2015 International Nuclear Atlantic Conference - INAC 2015
São Paulo, SP, Brazil, October 4-9, 2015
ASSOCIAÇÃO BRASILEIRA DE ENERGIA NUCLEAR - ABEN
ISBN: 978-85-99141-06-9
TWO-DIMENSIONAL THERMAL ANALYSIS OF A FUEL ROD BY
FINITE VOLUME METHOD
Rhayanne Y. N. Costa1, Mário A. B. da Silva1 and Carlos A. B. de O. Lira1
1
Departamento de Energia Nuclear
Universidade Federal de Pernambuco
Rua Professor Luiz Freire, 1000
50740-545 Recife, PE
[email protected] [email protected] [email protected] ABSTRACT
In a nuclear reactor, the amount of power generation is limited by thermal and physic limitations rather than by
nuclear parameters. The operation of a reactor core, considering the best heat removal system, must take into
account the fact that the temperatures of fuel and cladding shall not exceed safety limits anywhere in the core. If
such considerations are not considered, damages in the fuel element may release huge quantities of radioactive
materials in the coolant or even core meltdown.
Thermal analyses for fuel rods are often accomplished by considering one-dimensional heat diffusion
equation. The aim of this study is to develop the first paper to verify the temperature distribution for a two-
dimensional heat transfer problem in an advanced reactor. The methodology is based on the Finite Volume
Method (FVM), which considers a balance for the property of interest. The validation for such methodology is
made by comparing numerical and analytical solutions. For the two-dimensional analysis, the results indicate
that the temperature profile agree with expected physical considerations, providing quantitative information for
the development of advanced reactors.
1. INTRODUCTION
The power generation in any thermal facility depends on the amount of vapor generated for
the turbines, which is related to the energy generated by the energy source. In a nuclear
facility, the output energy of a reactor is limited not on the fuel capacity but on the materials'
limits. To avoid radioactive leakage, the temperature limits must be always respected thus, it's
important the constant knowledge of thermal distribution as precisely as possible.
The equations describing nuclear reactions and thermohydraulics are too complex to be
solved analytically, being necessary the adoption of a numerical method for obtainment of a
solution for this problem. The method adopted for this study is the Finite Volume Method
(FVM), expected to provide better results than other methods giving that FVM have a
conservative approach, which may avoid discrepancies or discontinuities especially on the
interface superficies [1].
Finally, this is the first work developed with the adoption of FVM method. Although Finite
(or Control) Volume Method is an important and robust numerical method, its complexity and
the tradition of other methods made it underappreciated until now. It's expected that the
results obtained to be better (minor errors) and physically more representative due to its
conservative principles. Consequently, the possibilities for the continuity of this work are
great, improving the analysis, adding complexity levels and finally reaching a more realistic
representation of those systems.
2. FINITE VOLUME METHOD
The Transport Equation (Equation 1) describes a system over a conservative analysis
presenting flows within a volume. The terms represent, in order, the rate of variation in
property ϕ, the convective effect, the diffusive term and the source term. In Equation 1, is
the density, is the property under study, is the velocity vector, is the diffusion
coefficient and is the source term.
(1)
The application of Finite Volume Method starts with the discretization of total volume into
control volumes that can be analyzed individually applying the major equation as Equation 1
into each control volume, which is affected by their immediately neighbors and the
generation source.
In order to obtain the temperature distribution of a one-dimensional and steady system, under
only diffusive effects, Equation 1 can be rewrote as Equation 2.
(2)
To solve a system such as the one represented by Figure 1 (one-dimension rod with
extremities’ temperatures known with volumetric source strength constant), Equation 2 may
be applied allowing mathematics development until Equation 3. For a diffusion coefficient
equals to the conductivity of a material, Equation 3 is the basic equation for a pure
conductive system described by Incropera [2].
Figure 1. One-dimensional system under diffusion with extremities’ temperatures
known [1].
(3)
INAC 2015, São Paulo, SP, Brazil.
Equation 3 gives the numerical solution for the system represented by Fig. 1, with S being the
area of heat conduction and T being the temperature and e and w indicate the east and west
neighbors, respectively.
The solution for this system is given by Equation 4 with its coefficients (Asubscript) presented
in Table 1.
(4)
Table 1. Coefficients for FVM for one-dimensional steady system, from Figure 1.
Extremidade 1
Pontos
Intermediários
Extremidades 2
3. VALIDATION
The system proposed is a typical nuclear reactor fuel rod, a cylindrical rod. The validation for
this paper is the application of the Control Volume Method into a simpler system, which can
also be solved analytically to be compared. The system for validation is a fuel bar with
perfectly connected clad under steady state and pure diffusion processes and no cooling fluid,
but known clad's external wall temperature, that can be represent through Figure 2. Equation
1 was rewrite for cylindrical coordination and solved both analytically and numerically and
the results were compared.
Figure 2. One-dimension cylindrical discretized control volumes [3].
INAC 2015, São Paulo, SP, Brazil.
3.1. Objective
The system described in section 3 was studied for two main reasons:
i. guarantee that the method could be applied in thermal systems as a nuclear fuel rod
representing this system appropriately, and;
ii. to verify if the approximation of the volume integration of the source term could be
done by considering the source as a constant.
3.2. Method
The application of FVM into Equation 1 in cylindrical coordinates allowed the solution
through numerical method. The analytical method was also obtained. The solution equations
and their respective limits of application can be found in Table 2, for both analytical and
numerical solutions.
Table 2. Equations for defined limits for analytical/numerical solutions for Validation.
Analytical Numerical
r=0
0<r≤a
r=a
a<r<b
The conditions that allow the determination of the coefficients for analytical solution are the
boundary conditions given:
i. Continuity of temperature at ;
ii. Continuity of fluid flow at ;
iii. Temperature calculated through each possibility (from fuel to interface and from clad
to interface) for should be equal;
iv. Temperature at is known.
Both solutions were found through MatLab and the error between then was calculated
through Equation 5.
(5)
INAC 2015, São Paulo, SP, Brazil.
Table 3 presents the coefficients for analytical solution and Table 4, the coefficients for
numerical solutions.
Table 3. Coefficients for analytical solution for Validation.
C1 0
Table 4. Coefficients for numerical solution for Validation.
r=0 1 0 1
0<r<a
r=a
a<r<b 0
4. TWO-DIMENSIONAL STEADY SYSTEM
The problem studied in this paper is the determination of the temperature in the center of the
control volumes of a fuel bar perfectly connected to the clad and cooled by water flowing
upward. Equation 1 cannot be solved analytically for this analysis, and then it is necessary to
INAC 2015, São Paulo, SP, Brazil.
apply a numerical method such as FVM. Figure 3 represents the system for two-dimensional
thermal analysis, with water entering south the rod in the same temperature (290°C).
Fluid exit FLUID - Equation 8
Extremity north
FLUID - Equation 9
Axial's interior points
Extremity south FLUID - Equation 7
Fluid entrance
Figure 3. Two-dimension cylindrical discretized control volumes.
4.1. Objective
The two-dimension analysis aims three priority points:
i. apply Finite Volume method in a two-dimension analysis for a nuclear fuel rod, as the
first of many future papers in this promising area;
ii. achieve physically logical results for this analysis so it can be compared and analyzed
in the near future, and;
iii. achieve a result congruent to the knowledge in nuclear community.
4.2. Method
The application of FVM into Equation 1 in cylindrical coordinates allows the achievement of
a solution given by Equation 6 (a more complete Equation 4) with coefficients given in
Tables 5 and 6.
(6)
Table 5 gives the coefficients of all points in the north extremity of the rod, represented by
the points indicated through the highest arrow in Figure 3. The same coefficients might be
used to the points in the south extremity of the rod by alternating the south and north
coefficient.
INAC 2015, São Paulo, SP, Brazil.
Table 5. Coefficients for numerical solution for fuel rod's axial extremity north.
r=0 0
0<r<a
r=a
a<r<b 0
r=b 0
Table 6 gives the coefficients of all points inside the fuel rod the north extremity of the rod,
represented by the points indicated through the highest arrow in Figure 3.
INAC 2015, São Paulo, SP, Brazil.
Table 6. Coefficients for numerical solution for fuel rod's axial interior points.
r=0 0
0<r<a
r=a
a<r<b 0
r=b 0
The fluid temperature can be calculated through Equations 7, 8, and 9, gathered in Table 7.
Figure 3 indicates when each of those Equations is used: Equation 7 gives the temperature of
the point immediately north the entrance point, neighbor to the south extremity points.
Equation 8 gives the temperature of the exit point of the fluid and Equation 9 gives the
temperatures of fluid within those points. In Equations 7 to 9, the sum factor indicates all the
points south than the point to be determined except the entrance temperature, which is added
out of the sum sign.
Table 7. Equations to define fluid temperature in two-dimensional system analysis.
(6)
(7)
(8)
INAC 2015, São Paulo, SP, Brazil.
The temperature of the fluid entering and exiting the limits of the control volumes in the rod
were calculated by convection, but the heat used to increase the temperature of such points
was given due to convection from the near wall fluid and the mean temperature entering and
exiting. Once the axial dimension was divided in many points, the approximation used was
that at mean point the temperature was the mean between the entrance and exit ones. That
approximation should be good enough for a large number of points.
5. RESULTS
To define values, a few properties were took from El-Wakil [4] and gathered at Table 8.
Those were the entrance constants at MatLab software to display results.
Table 8. Physical properties and operational constants for MatLab entrance.
PHYSICAL PROPERTIES
Fuel rod's length L (m) 3.7
Fuel’s thermal conductivity - (W/m·K) 2.08
Clad’s thermal conductivity - (W/m·K) 13.8
Fuel radius - a (m) 0.0045
Clad thickness - b (m) 0.0006
Volumetric source strength - (W/m3) 5·108
Temperature of clad’s wall - Tb (°C) 400
Thermal diffusivity - D (cm) 0.16
Macroscopic absorption cross-section - (cm-1) 0.0197
Number of nodes in fuel - N 40
Number of nodes in clad – N1 10
Number of nodes axial - NZ 101
Specific heat of water – cp (J/kg·°C) 5986
Convection coeficiente – h (W/m2·°C) 17400
Fluid entrance temperature (°C) 290
Mass flow (kg/s) 0.39
5.1. Validation
The one-dimensional case was solved by both analytical and numerical methods and the
solution was plotted into the graphic in Figure 4, with aid of MatLab Software. Also, the
maximum percentage error was evaluated through Equation 5.
INAC 2015, São Paulo, SP, Brazil.
Figure 4. Temperature distribution for one-dimension analysis for a fuel rod, both
analytically and numerically.
The graphic shows a very good agreement between the two solutions and the maximum
percentage error found was , a very little one.
5.2. Two-Dimensional Analysis
The temperature was calculated for the two-dimensional case with the equations and
coefficients compiled with MatLab. For one-dimensional systems, it's an easily solved system
by inverting a tri-diagonal matrix. Once two-dimension systems result in penta-diagonal
matrix, a solution is to suppose the values of two temperatures, such as north and south, and
use those variables as entrance data, making the independent coefficient compound by
. That will allow the obtainment of a tri-diagonal matrix easily solved.
Now, it's necessary to guarantee that it converges and that the software can solve a system
like this in acceptable time.
Figure 5 presents a physically possible temperature distribution with the highest temperature
situated in the center of the rod (r = 0 and z = 0) being approximately 1050°C, and with a
little deflection of the maximum temperatures to north from axial middle point.
INAC 2015, São Paulo, SP, Brazil.
Figure 5. Temperature distribution obtained through two-dimensional analysis with
FVM for a nuclear fuel rod.
Figure 6 presents the axial temperature distribution of clad outer wall and cooling fluid, with
profile in accordance to well established literature [4]. The high variation shown in clad's
temperature is something to be studied in the future.
Figure 6. Clad outer wall and cooling fluid temperature profile for a two-dimensional
finite volume analysis.
INAC 2015, São Paulo, SP, Brazil.
6. CONCLUSIONS
This work intended to calculate the two-dimensional temperature distribution of a nuclear
fuel rod cooled by water flowing upward and it was successfully concluded. The validation
presents a near-zero error, very satisfactory, showing that the model can be adopted for the
proposed system and the approximation for the integration of the source term integrated in
volume as the source term times the control volume is also acceptable. The two-dimensional
analysis presents a temperature profile as expected, with higher temperatures at center. The
second graphic presents the expected profile for both fluid and outer clad wall.
It's necessary, though, to evaluate if the constants adopted are physically consistent, since a
higher convection coefficient would reduce the maximum temperature in the outer wall of the
clad and the dislocation of the maximum temperature from the axial middle due to fluid
would be more evident, for example.
Besides that, for future works, a fluid stress can be exploited, a transient scenario could be
studied and a more real system can be described, considering the gap region between fuel and
clad or a more realistic and detailed configuration of fluid and fuel rod.
ACKNOWLEDGMENTS
The authors thank the partial support of CNPq (Conselho Nacional de Desenvolvimento
Científico e Tecnológico) and of INCT (Instituto Nuclear de Ciência e Tecnologia de
Reatores Nucleares Inovadores).
REFERENCES
1. Veersteg, H K, Malalasekera, W, An Introduction to Computational Fluid Dynamics - The
Finite Volume Method, Pearson Education Limited, Harlow, England (2007).
2. Incropera, F. P., Dewitt, D. P., Bergman, T. L., & Lavine, A. S. Fundamentos se
Transferência de Calor e de Massa, LTC, Rio de Janeiro (2008).
3. Silva, M A B, Narain, R, "A quantitative estimate on the heat transfer in cylindrical fuel
rods to account for flux depression", Progress in Nuclear Energy, Volume 60, pp. 29-34
(2013).
4. El-Wakil, M M, Nuclear Heat Transport, International Textbook Company, Scranton,
United States (1971).
INAC 2015, São Paulo, SP, Brazil.
