Professional Documents
Culture Documents
Effect of power variations across a fuel bundle and within a fuel element
on fuel centerline temperature in PHWR bundles in uncrept and crept
pressure tubes
E.N. Onder a,⇑, D. Roubtsov a, Y.F. Rao a, B. Wilhelm b
a
Canadian Nuclear Laboratories, Chalk River, Ontario K0J 1J0, Canada
b
University of Guelph, Guelph, Ontario N1G 2W1, Canada
a r t i c l e i n f o a b s t r a c t
Article history: The neutron flux and fission power profiles through a fuel bundle and across a fuel element are important
Received 9 November 2016 aspects of nuclear fuel analysis in multi-scale/multi-physics modelling of Pressurized Heavy Water
Accepted 15 December 2016 Reactors (PHWRs) with advanced fuel bundles. Fuel channels in many existing PHWRs are horizontal.
Available online 5 January 2017
With ageing, pressure tubes creep and fuel bundles in these pressure tubes are eccentrically located,
which results in an asymmetric coolant flow distribution between the top and bottom of the fuel bundles.
The diametral change of the pressure tube due to creep is not constant along the fuel channel; it reaches a
maximum in the vicinity of the maximum neutron flux location. The cross-sectional asymmetric posi-
tioning of fuel bundles in a crept pressure tube contributes to an asymmetric power distribution within
a ring of fuel elements. Modern reactor physics lattice codes (such as WIMS-AECL1) are capable of predict-
ing the details of power distribution from basic principles. Thermalhydraulics subchannel codes (such as
ASSERT-PV) use models to describe inhomogeneous power distribution within and across fuel elements
(e.g., flux tilt model, different powers in different ring elements, or radial power profiles). In this work, phy-
sics and thermalhydraulics codes are applied to quantify the effect of eccentricity of a fuel bundle on power
variations across it and within a fuel element, and ultimately on the fuel temperature distribution and fuel
centerline temperature, which is one of the indicators of fuel performance under normal operating condi-
tions (NOC). Due to fissioning, Xe and Kr gases are produced, which contribute to the internal gas pressure
within an element. The increase in gas pressure may jeopardize the fuel element integrity (sheath rupture).
On the other hand, fission gas is produced as a function of neutron flux/power and diffuses to the grain
boundaries and fuel-pellet interface as a function of fuel temperature. Therefore, it is important to know
the fuel temperatures and distributions within the fuel.
Ó 2017 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.anucene.2016.12.016
0306-4549/Ó 2017 Elsevier Ltd. All rights reserved.
246 E.N. Onder et al. / Annals of Nuclear Energy 102 (2017) 245–254
candidates for advanced fuels that can be employed in an 3. Neutron physics and thermalhydraulics codes
advanced PHWR due to lower linear element ratings in outer ring
elements at comparable bundle power to a 37-element bundle, 3.1. WIMSAECL physics code
resulting in lower fuel temperatures and lower Fission Gas
Releases (FGR). WIMS-AECL (Altiparmakov, 2008) is a two-dimensional lattice
In all previous analyses (Bae and Park, 2011), fuel temperatures physics code that uses multi-group integral transport theory and
and temperature distributions within a fuel element in the same collision probability methods to solve for the neutron flux distribu-
ring are assumed to be identical. However, with the changes in ele- tion and reaction rates within a given lattice cell design. The code is
ment power in the same ring (e.g., top and bottom elements in the used with its 89-group nuclear data library based on ENDF/B-VII.0
outer ring) due ageing of nuclear reactors (i.e., channels in PHWR), evaluation. WIMS-AECL is capable of performing fuel depletion/
this assumption does not hold true. Consequently, the aim of this burn-up calculations in the reactor lattice and it calculates the
study is to quantify the element power differences in the same power density in the fuel pins at each burn-up step. The current
ring, which are always assumed to be same, and fuel (centerline) release version of WIMS-AECL is WIMS-AECL 3.1 (Altiparmakov,
temperatures and temperature distributions in the same ring in 2008). The WIMS-AECL neutron transport code has become a stan-
an aged reactor (such as fuel channel type reactor). Given that fis- dard tool for analyzing the lattice physics characteristics of Cana-
sion gas diffusion and hence FGR is a function of temperature, dian PHWR-type fuel bundles (Altiparmakov et al., 2010) and is
some differences in FGR can be expected in elements exposed to used in this study to calculate the element power distribution in
different powers in the same ring. the rings of a 43-element fuel bundle.
The objective of this study is to investigate the impact of non-
uniform power distributions across a fuel bundle and within a fuel
element on fuel centerline temperatures in uncrept and crept pres- 3.2. ASSERTPV thermalhydraulics subchannel code
sure tubes. Ultimately, the novelty of this study is the quantifica-
tion of important parameters, such as variations of element The ASSERT-PV code has been developed at CNL/AECL for sub-
powers and fuel centerline temperatures in the same ring; rather channel (SC) thermalhydraulics analysis of Canadian fuel bundles
than qualitatively indicating that power and fuel temperature vari- (Rao et al., 2014), (Rao and Hammouda, 2003). The current release
ations in the same ring due to ageing are not expected to be version of ASSERTPV is ASSERT-PV 3.2 (Rao et al., 2014; Rao and
significant. Hammouda, 2003) which was used to predict fuel temperatures
in the present analysis. ASSERT-PV meets the specific requirements
2. Background for subchannel thermalhydraulic analysis of single- and two-phase
flows in horizontally oriented Canadian PHWR fuel bundles. The
The existing fleet of Canadian PHWRs are pressure tube type code is generally used as a design assist tool for new designs of
reactors, where pressure tubes are located horizontally in calandria Canadian fuel bundles to predict sheath temperatures at the
tubes (see Fig. 1(a)). Fig. 1(b) shows the cross sectional geometry of post-dryout conditions, pressure drop and dryout power in Cana-
a 43-element bundle with appendages: end-plates, bearing pads, dian PHWR fuel channels (Rao and Leung, 2007) and provides sup-
Critical Heat Flux (CHF) enhancement buttons and spacers. The port for safety and licensing codes. It has been validated/assessed
43-element bundle consists of two types of fuel elements: the against experimental data for the 37-element and 43-element bun-
diameter of center element and inner ring elements is slightly lar- dles. It provides detailed flow and phase distributions in subchan-
ger than the diameter of middle and outer ring elements. nels of a fuel bundle, and predicts channel pressure drop, dryout
Fig. 2(a) shows the reference (uncrept) pressure tube whereas power and CHF location, post-dryout fuel-sheath temperatures,
Fig. 2(b) shows an axially non-uniformly crept pressure tube where fuel temperatures at different radial locations of the fuel pellet.
fuel bundles are eccentrically located. This eccentricity results in ASSERT-PV (Rao et al., 2014; Rao and Hammouda, 2003; Carlucci
an asymmetric flow distribution between the top and bottom of et al. (2004)) is based on ASSERT-IV (Carver et al., 1990), which
the fuel bundles (see the flow by-pass region in Fig. 2(b)), which in turn originated from the COBRA-IV computer program. The
in turn can result in power differences within the same ring two-phase flow model used in ASSERT-PV is based on an advanced
between the top and bottom of a bundle depending on the creep drift-flux model: a five-equation model that can consider thermal
ratio, which is defined as: non-equilibrium and the relative velocity of the liquid and vapour
phases. Thermal non-equilibrium is dealt with by two-fluid energy
CR ¼ ðDcreep Dref Þ=Dref ð1Þ
equations for the liquid and vapour while relative velocity is
where Dcreep is the local inner diameter (ID) in a crept pressure tube obtained from semi-empirical correlations. While retaining the
and Dref is the ID of the reference (uncrept) pressure tube. major features in ASSERT-IV, ASSERT-PV employs a more compre-
From foil activation experiments in nuclear reactors (Serdula hensive and robust numerical solution based on the PressureVeloc-
and Green, 1967) and from reactor physics simulations ity numerical method to enhance the modelling capability under
(Altiparmakov et al., 2010), the thermal neutron flux inside the relatively low flows, in addition to other improvements. The major
bundle in a pressure tube decreases toward the bundle center. As subchannel thermalhydraulics models in ASSERT-PV are described
a result, there is a neutron flux gradient in the fuel elements and in Rao et al. (2014), including the single- and two-phase inter-
a similar fission power distribution in the fuel pellets (see Fig. 3). subchannel turbulent mixing, void drift and diffusion, buoyancy
This effect is known as ‘‘flux tilt” (Tong and Weisman, 1979) and drift, CHF, post-dryout heat transfer, and the effects of bundle
can be modelled as follows: appendages.
The fuel temperature was calculated at different specified radial
powðr; hÞ ¼ q000 ð1 þ e ðr=Rf Þ cos hÞ ð2Þ
locations. The material properties of fuel and sheath are user-
where q000 is the power density, Rf is the fuel pellet radius, r and h are defined values, and for this analysis, the material properties were
local cylindrical coordinates associated with the fuel element, and obtained from IAEA-TECDOC-1496 (IAEA, 2006). In this study, cal-
e > 0 is the tilt parameter. The element power (per unit length) is culations were performed for CANDU fuel at the middle of burnup
given by (~90 MWh/kgU). Provided that the CANDU fuel mid-burnup is low,
an assumption was made to use the fresh fuel thermal
q0 ¼ q000 pR2f ð3Þ conductivity.
E.N. Onder et al. / Annals of Nuclear Energy 102 (2017) 245–254 247
Fig. 1. Canadian Pressurized Water Reactor with a Fuel Channel and a Fuel Bundle.
Fig. 3. Thermal neutron flux in a reactor lattice cell for 43-element bundle (WIMS-AECL 3.1).
248 E.N. Onder et al. / Annals of Nuclear Energy 102 (2017) 245–254
Fig. 4. (a) Direction of Flux Tilt, (b) Degree of Flux Tilt in Elements of Different Rings.
Fig. 6. Relative Power Distribution in Different Rings for Uncrept and 5.1% Crept Channels with Bundle Geometry Showing Locations of Elements in a Fuel Channel.
250 E.N. Onder et al. / Annals of Nuclear Energy 102 (2017) 245–254
Fig. 7. Critical (Hot) Element and Maximum Sheath Temperature Locations in a 43-Element Bundle with: (a) Uniform and Non-Uniform EPD in the Uncrept Channel and
Uniform EPD in the 5.1% Crept Channel, (b) Non-Uniform EPD in the 5.1% Crept Channel. (For interpretation of the references to colour in this figure legend, the reader is
referred to the web version of this article.)
Fig. 8. Coolant Temperatures in ubchannels Surrounding Critical Elements at the Axial Location where Fuel Centerline Temperatures Reach a Maximum.
is facing. The solution requires an iteration to achieve a consis- of the boundary conditions) to calculate the temperature
tency between the heat generation and transfer to the cladding distribution in the (cylindrical) fuel pellet by solving the (steady
and the convective heat transfer to the coolant, which is state) heat conduction equation
expressed using the Newton’s cooling law as
rðkrTÞ þ q000 ¼ 0 ð7Þ
~
q00 ¼ hðT w T co Þ~
n ð6Þ
Despite that higher power (or heat flux) is imposed to the fuel
where ~q00 is the heat flux, T co is the coolant temperature, h is the heat segments facing the outer subchannels (to account correctly for
transfer coefficient, (predicted using a Dittus–Boelter3 type equa- the effect of flux tilt that follows the neutron flux in heavy water
tion) and ~n is the unit vector perpendicular to the surface. The sheath pressure tube type reactors), slightly higher sheath temperatures
outer surface temperature, Tw, is estimated and later used to calcu- are observed in the segments facing the inner subchannels. This
late fuel pellet surface temperature, which in turn is used (as one is due to higher coolant temperatures in the inner smaller sub-
channels (as indicated in Fig. 7, relatively hotter coolants and
3
Nu = f(Re, Pr) = hD=kc ; where Nu is the Nusselt number, Re is the Reynolds
sheath temperatures are in subchannels 32/33/34 and 36/37/38
number, Pr is the Prandtl number, h is the heat transfer coefficient, D is the tube (in but not in the subchannel 35 in comparison with larger outer sub-
this case subchannel) diameter, kc is the thermal conductivity of fluid. channels 58/59 and 61/62).
E.N. Onder et al. / Annals of Nuclear Energy 102 (2017) 245–254 251
Fig. 9. Variation of Critical Element Sheath and Cross Section Average Coolant Temperatures along the fuel channel in Uncrept and 5.1% Crept Channels (CC).
Fig. 8 shows coolant temperatures in subchannels surrounding for the uniform EPD, coolant temperatures in the subchannels at
the critical elements at the axial location where the maximum fuel the outlet vary between 305.0 °C and 312.7 °C in the uncrept chan-
center temperatures are reached. Coolant temperatures are higher nel, and between 295.3 °C and 310.3 °C in the 5.1% crept channel.
in the inner subchannels than in the outer subchannels and are Maximum sheath temperatures occur in critical elements facing
higher for the 5.1% crept channel (uniform and non-uniform EPD) inner subchannels (as identified/coloured in amber in Fig. 7 with
than for the uncrept channel (uniform and nonuniform EPD). subchannels 32, 33, 34, 36, 27 and 38) and reach a peak value
The fuel pellet periphery temperatures at the segments facing downstream of the peak power location. Maximum sheath temper-
outer subchannels (i.e., facing subchannels 59, 60 and 61 in ature is reached in the 5.1% crept channel earlier as compared to
Fig. 7) are slightly higher (up to 6 °C) than those at the segments that in the uncrept channel (i.e., at an axial location of 408 cm in
facing inner subchannels. However, coolant temperatures in the the 5.1% channel compared to 448.5 cm in the uncrept channel).
inner subchannels are higher (up to 6 °C) than that in the outer This is due to a higher enthalpy imbalance in the 5.1% crept chan-
subchannels, which in turn results in slightly higher (up to 3 °C) nel. The maximum sheath temperature in the 5.1% crept channel
cladding surface temperatures at the outer-ring element segments (328 °C) is slightly higher than that in the uncrept channel
facing inner subchannels as compared to those at the segments (323 °C). At the outlet, the sheath temperature for the uncrept
facing outer subchannels. channel (314 °C) is slightly higher as compared to that in the
Fig. 9 shows the variation of cross section average coolant and 5.1% crept channel (311 °C), similar to the coolant temperatures.
critical sheath temperatures along the fuel channel for uncrept Fig. 10 shows the axial distribution of fuel centerline tempera-
and 5.1% crept channels (CC) together with the axial power distri- tures in the critical elements4 for elements with linear powers of 4–
bution and pressure tube creep. It should also be noted that there is 5 kW/m at the low power region (i.e., at the inlet) and 2835 kW/m at
no difference in coolant temperature variations along the fuel the high power region. Fuel centerline temperatures reach a maxi-
channel between uniform and non-uniform EPD cases. The differ- mum at the peak power location (323 cm). A fuel centerline tem-
ence in sheath temperatures between uniform and non-uniform perature of 1248 °C for the uniform EPD in the uncrept channel, of
EPD in the uncrept channel is insignificant. Coolant temperatures 1255 °C for the nonuniform EPD in the uncrept channel, of 1260 °C
increase gradually in both fuel channels. However, the average for the uniform EPD in the 5.1% crept channel and of 1302 °C for
coolant temperature at the channel outlet is lower in the 5.1% crept the nonuniform EPD in the 5.1% crept channel is reached in the crit-
channel than in the uncrept channel. This is due to the higher ical elements that are located in the outer ring with linear element
enthalpy imbalance in the crept channel downstream of the critical powers of 35 kW/m at the high power region.
location (i.e., beyond the peak in the axial power distribution and
downstream of the maximum pressure tube diametral creep).
Coolant temperatures in the subchannels of a bundle in the 4
It should be noted that critical elements refer to the elements with peak
uncrept pressure tube are more homogenously distributed: e.g., centerline temperatures, and these elements are located in the outer ring.
252 E.N. Onder et al. / Annals of Nuclear Energy 102 (2017) 245–254
Fig. 10. Variation of Fuel Centerline Temperatures along the fuel channel in Uncrept and 5.1% Crept Channels (CC).
There is no significant difference among uniform EPD (taken as As shown in Fig. 12, large variations in the fuel centerline tem-
the reference case) and non-uniform EPD in the uncrept channel perature are observed for the 5.1% crept channel, in particular for
and uniform EPD in the 5.1% crept channel. An increase of 7.3 °C the outer ring elements: centerline temperatures in outer ring (ele-
(or 0.59%), as compared to the reference case, is observed for the ments #23 through #43) vary around the average centerline tem-
non-uniform EPD in the uncrept channel, and of 11.4 °C (or perature of 1246 °C within ± 56 °C (i.e., 1302 °C at the bottom and
0.91%) for the uniform EPD in the 5.1% crept channel. In the case 1190 °C at the top). Consequently, no significant differences in FGR
of non-uniform EPD in the 5.1% crept channel, however, the tem- are expected between top and bottom elements for such differ-
perature increase is 54.1 °C (or 4.33%) as compared to the reference ences in fuel temperature.
case, and 42.7 °C (or 3.39%) as compared to the case of uniform EPD Similar to the bundle power profile p(i, j) (see Fig. 6), the center-
in the 5.1% crept channel line temperature maps shown in Fig. 12 allow a simple approxi-
It had been expected that the temperature distribution could mate description:
show angular dependency because the flux tilt is imposed to fuel
elements and nonuniform subchannel coolant temperatures occur T c ði; jÞ T c;i þ DT c;i cosðai ð2p=Ni Þ jÞ: ð8Þ
due to enthalpy imbalance and flow (re)distributions between sub-
where Tc,i and DTc,i are the parameters to be defined from the least
channels due to turbulence. Therefore, Fig. 11(a) is plotted for the
square method. For the uncrept pressure tubes (CR = 0), Tc(i, j) Tc,i
variation of fuel temperatures in the fuel pellet at the mid-channel
is a good approximation because, for a given fuel ring, the deviation
location (axial location where the highest fuel centerline tempera-
of Tc(i, j) from a constant (Tc,i) is ±0.1% in the 43-element fuel
tures are observed) as a function of distance from the fuel center in
bundles (e.g., a few °C for Tc 1000 °C). In the case of crept pressure
the segment facing outer subchannels and in the segments facing
tubes, the differences in Tc(i, j) in a given fuel ring can became more
inner subchannels. It should be noted that the temperatures of ele-
noticeable as DTc,i is expected to be increased with CR. For example,
ment 35 in Fig. 11(a) are averaged to be representative of two
for CR = 5.1% and non-uniform EPD, we have
effective segments, the inner and the outer. The inner segment
neighbours the three inner subchannels 36, 37 and 38, whereas DT c;inner =T c;inner 1:5%; DT c;middle =T c;middle
the outer segment neighbours the two outer subchannels 61 and
2:5%; andDT c;outer =T c;outer 4:6%;
62 (see Fig. 11(b)). In segments facing outer subchannels, slightly
higher temperatures are observed as compared to the segments i.e., roughly, DTc,i /Tc,i 1%; see Fig. 12.
facing inner subchannels. Table 1 summarizes temperature distri- However, detailed analyses of the dependence of Tc,i and DTc,i
butions in different segments (facing different subchannels) in on the dimensionless creep CR are left for future studies. On the
the critical elements. other hand, approximating p(i, j) and Tfuel(i, j) in the fuel bundles
Fuel temperature variations within each fuel ring are expected in crept PT with the formula given by
to follow local power variations (see Fig. 6), with the most pro-
nounced effect in a crept channel. Fig. 12 shows the variation of yði; jÞ yi ðCRÞ þ Dyi ðCRÞ cosðai ð2p=Ni Þ jÞ; Dyi ðCRÞ yi ;
fuel centerline temperatures with locations (i.e., top and bottom ð9Þ
and different rings). The largest temperature difference of 112 °C
is observed between top and bottom in the outer ring elements and with the linear scaling of Dyi / CR can be used for practical
with non-uniform EPD in the 5.1% crept channel. The temperature applications, such as PHWR analysis (rule of thumb) linear scaling
difference between top and bottom elements in the middle ring is of Dyi / CR can be useful in practical applications, such as PHWR
50 °C and in the inner ring is 36 °C. The temperature variations analysis (rule of thumb); y(i, j) = (p,Tfuel).
between top and bottom elements in the other configurations The explicit results as obtained here for p(i, j)/pbundle and Tc(i, j)
(i.e., 43-element bundle in the uncrept channel with uniform and at CR = 5.1% can be used as a reference point for the estimates
non-uniform EPD and in the 5.1% crept channel with uniform based on engineering judgment for creep ratio CR 1% as applied
EPD) are lower than 30 °C at any given ring. in nuclear engineering.
E.N. Onder et al. / Annals of Nuclear Energy 102 (2017) 245–254 253
Fig. 11. Variation of Fuel Temperatures with Normalized Radial Distance in Uncrept and 5.1% Crept Channels (CC) (a), and Segmentation of Fuel Element 35 (b).
Table 1
Summary of Temperature Distribution in the Critical Segment of Fuel Pellets for Uniform and Non-Uniform EPD in Uncrept and 5.1% Crept Channels.
Channel Creep (%) Element Power Distribution Critical Element SC # Inner or Outer SC Temperatures (°C)
Fuel Centre Mid-Radius Pellet Surface Sheath Surface
0 Uniform 35 36 Inner 1247.6 948.2 376.3 316.3
37 Inner 1247.4 934.4 373.7 315.5
38 Inner 1247.6 947.4 375.4 315.3
61 Outer 1248.1 989.1 379.9 314.1
62 Outer 1248.1 988.4 379.1 313.3
0 Non-uniform 32 32 Inner 1254.9 952.6 376.5 316.3
33 Inner 1254.7 938.9 374.1 315.7
34 Inner 1254.9 952.3 376 315.7
58 Outer 1255.4 993.7 379.9 313.7
59 Outer 1255.4 993.8 380.2 314.1
5.1 Uniform 35 36 Inner 1259.0 958.2 383.1 323.4
37 Inner 1258.8 944.8 381.4 323.6
38 Inner 1259.0 957.6 382.5 322.7
61 Outer 1259.5 999.3 387.1 321.7
62 Outer 1259.5 998.6 385.9 320.4
5.1 Non-uniform 34 35 Inner 1301.5 976.4 382.9 322.2
36 Inner 1301.6 982.4 385.1 324.0
60 Outer 1302.2 1028.6 389.5 322.1
61 Outer 1302.2 1029.3 390 322.5
Fig. 12. Centerline Temperature Variations between Top and Bottom Elements.
254 E.N. Onder et al. / Annals of Nuclear Energy 102 (2017) 245–254