Annals of Nuclear Energy 87 (2016) 400–410

Contents lists available at ScienceDirect

Annals of Nuclear Energy

journal homepage: www.elsevier.com/locate/anucene

Modeling the thermal stress of heat transfer tubes with tube support
plate gaps in a steam generator
Baozhi Sun ⇑, Jianxin Shi, Liu Yang, Longbin Yang ⇑, Yanjun Li, Shanghua Liu
College of Power and Energy Engineering, Harbin Engineering University, Nantong Street, Harbin 150001, PR China

a r t i c l e i n f o a b s t r a c t

Article history: This work presents a three-dimensional ‘‘unit pipe” model with and without gaps between tubes and
Received 4 April 2015 associated tube support plates in a steam generator. The model is used to simulate the thermal–hydraulic
Received in revised form 11 September 2015 characteristics of the steam generator and obtain the distributions of key parameters of the heat transfer
Accepted 15 September 2015
tubes, the primary and secondary sides. Using this model, we pass fluid calculation results to the struc-
ture through a flow-heat-solid coupling in workbench, conduct thermal stress calculations of the heat
transfer tubes, and analyze the influence of gaps on the system’s thermal–hydraulic characteristics and
Keywords:
resulting thermal stress. The results indicate that there is a rapid change in the secondary side’s velocity
Steam generator
Tube support plates’ gaps
and heat transfer coefficient at the tube support plates, and that the steam quality and fluid velocity in
Flow-heat-solid coupling the gaps is higher and lower, respectively, than in the flood holes. Moreover, the resulting thermal stress
Thermal–hydraulic experienced by the heat transfer tubes shows a periodic fluctuation around their circumference at the
Thermal stress tube support plates, while the thermal stress without gaps at the tube support plates is greater than
the thermal stress seen with gaps, and its circumferential volatility is approximately twice the volatility
with gaps present.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction media model was used by Ferng (2007) and Ferng and Chang
(2008) to investigate the law of two-phase flow and heat transfer
Heat transfer tube failure in a steam generator (SG) is the most of the secondary side. Meanwhile, Li et al. (2013), Sun and Yang
common accident in nuclear power plants that seriously affects (2013) and Yang et al. (2013) used a coupled heat transfer
their integrity and security. Failure is primarily due to Stress Cor- model to study the influence of TSPs on the thermal–hydraulic
rosion Cracking (SCC) for deposition of corrosion products along characteristics in a SG without considering the existence of gaps
with synergistic effect of stress stemming from a number of rea- at the TSPs, finding wall temperature at the contact points was
sons and thermally induced stress is one of them and here it is crit- significantly higher than at the flood holes because the TSPs were
ical to note that difference in temperature across the tube wall in direct contact with the tubes.
thickness leads to the thermal stress, and thus effect of tube sup- Strong thermal stress research has greatly enhanced the
port plate (TSP) gap is important for consideration of local effects development of the theory underlying thermal stress in heat
(Chen et al., 2011). Therefore, investigation of the influence of gaps transfer tubes, and the thermal stress of an elastic tube has been
on the thermal stress of the heat transfer tubes in a SG is of critical theoretically studied by many researchers (see e.g., Sauer (1996),
significance for maintaining safe operation of a pressurized water Guerreri and Cossa (1998), Orcan and Eraslan (2001), Cardella
reactor nuclear power plant. (2002), Eraslan and Orcan (2002)). Li et al. (1999) investigated
Thermal stress within a SG is closely related to its thermal– the thermal stress of tube plates in a heat exchanger using a finite
hydraulic characteristics (Kao et al., 1982). The fluid flow and heat element method, and proposed methods to reduce and eliminate
transfer between the primary and secondary sides is an especially thermal stress and improve tube plate structure. Thermal stress
complex two-phase flow, where the boiling heat transfer of the caused by fully developed laminar flows and pulsatile flows inside
secondary side has a large impact on wall temperature. A porous heating tubes has also been investigated by Al-zaharnah et al.
(2000, 2001a,b). Rahimi et al. (2003) conducted experimental
⇑ Corresponding authors. Fax: +86 0451 82519305 (B. Sun). Fax: +86 0451 research on the thermal stress distribution of a boiling bundle
82569633 (L. Yang). using a resistance wire heating method. Yapici and Albayrak
E-mail addresses: [email protected] (B. Sun), [email protected] (L. Yang). (2004) conducted a numerical investigation of inner wall thermal

http://dx.doi.org/10.1016/j.anucene.2015.09.019
0306-4549/Ó 2015 Elsevier Ltd. All rights reserved.
 B. Sun et al. / Annals of Nuclear Energy 87 (2016) 400–410 401

Nomenclature

a liquid phase q sensible heat transfer rate between the liquid phase and
b vapor phase vapor phase (kJ/(m3 s))
u volume fraction e thermal strain
q density (kg/m3) u displacement of micro unit in the x direction (m)
mba mass transfer rate from liquid phase to vapor phase v displacement of micro unit in the y direction (m)
(kg/(m3 s)) w displacement of micro unit in the z direction (m)
u velocity vector (kg/s) E Young’s modulus (GPa)
t time (s) r thermal stress (GPa)
le effective dynamic viscosity (kg/(m s)) l Poisson’s ratio
T temperature (K) n thermal expansion coefficient
p pressure (Pa) DT temperature variation (K)
g gravitational acceleration (m/s2) c shear strain
fab the force exerted by liquid phase to vapor phase (N) s shearing stress (GPa)
fba the force exerted by the phase change from vapor phase G shear modulus (GPa)
to liquid phase (N) r2 Laplacian
H enthalpy (kJ/kg) X, Y, Z volume force component (N/m3)
k thermal conductivity (W/(m K)) e volumetric strain

stress under non-uniform heat flux using commercial software, diameter of the heat transfer tubes, tube pitch, etc. The inlet and
comparing the distribution of thermal stress under different outlet positions of the primary and secondary sides in the unit-
heating methods and flow velocity. Özceyhan (2005) obtained pipe model are also consistent with the physical reference. The pri-
the thermal stress distribution of the heat transfer tubes under mary flow region is inside the tube, while the secondary flow
non-uniform heat flux via different heating methods using a model region is in between the tubes, taking into consideration the tube
based on the roughness of the heat transfer tubes and the velocity, wall thickness. In addition, we arrange nine pairs of four-leaf
and proposed methods to reduce thermal stress. Tashakor et al. plum-shaped support plates along the axis in the secondary flow
(2011) predicted positions prone to stress corrosion by analyzing region, such that they have potential gaps between the TSPs and
the positions of deposited impurities and corrosion products. the heat transfer tubes. The tube diameter is 19.05 mm, wall thick-
Garrido et al. (2012) conducted a numerical simulation of the heat ness is 1.09 mm, tube pitch is 27.43 mm, the height of the straight
transfer and thermal stress of heat transfer tubes, pressure pipe- tube section is 9 m, and the radius of the elbow section is 0.82 m.
line, and pump in a SG and predicted the range of fatigue load The thickness of the TSPs is 0.03 m, adjacent TSP pitch is 1 m and
under given conditions. the gaps between the TSPs and the heat transfer tubes is 0.3 mm.
As can be seen from the above analyses, SG investigations have We first partition all the blocks for the model, adjust the posi-
mostly focused on thermal–hydraulic characteristics or on thermal tions of the mesh nodes around each gap, make the mesh size
stresses induced in the heat transfer tubes without considering the increase proportionally layer-by-layer, and meet the requirements
effect of TSP gaps on the thermal–hydraulic characteristics of the of the expansion rate and number of grids using ICEM software.
primary and secondary domains and U-tube bundles within SG, The mesh system is shown in Fig. 2. A careful mesh independence
and the thermal stress have been typically analyzed by adding study was carried out in order to provide internally coherent
secondary or tertiary boundary conditions only considering numerical results. The eight mesh systems used in this study are
the influence of secondary side. However there is a strong summarized in Fig. 3. The final number of meshes is 5437326,
flow-thermal-solid coupling relationship among complex two- based on the mesh distortion, aspect ratio, expansion rate and
phase flow, the heat transfer and thermal stress in a SG, and the mesh independence.
gaps at the TSPs directly affect the thermal–hydraulic characteris-
tics of the SG, the thermal stress associated with the heat transfer
3. Mathematical formulation
tubes, and the coupling relationship. This paper conducts a numer-
ical simulation of a SG’s thermal–hydraulic characteristics that
Since there are common equations governing U-tube heat con-
considers the gaps between the tubes and the TSPs in the actual
duction and single-phase convection heat transfer in the primary
structure of the SG on the basis of Li et al. (2013), Sun and Yang
side, this paper does not provide any detail concerning these basic
(2013) and Yang et al. (2013), then passes these fluid calculation
equations. Instead, the following section primarily illustrates the
results to the structure and calculates the thermal stress of the
mathematical equations applied to the secondary domain and
heat transfer tubes. This method reveals the influence of gaps on
the thermal stress for the U-tube bundle.
the thermal–hydraulic characteristics in a SG as well as the
distribution of the thermal stress.
3.1. Liquid–vapor two-phase flow model

2. Physical model and mesh For our flow model, the steam quality at the outlet of secondary
side is about 0.25 and the flow pattern is only bubbly flow in the
This paper establishes a three-dimensional unit-pipe model, boiling process because of the low steam quality, thus we treat
including primary and secondary sides, heat transfer tubes, and the secondary side of a SG experiencing a two-phase flow
four-leaf plum-shaped support plates, using the structural and consisting of liquid and vapor phases as possessing continuous
operational parameters of a SG at the Daya Bay Nuclear Power and discrete phases, respectively. The governing equations for
Plant (DBNPP). As shown in Fig. 1, the geometry of the model is the liquid and vapor phases are established using the fluid pair
identical to the actual steam generator for elements such as the model (Sun and Yang, 2013).
402 B. Sun et al. / Annals of Nuclear Energy 87 (2016) 400–410

Fig. 1. Three-dimensional unit-pipe model.

Fig. 2. Local mesh schematic.

Continuity equation of the liquid phase: Momentum equation of the vapor phase:
@ @
ub qb ub Þ þ r  ðub ðqb ub ub  leb ðrub þ ðrub ÞT ÞÞÞ
ð
ðu q Þ þ r  ðua qa ua Þ ¼ mba ; ð1Þ @t
ð4Þ
@t a a ¼ ub ðqb g  rpb Þ þ f ba þ mba ub ;
Continuity equation of the vapor phase:
Energy equation of the liquid phase:
@
ðu q Þ þ r  ðub qb ub Þ ¼ mba ; ð2Þ @
ua qa Ha Þ þ r  ðua ðqa ua Ha  ka rT a ÞÞ
ð
@t b b @t
ð5Þ
¼ qab  mba Ha ;
Momentum equation of the liquid phase:



 Energy equation of the vapor phase:
@
@t
ð ua qa ua Þ þ r  ua qa ua ua  lea rua þ ðrua ÞT
ð3Þ @
ub qb Hb Þ þ r  ðub ðqb ub Hb  kb rT b ÞÞ
ð
¼ ua ðqa g  rpa Þ þ f ab  mba ua ; @t
ð6Þ
¼ qba þ mba Hb ;
 B. Sun et al. / Annals of Nuclear Energy 87 (2016) 400–410 403

0.30 605

Coolant Outlet Temperature on the Primary Side/K

Numerical Simulation Result Numerical Simulation Result
600
Outlet Steam Quality on the Secondary Side
Actual Operating Result Actual Operating Result
0.25
595

0.20 590

585
0.15
580

0.10 575

570
0.05
565

0.00 560
6
0 1 2 3 4 5 6 7 ×10 0 1 2 3 4 5 6
6
7 ×10
Grid Number Grid Number

(a) Outlet Steam Quality (b) Coolant Outlet Temperature

Fig. 3. Mesh independence validation.

The turbulence model, thermal phase change model, and temperature field to static structure analysis module to conduct
interphase mass, momentum, and energy transfer models of the the calculation of thermal strain and stress. Thus we can achieve
two-phase flow on the secondary side are described in detail by the one-way fluid-thermal-solid interaction numerical simulation
Sun and Yang (2013). through above steps. Specifically, the calculation region of the
steam generator includes secondary side, tube and primary side,
3.2. Thermal stress analysis model but only the mass, momentum and energy equations of vapor–liquid
two-phase in secondary side (Eqs. (1)–(6)) are shown due to space
According to Hooke’s law of elasticity (Li et al., 2003), the limitation. We know that Eqs. (1)–(6) are not enough for the
relationship between thermal stress and thermal strain can be calculation because the equations are not closed-form expression.
expressed as: We still need turbulence model and thermodynamic phase change
9 model which characterize the mass, momentum and heat transfer
ex ¼ @u
@x
¼ 1E ½rx  lðry þ rz Þ þ nDT >
= between vapor and liquid phases in secondary side. In addition, we
ey ¼ @@yv ¼ 1E ½ry  lðrz þ rx Þ þ nDT ; ð7Þ also need single-phase convective heat transfer differential equa-
>
;
ez ¼ @w
@z
¼ 1E ½rz  lðrx þ ry Þ þ nDT tion of primary side and heat conduction differential equation of
tube wall. The above equations not given in the manuscript are
sxy syz szx shown in previously published papers (see Sun and Yang (2013)).
cxy ¼ ; cyz ¼ ; czx ¼ ; ð8Þ
Thus we can obtain the flow field and temperature field of primary,
G G G
secondary sides and tube wall by solving the equations including
The equilibrium differential equations-displacement equations:
9 two-fluid model of the secondary side, single-phase convective
@e
ðk þ GÞ @x þ Gr2 u  b @T þX ¼0> > heat transfer differential equation of primary side and heat con-
@x =
duction differential equation of tube wall et al. The above work
@e
ðk þ GÞ @y þ Gr2 v  b @T þ Y ¼ 0 ; ð9Þ
@y >
> is a prerequisite before going for the thermal stress calculation.
;
ðk þ GÞ @x þ Gr w  b @z þ Z ¼ 0
@e 2 @T
After obtaining the temperature field, we can work out the thermal
stress distribution along x, y, z directions in static structure analy-
The deformation continuity equations-coordination equations:
sis module following above steps, and then calculate the thermal

9
r2 rx þ 1þ1l ¼ nE 11l r2 T þ 1þ1l @2 H @2 T >
>
>
stress over the length, thickness and circumference of the tube


@x2 @x2
 >
> through the conversion relation between Cartesian and cylindrical
>
@ T > >
r2 ry þ 1þ1l @@yH2 ¼ nE 11l r2 T þ 1þ1l
2 2
@y2 >
> coordinates.
>
>

 >
>
@2 T =
r2 rz þ 1þ1l @@zH2 ¼ nE 11l r2 T þ 1þ1l
2
@z 2
; ð10Þ 3.3. Boundary conditions
>
>
r2 sxy þ 1þ1l @2 H
@x@y
nE
¼  1þ l
@2 T
@x@y
>
>
>
>
> The boundary conditions in our model are imposed based on
>
>
r2 syz þ 1þ1l @2 H nE
¼  1þ l
@2 T
>
> domains and the interfaces between the primary side, secondary
@y@z @y@z >
>
>
;
r2 szx þ 1þ1l @2 H
@z@x
nE
¼  1þ l
@2 T
@z@x
side, and heat transfer tubes are set up as fluid–solid coupling
interfaces, and other symmetrical walls are set up as symmetrical
nE
Here, e = ex + ey + ez, b ¼ 12 l ; H ¼ 12l ðe  3nTÞ:
E boundaries. The fluid in the primary and secondary sides is liquid
Here we make a detailed description of solution process of and liquid and steam, respectively, as per the IAPWS IF97 database;
temperature field and thermal stress to elaborate how the Ts are specific parameters are shown in Table 1.
calculated over the length, thickness and circumference prior to
evaluating thermal stress distribution. After calculating the 3.4. Material properties
temperature field according to mass, momentum and energy
equations, we pass the temperature load of heat transfer tubes to To provide a static analysis of the heat transfer tubes in a SG, we
static-state thermal analysis module and this module pass the must first set the Young’s modulus and Poisson’s ratio material
404 B. Sun et al. / Annals of Nuclear Energy 87 (2016) 400–410

Table 1
Boundary conditions.

Calculation domain Boundary Name Unit Value

Primary side Inlet Mass flow rate kg/s 4.16
Temperature K 600.75
Outlet Pressure MPa 15.50
Secondary side Inlet Mass flow rate kg/s 1.887
Temperature K 543.45
Outlet Pressure MPa 6.81

Table 2
Correlation of material properties.

Physical parameter Unit Correlation

Thermal conductivity W/(m K) 0.0186T + 6.62 Fig. 4. Temperature curves in the flow direction for the primary and secondary
Young’s modulus GPa 0.0757T + 232.55 fluids, and the inner and outer walls of the heat transfer tubes.
Thermal expansion coefficient lm/(m °C) 0.004T + 12.36

properties of the tubes, as well as define the thermal expansion

coefficient and thermal conductivity when a thermal load is
applied. For this work, we fitted the thermal conductivity, elastic
modulus, and thermal expansion coefficient of Inconel 690 alloy
into the correlations described in Table 2 using experimental data
measured under different temperatures. We also used a Poisson’s
ratio of 0.3 for the material. Based on the diameter and wall thick-
ness of the heat transfer tubes, and referring to the relevant exper-
imental data (see e.g., Hongbing et al. (2003)), we calculated the
yield limit and ultimate strength of the material as 379 MPa and
700 MPa, respectively.

4. Results and discussion

For this work, we conducted a numerical simulation of the ther-

mal–hydraulic characteristics in a SG, with and without gaps Fig. 5. Steam quality curves in the axial direction on the secondary side.
between the TSPs and the heat transfer tubes, using CFX software.
The results show that these gaps significantly affect the local ther- The variation in steam quality along the axial direction of the
mal–hydraulic characteristics at the TSPs, but have a negligible heat transfer tubes on the secondary side is shown in Figs. 5 and
effect on global thermal–hydraulic characteristics such as steam 6. As can be seen from the figures, the steam quality of the cold
quality and surface heat transfer coefficient along the height of leg, hot leg, and mean value along the axial height is initially
the heat transfer tubes. Therefore, this paper mainly analyzes the constant before starting to increase. The single-phase flow heat
vapor–liquid two-phase characteristics of the secondary side with transfer region (preheating section) is about 2 m. When the tubes
gaps in a SG, and the effect of those gaps on local thermal–hy- start to curve, the parallel flow converts to a cross flow, resulting
draulic characteristics and the thermal stress of the heat transfer in improved heat transfer and a rapidly rising steam quality
tubes. (Damiani et al., 2013; Wang et al., 2014). When DBNPP is running
at full load, the circulation ratio is about 4, meaning that the steam
4.1. Two-phase flow and boiling heat transfer characteristics on the quality at the outlet is about 0.25. The result obtained from the
secondary side numerical simulation result is 0.24, consistent with the actual
operating parameters of the SG.
As can be seen from Fig. 4, the coolant temperature on the pri- Fig. 7 shows the variation in surface heat transfer coefficient
mary side linearly decreases in the flow direction with a corre- between the secondary fluid and the outer wall of the heat transfer
sponding outlet temperature of 570.89 K while the measured tubes. The figure shows that the surface heat transfer coefficients
temperature at the outlet of the physical reference can approach of the cold leg, the hot leg, and the average value SG of cold and
565.55 K when DBNPP is running at full load (Ding (2003)). This hot leg all gradually increase; the mass flow rate, velocity, and heat
result proves the accuracy of our coupled heat transfer numerical transfer are all greater in the hot leg because the secondary side
simulation. The fluid temperatures of the cold and hot legs fluid of the hot and cold legs is allocated in a 4:1 ratio. At each
gradually rise in the preheating section, then reach a saturated TSP, the surface heat transfer coefficient of the secondary side fluid
temperature following a slightly superheated state before remain- first sharply increases before decreasing rapidly. The main reason
ing constant. Cong et al. (2013) and Wang et al. (2014) justify the for this spike is that the cross-sectional area of the flow decreases
discussion of temperature. However, the average surface heat at the four-leaf plum-shaped support plates as the flow velocity
transfer coefficient changes rapidly at the TSP positions, as seen increases. This enhances the turbulence of the boundary layer
in Fig. 7, because the flow structure of the secondary fluid changes and makes the heat transfer more intense. However, this heat
at the TSPs, thus resulting in a rapid oscillation in the distribution transfer enhancement is not obvious with the expansion of the
of the outer wall temperature of the heat transfer tubes at these jet flow cross-section after flowing through the TSPs, as the surface
points. heat transfer coefficient rapidly decreases.
 B. Sun et al. / Annals of Nuclear Energy 87 (2016) 400–410 405

Fig. 6. Three-dimensional distribution of steam quality on the secondary side at the sixth TSP.

surface heat transfer coefficient, which calculated by the widely-

approved empirical correlation in engineering, all means the mean
value. Thus we compare the integral mean value with the empirical
result (in Fig. 7) and not the average one to verify the numerical
results of surface heat transfer coefficient. The relative deviation
between mean value and Rohsenow’s empirical correlation is
4.4%, which ensures the accuracy of our numerical results.

4.2. The effect of gaps on two-phase flow and boiling heat transfer

Fig. 8(a and b) indicate that the distribution of the vapor phase
velocity at the eighth TSP is similar to that of the liquid phase
velocity, even though the vapor phase velocity is significantly
greater than the liquid phase velocity. The velocities at the flood
holes and at the gaps are about 12 m/s and 4 m/s, respectively.
As can be seen from Fig. 8(c), the steam quality at the gaps is
Fig. 7. Heat transfer coefficient curves in the axial direction on the secondary side. higher than at the flood holes. The reason for this difference is that
the mass flow rate at the gaps is lower (Fig. 8(a)), which means
most of the liquid moving through the gaps is vaporized. Since
Since the study object of the manuscript is the natural circula- the vapor phase viscosity is less than the liquid phase viscosity,
tion U-shaped steam generator running in Daya Bay Nuclear Power more vapor can enter each gap, thus resulting in a higher steam
Plant (DBNPP), it is unrealistic to conduct experimental study for quality at the gaps. Fig. 8(d) shows the distribution of the steam
this situation. The operational data found in DBNPP is limited that quality beyond the eighth TSP. Part of the vapor gathers at the flow
only includes mass flow, pressure, temperature, steam quality et al. channel center due to wall lubrication and buoyancy forces when
at inlet and outlet of primary and secondary sides. Therefore we the fluid flows through the TSP, which leads to a higher steam
compare the outlet temperature of primary and secondary sides, quality at the center of the flow channel.
outlet steam quality of secondary side with operational data hop- Fig. 9 shows the change in outer wall temperature in the
ing to verify our mathematical model and method. However, above circumferential direction with and without gaps at the position
verification method only verify the macro-parameters at inlet and of the sixth TSP. Here, h represents the circumferential position
outlet, and is lack of the verification of key parameters related to of the central tube (Fig. 10). Note that the outer wall temperature
heat transfer such as surface heat transfer coefficient. So Fig. 7 also at the gaps is significantly higher than at the flood holes, giving
shows the comparison between the numerical simulation result the wall temperature curve a periodic oscillation around the
and Rohsenow’s empirical correlation (Rohsenow, 1985) result circumference of the tube. The main reason for this result is that
with respect to the average surface heat transfer coefficient of the flow rate through the gaps only accounts for 3–8% of the total
the secondary side. In Fig. 7, SG represents the average heat trans- flow rate; i.e., most of the fluid travels through the flood holes.
fer coefficient of hot and cold leg, the mean value represents that Since the flow rate is greater, the heat transfer is more intense,
integrate the heat transfer coefficient along the entire heat transfer and so the wall temperature is lower at the flood holes compared
area of the tubes and then average the integral value. Now the to its temperature at the gaps. What can also be seen from the
406 B. Sun et al. / Annals of Nuclear Energy 87 (2016) 400–410

(a) Liquid phase velocity distribution contour (b) Vapor phase velocity distribution contour

(c) Steam quality distribution contour at the (d) Steam quality distribution contour beyond the
eighth TSP eighth TSP
Fig. 8. Flow and heat transfer characteristics at the eighth TSP.

stronger than the case of no gap. Therefore there is the difference

With Gaps in base line temperature of 1 K between the cases.
Without Gaps The cross-sectional temperature distribution contours of the
hot leg heat transfer tubes at the sixth TSP are shown in Fig. 10.
The inner wall temperature is higher than the outer wall tempera-
Temperature (K)

ture, while the inner wall undergoes a compressive stress and the
outer wall a tensile stress. The wall temperature at the gaps is
higher than at the flood holes, and the tube wall experiences com-
pressive and tensile stress at the gaps and flood holes, respectively.
These stresses result in a complex stress distribution at the TSPs.
Moreover, the change in wall temperature around the circumfer-
ence of the tube when there are no gaps is more intense, increasing
the circumferential thermal stress of the heat transfer tubes.
Therefore, blocking the gaps not only causes a decline in thermal
power but also increases the circumferential thermal stress, thus
accelerating stress corrosion of the heat transfer tubes.
θ (°)
The variation in circumferential outer wall temperature of the
Fig. 9. Outer wall temperature curves around the tube circumference at the sixth hot leg from the fifth to seventh TSP is clearly shown in Fig. 11.
TSP. The temperature amplitude progressively increases from the fifth
to seventh TSP; as the liquid is heated, more steam is generated,
which results in either alternating wet and dry flows appearing
figure is that the wall temperature amplitude for the structure at the heat transfer surface of the gaps, or steam cushions forming
without gaps is approximately twice that of the structure with in the gaps. This weakens the heat transfer and its role in cooling at
gaps; the wall temperature is lower at the gaps and distributes the gaps, resulting in an increase in the difference in circumferen-
more uniformly around the circumference of the heat transfer tial wall temperatures.
tubes because the fluid flows through the gaps, producing a normal Compared with the results from previous analyses, we can see
boiling heat transfer at the surface, where the boiling heat transfer that the velocity is lower and steam quality is greater at the gaps
is greater than for the structure without gaps. The fluid of compared to these parameters at the flood holes, resulting in a lack
secondary side flows through the gap in lower velocity and has a of liquid in the gaps. Meanwhile, the wall temperature at the gaps
certain cooling effect on the tube, the overall cooling effect is is significantly higher and liquid is easily evaporated at these
 B. Sun et al. / Annals of Nuclear Energy 87 (2016) 400–410 407

Fig. 10. Temperature distribution contours at the sixth TSP with and without gaps.

565
The Fifth TSP
The Sixth TSP
Outer Wall Temperature (K)

564 The Seventh TSP

563

562

561

560
-180 -120 -60 0 60 120 180
θ (°) Fig. 13. Steam quality distribution contour at the fifth TSP.

Fig. 11. Circumferential outer wall temperature at different TSPs.

between the gap locations (between angles, e. g. 150 and 110,
it is at 130, 50 and 10, it is at 20, etc.), and correspondingly
the intermediate peak/crest of wall temperature between the gap
locations. This is because steam quality is relatively small between
the gap locations compared to other locations, which indicates the
number of bubbles generated between the gap locations (between
angles, e. g. 150 and 110, it is at 130, 50 and 10, it is at 20,
etc.) is relatively few compared to other locations. While the
turbulence of bubble’s generation and departure from the wall will
enhance heat transfer. Thus the surface heat transfer coefficient is
relatively weak between the gap locations compared to other
locations. Thereby we can observe the intermediate peak/crest
between the gap locations (between angles, e. g. 150 and 110,
it is at 130, 50 and 10, it is at 20, etc.) for all the three TSPs
(Fig. 11).
It can be seen from Fig. 12 that, there really exists the
quantitative relation between steam quality and temperature dis-
tribution and its relevance in considering local effects like TSP for
the specific steam generator studied in the manuscript, but the
Fig. 12. The comparison between circumferential steam quality and wall temper- quantitative relation is affected by many factors (such as size of
ature at fifth TSP.
the gap, the form of TSP etc.). So further research need to be
done to figure out the common quantitative relation between
steam quality and temperature distribution and its relevance in
points, thereby accelerating the concentration and deposition of considering local effects like TSP.
impurities, which may increase the stress corrosion.
The steam quality distribution contour at the fifth TSP (Fig. 13)
and the comparison between circumferential steam quality and 4.3. Thermal stress analysis of the heat transfer tubes
wall temperature at fifth TSP (Fig. 12) are shown to satisfactorily
explain the quantitative relation between steam quality and tem- Fig. 14 indicates that the thermal stress of the hot leg slowly
perature distribution. It can be found from these figures that, the decreases along the axis of the tubes. The fluid velocity, heat
TSP and the gap leads to intermediate nadir of steam quality transfer coefficient, and local capacity of heat transmission all
408 B. Sun et al. / Annals of Nuclear Energy 87 (2016) 400–410

56 Wobo Zhang (2007) found although thermal stress does not exceed
52 Outer Wall the yield strength and ultimate strength, but the sudden change of
thermal stress combined with mechanical stress still has great
48 Inner Wall
influence on the overall SCC behavior.
Thermal Stress (MPa)

44 Fig. 16 shows the change in outer wall thermal stress around

40 the circumference of the tube at the first TSP of the hot leg both
with and without gaps. In the figure, both trends are basically
36
the same, showing a periodicity. From the thermal–hydraulic char-
32 acteristics in a SG (Fig. 8(a and b)), we can see that the flow rate is
28 greater, the velocity is higher, the heat transfer is more intense,
and the fluid takes more heat away from the wall at the flood holes
24
compared to the gaps, resulting in a lower outer wall temperature
20 and a larger temperature difference. Thus, the thermal stress of the
16 outer wall at the flood holes is larger than the thermal stress at the
-1 0 1 2 3 4 5 6 7 8 9 10
gaps. For the structure without gaps, the heat transfer between the
tubes and TSPs is through thermal conduction, so the capacity of
Axial Height (m) heat transmission is significantly less than the heat transmission
Fig. 14. Change in thermal stress of the hot leg along the axial height of the tube. caused by convection for the structure with gaps. Therefore, more
heat gathers on the wall, causing an increase in outer wall temper-
ature and a variation in temperature amplitude of the outer wall
increase at each TSP, causing the temperature difference between around the circumference of the tube, resulting in a rapid change
the inner and outer walls to decrease and the thermal stress to in thermal stress in the circumferential direction. This variation
drop to a local minimum. However, the velocity decreases, heat in amplitude is about twice that for the structure with gaps. Obvi-
transfer weakens, and the temperature difference between the ously plugging the gaps between the TSPs and tubes in the actual
inner and outer walls increases with the expansion of the flow operation process of a SG will lead to heat transfer deterioration
channel as the fluid flows through the TSPs, increasing the thermal and an increase in the variation in outer wall temperature ampli-
stress. The inlet of the primary and secondary side fluids is at the tude around the circumference at each TSP. This will produce
bottom of the hot leg in a SG, and the temperature difference is greater thermal stress, resulting in more pitting and denting of
greatest at this position. This results in a greater temperature dif- the heat transfer tubes, accelerating the stress corrosion, and
ference between the inner and outer walls, and the heat transfer affecting the operational security and stability of the SG.
of the tubes in the z-axis direction is limited due to this position Fig. 17 shows the variation in circumferential inner wall ther-
constraint. Thus, the maximum thermal stress of the hot leg occurs mal stress at the first TSP of the hot leg both with and without
at the end position of the heat transfer tubes and is about 51 MPa. gaps. It appears that the inner wall temperature also fluctuates cir-
The thermal stress of the cold leg and hot leg, respectively, ranges cumferentially due to the outer wall temperature, resulting in a
from 2.1 to 16.2 MPa for the cold leg and 16.2 to 50.7 MPa for the periodic oscillation in the thermal stress circumferentially around
hot leg. Note that thermal stress in the hot leg is significantly the tube. The amplitude of this thermal stress variation for the
greater than the thermal stress in the cold leg, and we can predict inner wall is less than that for the outer wall since the circumfer-
that the degradation effects caused by thermal stress on the heat ential temperature distribution of the primary side fluid is rela-
transfer tubes will be greater for the hot leg than the cold leg. tively uniform at the TSPs.
Given c represents the angle varying from the hot leg in the cir- The circumferential change in thermal stress of the inner and
cumferential direction at the elbow section in Fig. 15, we can see outer walls both show a periodic oscillation, but with opposite
that the thermal stress is lower at the elbow section near the cold trends, seen in Figs. 18 and 19. The amplitude of the oscillations
leg, and higher near the hot leg. The expansion of the pipe’s geom- on the inner wall is significantly less than that seen on the outer
etry at the elbow section produces a turbulent flow in the fluid, wall. The thermal stress of the inner wall decreases as the thermal
which enhances the heat transfer coefficient and causes the stress of the outer wall increases, and is less than the outer wall
temperature difference between the inner and outer walls to thermal stress at the flood holes. This occurs because the inner wall
drop, resulting in a corresponding decrease in thermal stress. temperature is higher than the outer wall temperature; the inner

With Gaps
Without Gaps
Thermal Stress (MPa)

θ (°)

Fig. 16. Variation in outer wall thermal stress around the tube circumference at the
Fig. 15. Change in thermal stress along the curve of the elbow section of the tube. first TSP of the hot leg.
 B. Sun et al. / Annals of Nuclear Energy 87 (2016) 400–410 409

The temperature difference between the inner and outer walls

With Gaps
Without Gaps decreases along the axial height of the heat transfer tubes in the
hot leg, as does the thermal stress. As shown in Fig. 20, the average
Thermal Stress (MPa)

thermal stress at the third, fourth, and fifth TSPs is smaller at each
successive TSP i.e. the average thermal stress at upper TSPs is smal-
ler than at lower TSPs but the change in amplitude of the thermal
stress circumferentially around the tube does not decrease. This is
related to the effect of the TSPs on heat transfer and indicates that
the local effect of the TSPs on the thermal stress of the heat transfer
tubes is basically consistent.
As can be seen in Fig. 21, the circumferential variation in outer
wall thermal stress at the ninth TSP is different from the circumfer-
ential change in thermal stress seen at the other TSPs. At the ninth
TSP, the thermal stress curve is smoother in the crests while small
θ (°)
fluctuations appear in the troughs. This occurs because the elbow
Fig. 17. Variation in inner wall thermal stress around the tube circumference at the tube section causes deformation due to thermal stress that gener-
first TSP of the hot leg. ates a force on the straight tube section. Thus, the forces acting on
the straight tube section are more complex at the ninth TSP, chang-
ing the distribution characteristics of the heat transfer tube stress.
It can be found from the numerical results that the cross-
sectional average steam quality changes along the length (Fig. 5),
and the TSPs have dramatically influence on local circumferential
steam quality of cross section where the TSPs are located
Thermal Stress (MPa)

(Figs. 6 and 8c and d). The local variation of circumferential steam

quality directly affects the wall temperature distribution (Fig. 12),
thereby affects the local thermal stress (Figs. 16–21). The variation
of steam quality means the number of bubbles generated in

Outer Wall The Third TSP

Inner Wall The Fourth TSP
The Fifth TSP
Thermal Stress (MPa)

θ (°)

Fig. 18. Circumferential thermal stress with gaps.

Outer Wall
Inner Wall
Thermal Stress (MPa)

θ (°)

Fig. 20. Circumferential variation in the outer wall thermal stress of the hot leg at
the third, fourth, and fifth TSPs.

θ (°)
Thermal Stress (MPa)

Fig. 19. Circumferential thermal stress without gaps.

The First TSP
The Ninth TSP

wall and outer wall suffer compressive stress and tensile stress,
respectively, in the radial direction. The circumferential change
in amplitude of the outer wall temperature and the outer wall tem-
perature difference in both the radial and circumferential direc-
tions are larger, while the circumferential change in amplitude of
the inner wall temperature is less and inner wall temperature
θ (°)
difference mainly exists in the radial direction, resulting in the
different changes in inner and outer wall thermal stresses Fig. 21. Circumferential variation in the outer wall thermal stress of the hot leg at
occurring around the circumference of the tube. the first and ninth TSPs.
410 B. Sun et al. / Annals of Nuclear Energy 87 (2016) 400–410

secondary side changes, the turbulence induced by bubble’s Cardella, A., 2002. Analytical methodology and boundary problem for computing
temperature and thermal stresses in tubes. Heat Technol. (Pisa, Italy) 20, 61–67.
generation and departure from the wall will enhance heat transfer.
Chen, Y.Q., Gui, C., Wang, X.Y., 2011. Analytical investigation on steam generator
The more bubbles, the more severe turbulence and intense heat tube fretting due to foreign object. Nucl. Power Eng. 32, 21–24 (in Chinese).
transfer. More intense heat transfer means better cooling on the Cong, T., Tian, W., Qiu, S., et al., 2013. Study on secondary side flow of steam
wall, results lower wall temperature. It can also be seen from mass, generator with coupled heat transfer from primary to secondary side. Appl.
Therm. Eng. 61 (2), 519–530.
momentum and energy conservation equations of vapor–liquid Damiani, L., Montecucco, M., Prato, A.P., 2013. Conceptual design of a bayonet-tube
two-phase in mathematical model, that the variation of steam steam generator for the ALFRED lead-cooled reactor. Nucl. Eng. Des. 265, 154–
quality will lead to change in velocity, thereby affect the heat 163.
Ding, X.S., 2003. Natural circulation steam generator operating principles and
transfer (i.e., the surface heat transfer coefficient changes), which operating parameters. Nucl. Power Plant (China) 2, 15–21.
lead to changes in wall temperature and local thermal stress Eraslan, A.N., Orcan, Y., 2002. Computation of transient thermal stresses in elastic-
distribution at this position. plastic tubes: effect of coupling and temperature-dependent physical
properties. J. Therm. Stresses 25, 559–572.
Ferng, Y.M., 2007. Investigating the distribution characteristics of boiling flow and
5. Conclusions released nuclide in the steam generator secondary side using CFD methodology.
Ann. Nucl. Energy 34, 724–731.
Ferng, Y.M., Chang, H.J., 2008. CFD investigating the impacts of changing operating
There is a rapid change in the secondary side’s velocity and heat conditions on the thermal-hydraulic characteristics in a steam generator. Appl.
transfer coefficient where the TSPs are placed, and the steam qual- Therm. Eng. 28, 414–422.
ity and fluid velocity is higher and lower, respectively, in the gaps Garrido, O.C., Cizelj, L., Simonovski, I., 2012. Modular 3-D solid finite element
model for fatigue analyses of a PWR coolant system. Nucl. Eng. Des. 246, 115–
between the TSP and heat transfer tube than in the flood holes. 122.
A periodic oscillation in the outer wall temperature occurs in Guerreri, G., Cossa, L.M., 1998. Temperature stresses in thick circular cylindrical
the circumferential direction of the heat transfer tubes at the TSPs, shells used in nuclear reactors with exponential heat generation within the
shell wall thickness. Proc. Inst. Mech. Eng. C J. Mech. Eng. Sci. 212, 555–565.
such that the wall temperature at the gaps is slightly higher than Hongbing, Z., Shoujun, L., Yaohe, H., 2003. Foreign research status of Inconel 690
the wall temperature at the flood holes. However, at contact points used in steam generator heat transfer tubes. Special Steel Technol. (China) 4, 2–
the wall temperature is significantly higher than at the flood holes 11.
Kao, T.T., Cho, S.M., Pal, D.H., 1982. Thermal modeling of steam generator tubing
for the structure without gaps, causing rapid changes in the wall under CHF-induced temperature oscillations. Int. Heat Mass Transfer 25 (6),
temperature circumferentially around the tube and increasing 781–790.
the temperature difference within the circumferential distribution. Li, Y., Yang, Y., Liu, M.S., Dong, Q.W., Wang, D.B., Ling, X., 1999. Numerical
simulation of thermal stress in tube-sheet of heat transfer equipment. Int. J.
The inner wall thermal stress is slightly higher than the outer
Press. Vessels Pip. 76, 671–675.
wall thermal stress, but the change in thermal stress for both wall Li, W.C., Huang, B.H., Bi, Z.B., 2003. Theoretical Analysis and Application of Thermal
surfaces along the axial height of the tubes are consistent. Thermal Stress. China Electric Power Press, 69-70.
Li, Y., Yang, Y., Sun, B., 2013. Numerical investigation of thermal–hydraulic
stress increases sharply at the TSPs, and the maximum thermal
characteristics in a steam generator using a coupled primary and secondary
stress of about 51 MPa occurs at the bottom of the hot leg. side heat transfer model. Ann. Nucl. Energy 55, 258–264.
The circumferential change in inner and outer wall tempera- Orcan, Y., Eraslan, A.N., 2001. Thermal stresses in elastic-plastic tubes with
tures experience opposite trends. Moreover, the inner wall thermal temperature-dependent mechanical and thermal properties. J. Therm. Stresses
24, 1097–1113.
stress is slightly higher than the outer wall thermal stress and the Özceyhan, V., 2005. Conjugate heat transfer and thermal stress analysis of wire coil
circumferential change in amplitude is less than what is seen for inserted tubes that are heated externally with uniform heat flux. Energy
the outer wall. Convers. Manage. 46, 1543–1559.
Rahimi, M., Owen, I., Mistry, J., 2003. Thermal stresses in boiler tubes arising from
high-speed cleaning jets. Int. J. Mech. Sci. 45, 995–1009.
Acknowledgements Rohsenow, W.M., 1985. Handbook of Heat Transfer, Fundamentals, second ed.
McGraw-Hill Book Company, New York, 12.2–12.18.
Sauer, G., 1996. Simple formulae for the approximate computation of axial stresses
This work was sponsored by the National Natural Science Foun- in pipes due to thermal stratification. Int. J. Press. Vessels Pip. 69, 213–223.
dation of China (Nos. 51479040, 51579048) and Natural Science Sun, B., Yang, Y., 2013. Numerically investigating the influence of tube support
Foundation of Heilongjiang Province of China (Nos. E201422, plates on thermal–hydraulic characteristics in a steam generator. Appl. Therm.
Eng. 51, 611–622.
E201346), and Postdoctoral Research Startup Foundation of
Tashakor, S., Jahanfarnia, G., Kebriaee, A., 2011. Numerical model for estimation of
Heilongjiang Province of China (No. LBH-Q14036) which are corrosion location in nuclear power plant steam generators. Nucl. Eng. Des. 241,
gratefully acknowledged. 95–100.
Wang, C., Cong, T., Qiu, S., et al., 2014. Numerical prediction of subcooled wall
boiling in the secondary side of SG tubes coupled with primary coolant. Ann.
References Nucl. Energy 63, 633–645.
Yang, Y., Sun, B., Li, Y., Yang, L., Zheng, L., 2013. CFD investigation of thermal–
Al-Zaharnah, I.T., Yilbas, B.S., Hashmi, M.S.J., 2000. Conjugate heat transfer in fully hydraulic characteristics for a steam generator with and without tube support
developed laminar pipe flow and thermally induced stresses. Comput. Methods plates. Proc. Inst. Mech. Eng. C J. Mech. Eng. Sci., 0954406213479740
Appl. Mech. Eng. 190, 1091–1104. Yapici, H., Albayrak, B., 2004. Numerical solutions of conjugate heat transfer and
Al-zaharnah, I.T., Hashmi, M.S., Yilbas, B., 2001a. Thermal stresses in thick-walled thermal stresses in a circular pipe externally heated with non-uniform heat
pipes subjected to fully developed laminar flow. J. Mater. Process. Technol. 118, flux. Energy Convers. Manage. 45, 927–937.
50–57. Zhang, Wobo, 2007. Breakage characteristic of thermal–mechanical couple
Al-Zaharnah, I., Yilbas, B.S., Hashmi, M.S.J., 2001b. Pulsating flow in circular pipes— alternating fatigue for gasoline engine piston ring bank. Chin. J. Mech. Eng. 43
the analysis of thermal stresses. Int. J. Press. Vessels Pip. 78, 567–579. (10), 89–93.