑑𝜙 ȁ𝑟=𝑅𝑖𝑛𝑡 = −𝐷 2 𝑑𝜙 ȁ𝑟=𝑅𝑖𝑛𝑡
the current: −𝐷
𝑑𝑟 𝑑𝑟
(1) 1 1
1 𝐵𝑚 𝑅𝑖𝑛𝑡 cos 𝐵𝑚 𝑅𝑖𝑛𝑡 −sin 𝐵𝑚 𝑅𝑖𝑛𝑡
LHS: −𝐷 1 𝑑𝜙 ȁ𝑟=𝑅𝑖𝑛𝑡 = −𝐷 1 𝐶1
𝑑𝑟 𝑅𝑖𝑛𝑡 2

2 2
2 exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 +1
2 𝑑𝜙 ȁ𝑟=𝑅𝑖𝑛𝑡 = 𝐷 2
RHS: −𝐷 𝐶4 1 2
𝑑𝑟 𝑅𝑖𝑛𝑡
2 2 2 1 (1) 1 1
𝐷 𝐶4 exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 + 1 + 𝐷 𝐶1 𝐵𝑚 𝑅𝑖𝑛𝑡 cos 𝐵𝑚 𝑅𝑖𝑛𝑡 − sin 𝐵𝑚 𝑅𝑖𝑛𝑡 =0
(𝐸𝑞. 2)
 Solving the System of Equations
1 2
So far we have found: C1 sin 𝐵𝑚 𝑅𝑖𝑛𝑡 − 𝐶4 exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 = 0 (1)

2 2 2 1 (1) 1 1
𝐷 𝐶4 exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 + 1 + 𝐷 𝐶1 𝐵𝑚 𝑅𝑖𝑛𝑡 cos 𝐵𝑚 𝑅𝑖𝑛𝑡 − sin 𝐵𝑚 𝑅𝑖𝑛𝑡 = 0 (2)

• Eqs. (1) and (2) are homogeneous linear equations in the unknowns 𝐶1 and 𝐶2
• All other terms in each equation are known (material properties or dimensions)
• We can write this system of linear equations as a 2x2 matrix:
1 2
sin 𝐵𝑚 𝑅𝑖𝑛𝑡 − exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 𝐶1 0
=
𝐷 1 (1) 1 1
𝐶1 𝐵𝑚 𝑅𝑖𝑛𝑡 cos 𝐵𝑚 𝑅𝑖𝑛𝑡 − sin 𝐵𝑚 𝑅𝑖𝑛𝑡 𝐷 2 2
exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡
2
𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 + 1 𝐶4 0

• We can represent this symbolically as: 𝐴2x2 𝐶Ԧ2 = 02

• For the system to have a non-trivial solution, the determinant of 𝐴2x2 must be 0:
det(𝐴2x2 ) = 0
 Obtaining the Criticality Equation
The critical determinant, det(𝐴2x2 ) is relatively easy to compute by hand:
1 2
sin 𝐵𝑚 𝑅𝑖𝑛𝑡 − exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡
det (1) 1 1 2 2
=0
𝐷 1 𝐶1 𝐵𝑚 𝑅𝑖𝑛𝑡 cos 𝐵𝑚 𝑅𝑖𝑛𝑡 − sin 𝐵𝑚 𝑅𝑖𝑛𝑡 𝐷 2 exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 +1

2 2 2 1
𝐷 exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 + 1 sin 𝐵𝑚 𝑅𝑖𝑛𝑡 +
1 (1) 1 1 2
𝐷 𝐶1 𝐵𝑚 𝑅𝑖𝑛𝑡 cos 𝐵𝑚 𝑅𝑖𝑛𝑡 − sin 𝐵𝑚 𝑅𝑖𝑛𝑡 exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 = 0

1 2
Dividing both sides by sin 𝐵𝑚 𝑅𝑖𝑛𝑡 exp −𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 and rearranging the expression:
1 1 𝐷2 2
𝐵𝑚 𝑅𝑖𝑛𝑡 𝐶𝑜𝑡 𝐵𝑚 𝑅𝑖𝑛𝑡 − 1 = − 1 𝐵𝑒𝑓𝑓 𝑅𝑖𝑛𝑡 + 1 (The Criticality Eq. for this Reactor)
𝐷
• LHS = rapidly varying function of reactor material properties and dimensions
• RHS = slowly varying function of these parameters
 An Aside, Reviewing Decay Chains
We can use conservation to write ODEs describing the time-evolution of several
nuclide populations in a decay chain. Consider, by way of example:
λ𝑋 λ𝑌
𝑋 𝑌 𝑍, 𝜆𝑖 = decay constant of the ith nuclide units: s −1
If we also allow a production rate, 𝑅𝑖 (atoms/sec) for each type-i nuclide the
system of ODEs describing the number of nuclides of each type present are:
𝑑𝑁𝑋
= 𝑅𝑋 − 𝜆𝑋 𝑁𝑋 , Nx 0 = NX0
𝑑𝑡
𝑑𝑁𝑌
= 𝑅𝑌 + 𝜆𝑋 𝑁𝑋 − 𝜆𝑌 𝑁𝑌 , 𝑁𝑌 0 = 𝑁𝑌0
𝑑𝑡
𝑑𝑁𝑍
= 𝑅𝑍 + 𝜆𝑌 𝑁𝑌 , 𝑁𝑍 0 = 𝑁𝑍0
𝑑𝑡
• Each ODE is conservation equation: “change rate = gain rates - loss rates”.
• Radioactive decay rate of a nuclide of type 𝑖 is 𝜆𝑖 𝑁𝑖 (atoms/sec).
Thank you for your time! Any questions?