Annals of Nuclear Energy 139 (2020) 107278

Contents lists available at ScienceDirect

Annals of Nuclear Energy

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

Pre-CHF boiling heat transfer performance on tube bundles with or

without enhanced surfaces - a review
Shuai Ren b, Wenzhong Zhou a,b,⇑
a
Sino-French Institute of Nuclear Engineering and Technology, Sun Yat-Sen University, Zhuhai 519082, PR China
b
Department of Mechanical Engineering, City University of Hong Kong, Hong Kong, China

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

Article history: Boiling heat transfer over tube bundles has been extensively applied to various industries with a high
Received 16 April 2019 demand for efficient heat transfer. This work presents a review of recently published studies on pre-
Received in revised form 16 December 2019 CHF boiling heat transfer across plain and enhanced tube bundles. Bundle effect and heat transfer
Accepted 18 December 2019
enhancement by modified heating surfaces under various operating and geometric parameters are ana-
lyzed. Flow regime maps for boiling two-phase flow in horizontal and vertical bundles are critically
described separately. The local boiling heat transfer performance is affected by the non-uniform heat flux
Keywords:
distribution in a bundle. A decreasing heat flux distribution along the bundle height can enhance the bun-
Boiling
Bundle effect
dle effect. The effect of the pitch to diameter ratio on bundle effect also depends on the heat flux distri-
Heat transfer enhancement bution. Significant influences of the bundle inclination angle and elevation angle on the boiling heat
Enhanced surfaces transfer were observed by researchers. Complex bundle effect was found in special shape bundles, such
Two-phase flow as V-shape, C-shape, and U-shape bundles, which suggests applying different HTC correlations to differ-
ent regions in a bundle. Moreover, the bundle boiling behaviors under sub-atmospheric and sub-critical
pressures have been examined. The heat transfer performance in tube bundles with enhanced surfaces is
significantly impacted by the surface characteristics and the imposed heat flux. Bundle effect is still
prominent, and the surface enhancement reduces along the bundle height. A mixed bundle with
enhanced tubes only in the lower part can achieve the same heat transfer performance as a fully
enhanced bundle.
Ó 2019 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. Flow patterns of two-phase flow in tube bundles involving boiling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
3. Boiling heat transfer over plain tube bundles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.1. Bubble departure characteristics in bundle boiling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2. Bundle effect in horizontal bundles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.3. Effects of bundle layout on bundle boiling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.4. Pressure effects on bundle boiling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.5. Bundle effect in vertical and special-shaped bundles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4. Boiling heat transfer over tube bundles with enhanced surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
6. Recommendations for future studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Declaration of Competing Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

⇑ Corresponding author at: Sino-French Institute of Nuclear Engineering and Technology, Sun Yat-Sen University, Zhuhai 519082, PR China.
E-mail address: [email protected] (W. Zhou).

https://doi.org/10.1016/j.anucene.2019.107278
0306-4549/Ó 2019 Elsevier Ltd. All rights reserved.
2 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

Nomenclature

a empirical coefficient in Eq. (1) Tsup wall superheat temperature, (Tw-Tsat), K

b
0 0 in Eq. (2)
empirical coefficient Tw wall temperature, K
q
Bo boiling number, Gh u velocity, m/s
2 
fg qu D
Bd bond number WeL liquid Weber number, L rL o
Cp specific heat, J/kg∙K x vapor quality, (–)  
0:5
0:1
0:9 qG lL
Dd bubble departure diameter, mm Xtt Martinelli number, 1x
x q l
L G
f bubble departure frequency, Hz YIB boiling intensity parameter, (–)
f2/ two-phase friction factor
F enhancement factor, (–)
Greek symbols
Fpb hnpb,actual/hcooper
C falling film flow rate, kg/m∙s
Fr Froude number, (G2 =ðq2L gDÞ)
a thermal diffusivity, m2/s
G mass flux, kg/m2s
l dynamic viscosity, Pa∙s
h heat transfer coefficient, w/m2K
e vapor void fraction, (–)
hfg latent heat of vaporization,

 J/kg
qL C p T sup
k thermal conductivity, W/m∙K
Ja Jakob number, qG hfg r surface tension, N/m
jl superficial liquid velocity, m/s / hTP/hCVL
jg superficial vapor velocity, m/s /0 value of / at x = 0
L length of the tube, m
m_ mass flow rate, kg/s Subscripts
NR the number of tube rows B boiling
Nu Nusselt number, (hD/k) b bundle
KBB bundle factor CVL liquid phase convection
P tube pitch, mm  0 0 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi dry dryout
Peb Peclet number, aqq h r
gðq q Þ f fluid
G fg L G

Ph horizontal tube pitch, mm G gas

Pv
mm
vertical tube pitch,  in inlet
C l L liquid
Pr Prandtl number, pk
p pressure, Pa npb nucleate pool boiling
pr reduced pressure, p/pcrit pb pool boiling
pcrit critical pressure, Pa TP two phases
q’’ heat flux, W/m2 u lowest tube row
Ra average tube surface
roughness,
 lm v vapor
Re Reynolds number, qLluD


ReL Reynold number of single phase, Gð1xÞD Acronyms
lL
S suppression factor, (–) CHF critical heat flux
T temperature, K DNB departure from nucleate boiling
Tsat saturated temperature, K HTC heat transfer coefficient
Tsub subcooled temperature, K

1. Introduction latent heat transfer due to bubble formations and departure, and
bubbles originated on the lower tubes (bottom part) sliding along
As an efficient heat transfer mode, boiling has been extensively the sides of the upper tubes (upper part), which work together and
applied to various industries, such as petrochemical, nuclear power result in the so-called bundle effect. The two-phase flow in the
generation, refrigeration and air conditioning, seawater desalina- upper part of the bundle is affected by the lower part of the bundle,
tion, and food processing. A single tube is seldom used in boiling and thus when the same heat flux is applied to each tube of a bun-
applications, and instead in-line or staggered tubes are arranged dle, the boiling heat transfer coefficient (HTC) is observed to be
in a bundle and submerged in the boiling liquid. It has been widely increasing along with the bundle height. This enhancement will
established that heat transfer can be enhanced when occurring disappear when heat flux further increases and fully developed
over a tube bundle compared to over one single tube under the nucleate boiling is reached. The bundle effect is usually quantita-
equivalent conditions. Two-phase shell and tube exchanger is tively defined as the ratio of the overall or averaged HTC of the tube
one of the most common heat transfer devices which employs tube bundle to the HTC of an isolated single tube under the same heat
bundles as the heat transfer structure that allows fluids with differ- flux. The main influential variables on the bundle effect are fluid
ent temperature exchanging heat through the tube surfaces. To materials, bundle geometry, tube’s surface characteristics, operat-
effectively remove the generated vapor and maintain stable oper- ing conditions, nucleate boiling contribution, the corresponding
ation, shell side boiling is usually adopted. flow regimes, and the onset of dryout.
The vapor bubbles generated on the lower tubes (horizontal The bundle effect tends to be more prominent on plain or low
tube bundle) or the bottom part of the tubes (vertical tube bundle) finned tube bundles than modern enhanced tube bundles which
flow upwards through the entire tube bundle, thereby establishing are dominated by heat flux and subsurface structures (Liu and
a vapor–liquid two-phase flow which influences the heat transfer Qiu, 2002; Gorgy and Eckels, 2012). Cornwell and Schüller (1982)
on individual tubes. The two-phase heat transfer in a bundle con- and Cornwell (1990a,b) did several fundamental studies on bubbly
sists of contributions from single-phase convective heat transfer, flow boiling on the smooth tube bundle. Cornwell and Schüller
 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278 3

(1982) experimentally investigated the bubble sliding and growth involved, there are bubble formation and movement on the tube
near the top of a horizontal reboiler tube bundle using high-speed walls, which make the two-phase flow through the tube bundle
photography and confirmed that sliding bubbles could account quite intricate. Moreover, the distribution of the heat transfer coef-
for the enhancement of heat transfer observed at the upper tubes ficients over a tube bundle is not uniform but depends on the
of bundles. Cornwell (1990b) dissociated the total heat transfer of specific position and bundle layout, so that the corresponding bub-
nucleate boiling in three constituent parts: liquid convection, bub- ble dynamics and two-phase flow could also vary according to dif-
ble formation and growth, and ‘‘sliding bubble” mechanism. ferent locations in the bundle. The internal structures of two-phase
Cornwell (1990a) further examined the effects of sliding bubbles flow are classified by the flow regimes or flow patterns. In pool
on heat transfer in tube bundles using the dryness fraction and ana- boiling, according to the different regions in boiling curve
lyzed the corresponding local mechanism. It was found that the dis- (Nukiyama, 1966), there are natural convection regime, partial
ruption of the liquid boundary layer played a more critical role than nucleate boiling regime, fully developed nucleate boiling regime,
evaporation of the microlayer under sliding bubbles in the heat transition boiling regime and film boiling regime. Under the flow
transfer enhancement. Hahne and Müller (1983) and Hahne et al. boiling condition, with the increasing vapor phase quality, the boil-
(1991) studied saturated pool boiling over finned tube bundles. ing hydrocarbon vertical upward flow could transit from bubbly
Both the pool geometry and bundle configuration effects were dis- flow regime successively to intermittent flow regime and annular
cussed. Liu and Qiu (2004, 2006) investigated the boiling of pure dispersed flow regime in a horizontal tube bundle (Aprin et al.,
water and water-salt mixture over smooth tubes and enhanced 2007).
tubes in in-line and staggered bundles with different tube spacing. Various heat transfer mechanisms between the two-phase mix-
The bundle enhancement was observed on a compact tube bundle, ture and the wall, as well as between the two phases, depend on
and the boiling heat transfer rate increased with decreasing tube the flow regimes. For example, in the bubbly flow regime, the heat
spacing. Ribatski et al. (2008) recommended correlation for bundle transfer is governed by nucleate boiling on the heating surface;
effect in wide ranges of system parameters. Some representative while in the dispersed flow regime, the convective evaporation at
studies on boiling heat transfer in tube bundles are listed in Table 1. the liquid–vapor interface prevails and the bubble nucleation on
Many literature reviews related to boiling phenomena over a the heating surface is negligible (Aprin et al., 2011). Therefore,
single tube, cylindrical surfaces or tube bundles were published the regime-dependent heat transfer coefficient correlations
in the past, such as Collier and Thome (1994), Browne and Bansal together with two-phase flow regime criteria are adopted by some
(1999), Casciaro and Thome (2001), Ribatski and Thome, (2005, researchers, as listed in Table 2.
2007), and Swain and Das (2014). There are also some review Recently, Shah (2017) modified his previous dimensionless cor-
works released in the past decade that focused on some specialized relation (Shah, 2007) for local heat transfer coefficients during sat-
areas. Ciloglu and Bolukbasi (2015) reviewed the pool boiling of urated flow boiling in plain and enhanced tube bundles. Different
nanofluids. Abbas et al. (2017c) studied correlations for outside heat transfer regimes are distinguished by a boiling intensity
boiling of ammonia on single tube and bundle. Gorenflo et al. parameter YIB,MOD which is defined in Eq. (1), and the heat transfer
(2014) reviewed prediction methods for pool boiling heat transfer. coefficient was calculated by Eqs. (2)–(4):
Leong et al. (2017) summarized the boiling of dielectric fluids on
enhanced surfaces. The latest review regarding boiling over tube Y IB;MOD ¼ BoFr 0:3 ð1Þ
bundle was conducted by Swain and Das (2014) which intensively
discussed the effects of tube spacing, bundle geometry, surface
hTP ¼ F pb hcooper ð2Þ
characteristics and operating conditions on HTCs and bundle effect.
Swain and Das (2014) also reviewed the industrial scale studies on
boiling inside two-phase shell and tube heat exchangers. u ¼ u0 ð3Þ
Many studies with more profound and meticulous views on
boiling heat transfer over tube bundles were published during
2:3
the past couple of years. The previous reviews about bundle boiling u¼ 0:08
ð4Þ
mainly concentrated on experimental studies on horizontal bun- Z Fr 0:22
dles, while more investigations on boiling over vertical bundles, When YIB > 0.0008, the heat transfer is dominated by nucleate
inclined bundles, and special-shaped bundles were reported in boiling and hTP is calculated using (Eq. (17)). When YIB  0.0008,
recent years based on the increasingly diverse industrial applica- the heat transfer is governed by both nucleate boiling and convec-
tions. For the sake of a more comprehensive cognition of the boil- tion and hTP is determined by the maximum between (Eq. (2)), (Eq.
ing phenomena, updated information of the recent development in (3)) and (Eq. (4)). /0 is the value of / when x = 0. Fpb = hpb,actual/
the area of bundle boiling is required and will be helpful to the hcooper. hcooper is the Cooper (1984) correlation and hpb,actual is the
devising of future research work as well as the state-of-the-art same as hcooper unless pool boiling test data are available for the
design of efficient and economical heat transfer devices. This paper studied tubes or there is a reason to believe that another correla-
presents a review of studies published during the past five years on tion may be more appropriate than the Cooper correlation. Shah
pre-critical heat flux (pre-CHF) boiling heat transfer over tube bun- (2017) examined the newly modified correlation by 51 data sets
dles, covering natural and forced convection through multiple bun- and achieved good agreements with a mean absolute deviation of
dle geometries under a wide range of operating conditions. The 15.2%.
publications that have already been reviewed by Swain and Das Ribatski and Thome, (2005, 2007) summarized published inves-
(2014) and other authors will not be intensively discussed in the tigations on flow regime evaluation methods of horizontal tube
present review. bundles. These investigations were conducted through either
visual observations or time-series signals measurements. Owing
to the existence of tubes, the visualizing study of vapor–liquid
2. Flow patterns of two-phase flow in tube bundles involving two-phase flow patterns in a tube bundle is more difficult than
boiling the study of single tube external flow patterns or internal flow pat-
terns. The visual observation can only reveal the two-phase flow
The configuration of a tube array constrains the motion of gas near the outside shell wall, and different flow patterns can exist
bubbles as they rise and hit the tube bundle, and when boiling is near the shell wall and inside the tube bundle, as observed in
Table 1

4
Some previous studies on boiling in tube bundles.

Author Boiling mode Fluids Tube bundle characteristics Tube surface Operating conditions Heating method Max. HTC (kW/m2K)
characteristics
Webb and Chien (1994) Flow boiling R113, R123 6  3 staggered horizontal, Plain Tsat = 18.9, 37.8 °C, Electrically heated ~2.75 (R113)
D = 16.8 mm, P/D = 1.42 G = 0.28–40 kg/m2s, ~4.25 (R123)
x = 0.1–0.9
Gupte and Webb Flow boiling R11, R123, R134a 6  3 staggered horizontal, Finned, GEWA-SE Tsat = 4.4, 26.7⁰C, Electrically heated ~9 (R11, finned)
(1994, 1995a,b) D = 18.9 mm, P/D = 1.25 and Turbo-B q = 15–45 kW/m2, ~18.6 (R134a, GEWA-SE),
G = 7–30 kg/m2s, ~13.5 (R11, Turbo-B),
x = 0–0.9 ~10.5 (R123, Turbo-B)
Dowlati et al. (1996) Flow boiling R113 20  5 in-line horizontal, Plain Tsat = 48-61 °C, In-tube hot oil ~5.5
D = 12.7 mm, P/D = 1.3. q = 0–8 kW/m2,
G = 50–790 kg/m2s,
x = 0–0.5.
Ishibashi (2001) Pool boiling Water/salt mixture 7  3 staggered horizontal, Plain 1 atm, Electrically heated ~80 (P/D = 1.03)
D = 18 mm, P/D = 1.006, 1.03, 1.2 q = 0–140 kW/m2
Kim et al. (2002) Flow boiling R123, R134a 5  3 staggered horizontal, Surface with pores Tsat = 4.4, 26.7 °C, Electrically heated ~28.5 (R134a),
D = 18.8 mm, P/D = 1.26 and connecting q = 10–40 kW/m2, ~9.1 (R123)

S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

gaps G = 8–26 kg/m2s,
x = 0.1–0.9
Hsieh et al. (2003) Pool boiling R134a 5  3 in-line and staggered Plasma coated P = 537.06 kPa, Electrically heated ~7 (staggered)
horizontal, D = 20 mm, P/D = 1.5 copper q = 0.1–30 kW/m2
Liu and Qiu (2004) Pool boiling Water/salt mixture 6  3 staggered horizontal, Plain and roll- 1–3 atm, Electrically heated ~22 (P/D = 1.06, 3 atm)
D = 18 mm, P/D = 1.008–1.2 worked q = 0–140 kW/m2 ~16 (P/D = 1.06, 1 atm)
Robinson and Thome (2004a) Flow boiling R134a 8  3 staggered horizontal, Plain Tsat = 4.45 °C, In-tube hot water ~9.7
D = 19 mm, P/D = 1.17 x = 0.1–0.87,
q = 2–35 kW/m2,
G = 5–41 kg/m2.s
Robinson and Thome (2004b) Flow boiling R134a, R507a 8  3 staggered horizontal, Low finned Tsat = 4.4 °C (R134a), In-tube hot water ~11.5 (R507a)
D = 19 mm, P/D = 1.17 Tsat = 4.7 °C (R507a), ~17 (R134a)
x = 0.08–0.82,
q = 2–50 kW/m2,
G = 3–29 kg/m2.s
Robinson and Thome (2004c) Flow boiling R134a, R410a, 8  3 staggered horizontal, Turbo-BII Tsat = 4.4 °C (R134a), In-tube hot water ~25.5 (R507a)
R507a D = 19 mm, P/D = 1.17 Tsat = 4.7 °C (R507a), ~33 (R134a)
Tsat = 4.56 °C (R410a), ~29 (R410a)
x = 0.08–0.78,
q = 8–64 kW/m2,
G = 4–38 kg/m2.s
Burnside and Shire (2005) Flow boiling R113 17  5 in-line horizontal, D = 19 mm, Plain 1 atm, Electrically heated ~3.3
P/D = 1.34. q = 10–65 kW/m2,
Remax = 7800–27000
Gupta (2005) Pool & flow Water 5  3 in-line horizontal, D = 19 mm, Plain 1 atm, Electrically heated ~16 (pool boiling)
boiling P/D = 1.5 q = 10–40 kW/m2, ~13.5 (flow boiling)
G = 1–10 kg/m2s
Liu and Qiu (2006) Pool boiling Water 6  3 in-line horizontal, D = 18 mm, Plain 1–3 atm, Electrically heated ~30 (P/D = 1.1)
P/D = 1.008–1.2 q = 0–140 kW/m2
Liu and Liao (2006) Pool boiling Water 5  3 in-line horizontal, D = 18 mm, Plain 20, 50 and 100 kPa, Electrically heated ~31 (100 kPa, P/D = 1.1)
P/D = 1.03–1.2 q = 0–100 kW/m2 ~14.5 (50 kPa, P/D = 1.1)
~8 (20 kPa, P/D = 1.1)
Liao and Liu (2007) Pool boiling Water 5  3 staggered horizontal, Plain 20, 50 and 100 kPa, Electrically heated ~21 (100 kPa, P/D = 1.02)
D = 18 mm, P/D = 1.03–1.2 q = 0–100 kW/m2 ~12 (50 kPa, P/D = 1.06)
~10 (20 kPa, P/D = 1.06)
Schäfer et al., (2007) Pool boiling R134a 2  1 and 2  2 in-line horizontal, Plasma coated Tsat = -20- 20 °C, In-tube hot oil ~11.7 (Tsat = 20 °C)
D = 18 mm, P/D = 1.33 copper q = 2–100 kW/m2
Ribatski et al. (2008) Pool boiling R123 3  1 in-line horizontal, Plain Pr = 0.022–0.64, Electrically heated ~2.2
D = 19.05 mm, P/D = 1.32,1.53 and 2 q = 0.5–40 kW/m2
 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278 5

Noghrehkar et al. (1999). Therefore, using visual observation alone

is not sufficient for flow regime map development.
In order to obtain the two-phase flow regime map in an objec-
Max. HTC (kW/m2K)

~3.6 (n-pentane)
tive manner, a void fraction is usually employed to determine the
~11 (iso-butane)

~11 (P/D = 1.4)

local flow pattern, such as the probability density function (PDF)
~13 (propane)

~9.8 (R134a)
~6 (R236fa)
approach used in Hahne et al. (1990) and Noghrehkar et al.
(1999). Kanizawa and Ribatski (2016) studied the flow patterns

~5.5
during air–water upward flow across a horizontal triangular tube
bundle using the k-means clustering method. The flow regimes,
as shown in Fig. 1, were identified based on both pressure drop
In-tube hot water-glycol

and void fraction measured by a differential pressure transducer

In-tube hot water flow

and a capacitive sensoring system, respectively.

Electrically heated
Electrically heated
Various methods had been employed to measure the void frac-
Heating method

tion in tube bundles, such as the quick closing valve (QCV) tech-
nique used by Xu et al. (1998), gamma-ray densitometry used by
Dowlati et al. (1990, 1992), resistive probes used by Noghrehkar
flow

et al. (1999), optical probes used by Taylor and Pettigrew (2001),

capacitive sensors used by Kanizawa and Ribatski (2017a) and X-
Tsat = 5, 10, 15 and 20 °C,
iso-butane: P = 4–6 bar,

1 atm, q = 2–32 kW/m2

ray radiography used by Murakawa et al. (2018). Furuya et al.

propane: P = 6–12 bar,
Operating conditions

(2019) recently using X-ray CT system measured three-

q = 13–44 kW/m2;
G = 15–44 kg/m2s,

q = 12–45 kW/m2
q = 10–52 kW/m2
q = 3–53 kW/m2;
G = 9–45 kg/m2s,

G = 8–44 kg/m2s,

G = 4–36 kg/m2s,

dimensional void fraction distributions inside a 5  5 vertical rod

q = 0–21 kW/m2,
P = 0.2–0.5 bar,

bundle during pool boiling of water, condensed seawater and the

x = 0.1–0.9.
n-pentane:

mixture solution of seawater and borated water. The bubbles are

smaller in the salt waters than those in water. No obvious differ-
1 atm,

ence in the void fraction distributions was observed between the

condensed seawater and the mixture solution. For the horizontal
void distribution in the downstream region, the void fraction is
characteristics

higher in the central area than that near the shell wall for all three
Coated SS316
Tube surface

fluids. The incipience of boiling is shifted toward downstream, and

the void swell level is decreased in the condensed seawater and
Plain

Plain

Plain

mixture solution. The void fraction can also be predicted by the

correlations established based on Lockhart and Martinelli (1949)
7 vertical tubes, D = 19 mm, P = 1.66

parameter, as introduced in Kanizawa and Ribatski (2017b). The

D = 19.05 mm, P/D = 1.4, 1,7 and 2.0

value of vapor quality is required and can be determined from flow

rates measured using flowmeters.
Tube bundle characteristics

9  5 staggered horizontal,

8  3 staggered horizontal,

Commonly, the flow regimes were given as bubbly, intermittent

8x3 staggered horizontal,
D = 18.95 mm, P = 1.17
D = 19.05 mm, P = 1.33

and annular flows with some discrepancies, as shown in Fig. 1. Most

of the existing flow regime maps were developed for the air–water
fluids system while minimal attention has been paid to two-phase
flow involving boiling. Aprin et al. (2007) performed an experimen-
tal investigation on flow regimes identification during hydrocar-
bons two-phase upward flow boiling across a horizontal tube
bundle using PDF method. They (Aprin et al., 2007) found that the
transitions between flow patterns were independent of the liquid
superficial velocity. Three distinct flow regimes were classified such
as bubbly, intermittent and annular dispersed flows. Van Rooyen
R134a, R236fa

(2011) developed some visualized flow regime maps for both boil-
n-pentane,

iso-butane
propane,

ing refrigerants, and air–water flows in a regular triangular array

Water
Water
Fluids

using a borescope. Three flow regimes were distinguished as bub-

bly, intermittent and annular flows under adiabatic and diabatic
conditions. Compared to the flow pattern map identified by
Ulbrich and Mewes (1994) for upward air–water cross-flows in a
Boiling mode

Flow boiling

Flow boiling

Pool boiling
Pool boiling

horizontal tube bundle, it is observed that the flow regime transi-

tion boundaries for the boiling hydrocarbons and refrigerants are
different from those for air–water systems, as shown in Fig. 2.
Recently, Mao and Hibiki (2017) studied flow regime transition
criteria for upward two-phase cross-flow in horizontal tube bun-
dles by assuming the similarity between external and internal flow
Van Rooyen et al. (2012)
Aprin et al. (2007, 2011)

regime transition mechanisms and identified five different flow

Lakhera et al. (2012)

regimes as bubbly, cap bubbly, churn, finely dispersed bubbly

Gupta et al. (2010)

and annular flows. The effects of bubble nucleation on the diabatic

Table 1 (continued)

wall were considered in the flow regime transition, while with the
vapor void fraction increasing, these effects will become insignifi-
Author

cant. Moreover, the nucleation effects during the transition from

bubbly flow to cap bubbly flow are inconsiderable, because the pri-
mary mechanism is the steep rise of bubble coalescence rate at
6 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

Table 2
Regime-dependent heat transfer coefficient correlations.

Authors Characteristics Correlations

Jensen and Hsu (1987) R113, Superposition; n ¼ 1 C ¼ C 1 FrC 2 ln ðX tt Þ þ C 3 Fr C 4
2–5 bar, 00 B m_
Fr ¼ pffiffiffiffiffiffiffiffi
hnpb ¼ Aq qL gD
x = 0–0.36,
q00 = 1.6–44.1 kw/m2, hnpb from experimental data on a F 1 ¼ 0:678A0:354
G = 50–675 kg/m2s single tube with a simple power cur v e F 2 ¼ 1:86A1:93

13
NuCVL ¼ 0:137Re0:692
L Pr 0:34
L q
1=2
jL lL
A¼ B ¼ jL lqL
L
r


 L

S¼ kL
ð1  exp Fhcv L Y 0 ifB  F 1 ; Spray flow :
Fhcv L Y 0 kL
m
F ¼ ð/2L Þ2n m ¼ 0:692 C 1 ¼ 0:253; C 2 ¼ 1:50; C 3 ¼ 12:4;
n ¼ 0:674; if 2000 < ReL < 7000 C 4 ¼ 0:207; C 5 ¼ 0:205
n ¼ 0:191; if 7000 < ReL < 20; 000 ifB < F 1 & B < F 2 ; Slug flow :
 1:8
qG 
lL 0:2 C 1 ¼ 2:18; C 2 ¼ 0:643 C 3 ¼ 11:6;
/2L ¼ 1 þ XCtt þ XC25 X 2tt ¼ 1x
x qL lG
tt

0:5 C 4 ¼ 0:233; C 5 ¼ 1:09
Y 0 ¼ 0:041 gðq q Þr
L G

if B < F 1 & B  F 2 ; Bubble flow :

C 1 ¼ 0:036; C 2 ¼ 1:51; C 3 ¼ 7:79;
C 4 ¼ 0:057; C 5 ¼ 0:774
Webb and Chien (1994) R113, R123, Asymptotic, n = 3, S = 1, Superposition, n = 1 /2L ¼ 1 þ X8tt þ X12
tt
0.45–0.7 bar (R113),  1:8
qG 
lL 0:2
hnpb from cur v e fits of experimental
0.73–1.5 bar (R123), X 2tt ¼ 1x
x q l L G
x = 0.1–0.9, data on a single tube if ReL < 1000
G = 8–40 kg/m2s, hCVL  c ¼ 0:9; m ¼ 0:50; b ¼ 0:474
Žukauskas, (1972)

0:25
if 2000 < ReL < 7000
Nucv L ¼ cRem 0:36 PrL
L Pr L PrW




S ¼ FhckvLL Y 0 1  exp FhkcLv L Y 0 c ¼ 0:27; m ¼ 0:63; b ¼ 0:674
h 2 im if 7000 < ReL < 20; 000
/ ðPr þ1Þ 2b
F ¼ L 2L

0:5
c ¼ 0:27; m ¼ 0:63; b ¼ 0:191
Y 0 ¼ 0:041 gðq rq Þ
L G

if 20; 000 < ReL < 30; 000

c ¼ 0:27; m ¼ 0:63; b ¼ 0:182
Shah (2007) P = 0.3–7.8 bar, Pr = 0.005–0.189, Boiling intensity parameter: u0 ¼ maxð443Bo0:65 F pb ; 31Bo0:33 F pb ; 1Þ
G = 1.6–1391 kg/m2s, u ¼ hTP =hLT
Y IB ¼ F pb BoFr 0:3
q00 = 1–1000 kW/m2
F pb ¼ hpb;actual =hcooper ¼ 0:21Re0:62
hLT D
L Pr 0:4
kL
hcooper from Cooper ð1984Þ 1x0:8 0:4
Z¼ x Pr
Y IB > 0:0008; intense boiling reime :
hTP ¼ F pb hcooper
0:00021 < Y IB  0:0008;
conv ectiv e boiling regime :
Y IB  0:00021; conv ection regime :
u ¼ Z0:082:3
Fr 0:22
jg < 0:15m:s1
nucleate boiling dominant; h1¼ Cooper ð1984Þ
Aprin et al. (2011) n-pentane, propane, jg > 0:35 m:s1 : ReG ¼ V G qlG Dext:
G
iso-butane
conv ectiv e boiling dominant; _
V G ¼ qmx
Ge
lG CpG
Nu ¼ h2kDGext: ¼ 387p 0:17 Re0:34 0:33
G Pr G PrG ¼ kG
1 1
x=qG 
0:15m:s < J G < 0:35 m:s e¼
1:047 qx þ1x
qL þ m_
0:23
G

is intermediate regime; h3 ¼ maxðh1 ; h2 Þ

maximum packing of bubbles. Another surge in bubble coalescence bubbly” flow regime was not considered in Van Rooyen (2011). The
rate at maximum packing of small bubbles and cap bubbles indi- intermittent to annular flow transitions in the two criteria agree
cates the transition from cap bubbly flow to churn flow. Then if well with each other, but there is a certain discrepancy of the tran-
intensive turbulence is imposed on the two-phase flow, the flow sition from bubbly flow to intermittent flow in the two criteria
pattern will transit to finely dispersed flow. If ‘‘flow reversal” or (Fig. 3a). The prediction of Mao and Hibiki (2017) was not in coin-
‘‘onset of entrainment” occurs, the flow pattern will transit to cidence with Aprin et al. (2007) data (Fig. 3b), because the super-
annular flow. Mao and Hibiki (2017) compared the newly devel- ficial vapor velocity in local flow regime adopted by Aprin et al.
oped criteria with 12 published flow regime maps measured in (2007) was different from that in the global flow regime used by
the horizontal tube bundles, including the flow pattern maps pre- other researchers.
sented by Aprin et al. (2007) and Van Rooyen (2011), as shown in Regarding two-phase flow in vertical bundles, Williams and
Fig. 3. The ‘‘intermittent” flow regime in the Van Rooyen map (Van Peterson (1978) experimentally investigated upward flow of boil-
Rooyen, 2011) corresponds to the ‘‘cap bubbly” and ‘‘churn” flow ing water in a 1  4 heated rod bundle under high pressure of
regimes in Mao and Hibiki (2017) criteria and the ‘‘finely dispersed 2.76–13.79 MPa. Four flow patterns were identified, i.e., bubble
 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278 7

vation and multi-scale entropy analysis. Slug flow was not

observed under the experimental conditions. Lu et al. (2019)
observed four types of flow patterns in the water pool boiling in
a vertical tube bundle under flooded condition, i.e., bubble flow,
slug flow, churn flow and annular flow. Moreover, they found that
in one cross-section of the bundle at the same time, slug flow pre-
vails in the middle region while bubble flow dominates the outer
region. The authors attributed this phenomenon to the higher
heating intensity and the larger void fraction in the middle region
of the bundle. Liu and Hibiki (2017) developed new flow regime
transition criteria for upward two-phase flow in vertical rod bun-
dles based on a drift–flux model. They classified the flow structure
into six distinct flow regimes, i.e., bubbly, finely dispersed bubbly,
cap bubbly, cap turbulent, churn, and annular flows. The authors
pointed out that there is no slug bubble formation in the vertical
rod bundle because of the corresponding surface instability. Com-
pared with Venkateswararao et al. (1982) two-phase flow transi-
tion criteria which was developed from air-water flows, Liu and
Hibiki (2017)’s criteria achieved better agreement with the exper-
imental flow map data of Zhou et al. (2015), except a slight overes-
timation of the transition vapor superficial velocity from churn
flow to annular flow, as shown in Fig. 4.
From the above review, the objective approaches are suggested
Fig. 1. Schematics of flow patterns. (a) Bubbles, (b) Large bubbles, (c) Churn, (d) to determine the two-phase flow patterns in tube bundles. There
Dispersed bubbles, (e) Intermittent, (f) Annular. From Kanizawa and Ribatski are still no general criteria applicable to distinguish the two-
(2016).
phase flow regimes inside a tube bundle involving boiling. More
experimental database involving boiling two-phase flow over var-
ious bundle geometries and more test fluids (hydrocarbons and
refrigerants) are still required to build a deeper understanding of
flow boiling regimes transition over tube bundles.

3. Boiling heat transfer over plain tube bundles

In this section, investigations concerning pool boiling and

forced flow boiling heat transfer of common fluids (excluding
metal fluids) over plain tube/rod bundles are critically discussed.
The main characteristics of the reviewed experimental studies
are described in Table 3. It has been quite well established in the
literature that boiling heat transfer coefficient generally increases
with the increasing heat flux and the HTC increases from the bot-
tom tube towards the top tube in a horizontal bundle (Ribatski and
Thome, 2007; Swain and Das, 2014). Researchers have recently
placed more focus on the local boiling heat transfer characteristics
in bundle boiling under a wide range of geometric and operating
parameters.

3.1. Bubble departure characteristics in bundle boiling

The enhanced heat transfer in bundle boiling can be attributed

Fig. 2. Flow pattern map comparison between Aprin et al. (2007), Van Rooyen to the bubble motion and liquid convection originated from the
(2011) (adiabatic flow and diabatic flow), and Ulbrich and Mewes (1994). bottom tubes. The bundle effect, in turn, could influence the bubble
departure characteristics which are commonly represented by
bubble departure diameter, bubble departure frequency, and wait-
flow, froth flow, slug flow, and annular flow. Kaichiro and Ishii ing period. Generally, bubble departure diameter and frequency
(1984) developed new flow-regime criteria for upward gas–liquid increase with the heat flux and wall superheat but decrease with
flow in a vertical tube bundle by using relative velocity correla- the liquid subcooling, pressure, and mass flux (Mohanty and Das,
tions. The authors pointed out that direct geometrical parameters 2017). Several correlations were developed by researchers for bub-
such as the void fraction should be used for the two-fluid model ble departure diameter on a single heated surface based on the
formulation, especially for a subcooled boiling system. Zhou et al. consideration of the fluid thermal properties, force balance, operat-
(2015) studied the two-phase flow patterns of vertical steam- ing conditions and surface characteristics, such as Fritz (1935),
water flows in a 3  3 rod bundle arranged on a square pitch under Cole and Rohsenow (1969), and Kocamustafaogullari and Ishii
atmospheric pressure and relatively low mass flow condition. They (1983). Bubble departure parameters are usually measured using
classified the flow structure into four distinct flow regimes such as high-speed cameras and analyzed with the assistance of image
bubbly, bubbly-churn, churn and annular flows using visual obser- processing software. Due to the difficulty of getting images in the
8 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

(a) (b)

Fig. 3. Flow pattern map comparison between (a) Van Rooyen (2011) and Mao and Hibiki (2017); (b) Aprin et al. (2007) (propane system) and Mao and Hibiki (2017).
Adapted from Mao and Hibiki (2017).

Fig. 4. Flow pattern map comparison between Liu and Hibiki (2017), Venkateswararao et al. (1982), and Zhou et al. (2015) flow regime map data. Adapted from Liu and Hibiki
(2017).

tube bundle, the bubble departure parameters in bundle boiling quency on the only heated tube in the bundle is significantly
have rarely been reported. higher than that on an isolated single tube. The possible reason
Goel et al. (2018) recently investigated vapor bubble departure for this phenomenon is the existence of the tubes above the
characteristics in saturated nucleate boiling in a 5  3 staggered observed tube impedes the rising of convective plumes and thus
tube bundle. Four test cases were conducted by applying heat flux slightly increases the liquid temperature around the heated tube.
to only one isolated single tube and 1, 2 and 3 tubes in the bundle, Hotter liquid reduces the formation cycle of the thermal boundary
and then the bubble departure characteristics on the third tube layer on the heated tube and decreases the waiting period. Empir-
(from the top) and the isolated tube were measured, as shown in ical correlations were developed for bubble departure diameter
Fig. 5. The departure diameter and frequency were both found to and frequency incorporating the tube bundle effects as follows:
increase with the increasing heat flux, and wall superheat but to
decrease with the increasing mass flux, as depicted in Figs. 6 and
gD2d ðqL  qG Þ
7. Moreover, both the bubble departure diameter and frequency Bd ¼ ¼ 0:0056na Ja1:85 Pr2:5 ð1 þ ReÞ0:036 ð5Þ
were observed to increase with the increasing heated tube number
r
in its neighborhood. As regards the departure diameter, this is
because the heated tube nearby increases the bulk liquid temper- pffiffiffiffiffiffi  1=2
4g ðqL  qG Þ
ature and thus facilitates the growth of a bubble cap. The enhance- f Dd ¼ 0:0004nb Ja2:05 ð1 þ ReÞ0:031 ð6Þ
3qL
ment in bubble departure frequency is also due to the temperature
increase of the bulk liquid between the tubes, which speeds up the
re-establishment of the thermal boundary layer and decreases the n is the number of the heated tube in the bundle. However, the cor-
waiting period. When only one tube is heated in the bundle, the relations proposed are highly dependent on operating conditions,
bubble departure diameter is equivalent to the value under one heating surface characteristics, bundle arrangement, and even the
isolated single tube condition. However, the bubble departure fre- position of the studied tube in the bundle.
Table 3
Studies on boiling over plain tube bundles reviewed in Section 3.

Author Boiling mode Fluids Tube bundle characteristics Operating conditions Heating method Max. HTC (kW/m2K)
Lakhera et al. (2014) Pool boiling Water 8x3 staggered horizontal, 1 atm, q00 = 12–45 kW/m2 Electrically heated ~11 (P/D = 1.4)
D = 19.05 mm, P/D = 1.4, 1.7 and 2.0,
plain and coated
Chung et al. (2015) Pool boiling .Water 7 vertical tubes, D = 21.3 mm, 1 atm, Tinitial = 30 , 50 , and 70 , q00 = 39.8–99.6 kW/m2 Electrically heated Not provided
P/D = 1.53
Kang (2015a,b, 2016) Pool boiling Water 2 tandem tubes, D = 19 mm, 1 atm, q00 = 0–120 kW/m 2
Electrically heated ~18 (P/D = 1.5)

S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

P/D = 1.5–6
McNeil et al. (2015) Pool boiling Water two 6  3 inline horizontal bundle, 850 and 50 mbar, q00 = 10–70 kW/m2 Electrically heated ~6.1 (HP)
D = 28.5 mm, Ph/D = 2.42, Pv/D = 2.19 ~5.1 (LL LP)
~4 (HL LP)
Abbas and Ayub (2017) Flow boiling Ammonia 4-pass x3 triangular tubes, Tsat = 1.7 to 20 , q00 = 5–45 kW/m2 Hot water/ ~6 (Tsat = -1.7 )
horizontal, D = 19.1 mm, P/D = 1.25 glycol + modified
Wilson plot method
Ayub et al., (2017) Abbas Expansion Ammonia 4-pass x3 triangular tubes, Tsat = 1.7 to 20 , q00 = 5–45 kW/m2, Tsup = 2–10 Hot water/ ~4.6 (Tsat = -1.7 ,
et al., (2017a, 2017b) evaporation horizontal, D = 19.1 mm, P/D = 1.25 glycol + modified xin = 0.15, Tsup = 2 )
Wilson plot method
Kang (2017, 2018) Pool boiling Water 2 V-shape tubes, D = 19 mm, 1 atm, q00 = 0–120 kW/m2 Electrically heated ~18 (a = 2°)
Pave/D = 0.39–3.46 ~16 (/ = 90°)
00 2
McNeil et al. (2017) Pool boiling Glycerol- two 6  3 inline horizontal bundle, 50 mbar, q = 10–65 kW/m Electrically heated ~1.7
water D = 28.5 mm, Ph/D = 2.42, Pv/D = 2.19
Swain et al. (2017) Pool boiling Water 5  3 staggered horizontal, 1 atm, q00 = 12.28–37.63 kW/m2 Electrically heated ~4.3 (P/D = 1.25)
D = 20 mm, P/D = 1.25, 1.6 and 1.95
Swain and Das (2017) Flow boiling Water 5  3 staggered horizontal, 1 atm, q00 = 12.28–37.63 kW/m2, G = 20.25–192.9 kg/m2s Electrically heated ~4.3 (P/D = 1.25)
D = 20 mm, P/D = 1.25, 1.6 and 1.95
00 2 2
Wang et al. (2017) Flow boiling Water 2  2 inline vertical, Dequi. = 5.32 mm 11–19 Mpa, q = 200–600 kW/m , G = 700–1300 kg/m s Electrically heated ~61 (P = 19 MPa)
Zhang et al. (2017a) Flow boiling Water 7 vertical rods, D = 17.53 mm, 109–194 kPa, q00 = 50 kW/m2, G = 30–120 kg/m2s, Electrically heated ~14.5 (xin = 0.1,
P/D = 1.42 xin = 0.007–0.5 G = 120 kg/m2s)
Zhang et al. (2017b) Pool boiling Water 12 C-shape rods, D = 19 mm, P/D = 2 1 atm, q00 = 36–120 kW/m2 Electrically heated ~32 (vertical part)
Goel et al. (2018) Pool boiling, flow Water 5  3 staggered horizontal, 1 atm, q00 = 8–28 kW/m2, G = 120–303.5 kg/m2s, Electrically heated ~1.8
boiling D = 20 mm, P/D = 1.95 DTsat = 7–22 K
Zhang et al. (2018) Flow boiling Water 4  16 staggered horizontal bundle, 115–190 kPa, q00 = 20–55 kW/m2, G = 15–50 kg/m2s, Electrically heated ~9.9
D = 17.48 mm, P/D = 1.42 xout = 0.02–0.57
Lu et al. (2019) Pool boiling Water 9  9 Vertical tubes, D = 9.6 mm, q00 = 0.6–2.5 kW/m2, Electrically heated ~0.2 (q = 1.43 kW/m2)
P/D = 1.35, L = 4700 mm
Zhang et al. (2019) Flow boiling Water 4  25 staggered 45° inclined bundle, 117–260 kPa, q00 = 20–55 kW/m2, G = 15–50 kg/m2s, Electrically heated ~8.5
D = 17.48 mm, P/D = 1.42 xout = 0.05–0.79

9
10 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

tubes activated to the HTC for this upper tube activated alone in
the bundle. Kang (2015b) adopted this definition in the study on
saturated pool boiling of two tandem tubes horizontally placed
one above another and observed that when the two tubes were
applied the same heat flux, bundle effect decreased as the heat
fluxes increased and finally converged to 1. This phenomenon is
consistent with the results of the previous investigations that the
bundle effect generally vanishes at higher heat fluxes, owing to
the bubble coagulation and vapor blanketing on the upper tubes
(Swain and Das, 2014). When different heat fluxes were imposed
Fig. 5. Schematic of test cases in Goel et al. (2018). on the two tubes, throughout the heat fluxes tested, the increase
in heat flux of the lower tube increases the bundle effect. This bun-
dle effect is significant when the heat flux of the lower tube is lar-
5 ger than that of the upper tube and when the heat flux of the upper
Case 1, 59.14kg/m2s tube is less than 60 kW/m2, because the enhanced bubble nucle-
Case 2, 121.40kg/m2s ation on the lower tube and the consequential intensive bubbly
Case 3, 121.40kg/m2s flow prevail the two-phase heat transfer on the upper tube.
Case 4, 121.40kg/m2s Besides, when the same heat flux was applied to the two tubes,
Buble departure diameter (mm)

4
the bundle effect reached the minimum value.
Lakhera et al. (2014) studied the local heat transfer coefficient
of pool boiling in a bundle and found it to increase with the
3 increasing heat flux. A significant bundle effect was observed, as
shown in Fig. 8. The bundle effect (Memory et al. (1994) definition)
reached its maximum value of 5.83 on the seventh tube from the
bottom when heat flux was 14.1 kW/m2, and pitch to diameter
2
ratio was 1.4. The maximum enhancement of bundle average
Case 1, 147.84kg/m2s HTC (hbundle/hbottom) was 3.7 and occurred under the same condi-
Case 2, 303.51kg/m2s tion. The largest HTC did not show up on the topmost tube (the
Case 3, 303.51kg/m2s eighth) but on the seventh tube, possibly because the bubble
Case 4, 303.51kg/m2s plumes leave the upper portion of the topmost tube without the
1
5 10 15 20 25 confining of other tubes. Moreover, the heat transfer on the higher
Wall superheat (K) tubes is less sensitive to the heat flux variation, and this phe-
nomenon may be attributed to the decrease of bundle effect under
Fig. 6. Buddle departure diameter under different wall superheat and mass flux. higher heat fluxes.
Adapted from Goel et al. (2018). Lakhera et al. (2014) also investigated the effect of circumferen-
tial variation on the local heat transfer coefficient. The lowest HTC
is on the tube upper surface due to the coalescence of bubbles near
Case 1, 59.14kg/m2s the top surface, and the highest HTC is neat the tube lower surface
60
Case 2, 121.40kg/m2s owing to the striking bubbles from the lower tubes, which is con-
Case 3, 121.40kg/m2s sistent with the results in Hsu et al. (1993). Gupta (2005) also
50 Case 4, 121.40kg/m2s obtained the same trend of HTC circumferential distribution in a
Buble departure frequency (Hz)

5  3 in-line tube bundle, as illustrated in Fig. 9. The circumferen-

40 tial variation of HTC is more significant on the top tube than that

30

20

10 Case 1, 147.84kg/m2s
Case 2, 303.51kg/m2s
Case 3, 303.51kg/m2s
0
Case 4, 303.51kg/m2s

4 6 8 10 12 14 16 18 20 22
Wall superheat (K)

Fig. 7. Buddle departure frequency under different wall superheat and mass flux.
Adapted from Goel et al. (2018).

3.2. Bundle effect in horizontal bundles

Bundle effect is the main characteristic of boiling heat transfer

in tube bundles and a significant concern in the design of heat
transfer exchangers. Memory et al. (1994) defined bundle effect Fig. 8. Heat transfer coefficients along the height of plain tube bundles with data
as the ratio of the HTC for an upper tube in a bundle with all the from Lakhera et al. (2014) (P/D = 2) and Swain et al. (2017) (P/D = 1.95).
 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278 11

geometries and tube surface characteristics. Besides, the heat

1.6 transfer on the topmost tube also exhibited less sensitivity to the
change of heat flux compared to the middle three tubes. However,
1.4 different from the results from Lakhera et al. (2014), the increase of
heat flux did not enhance the HTC on the bottom tube. In the water
1.2 pool boiling data of an inline 5  3 bundle published in Gupta
(2005), there was a peak HTC value under heat flux of about
h/have

1.0 23 kW/m2, as shown in Fig. 10, but this phenomenon did not occur
in the studies of Lakhera et al. (2014) and Swain et al. (2017). This
0.8 is probably because the bundle layout in Gupta (2005) was in-line
while in the other two were staggered. Furthermore, under the
34.2 kW/m2, third tube from the bottom,
0.6 similar heat transfer structure and operating conditions, the HTC
(Lakhera et al., 2014)
results in Swain et al. (2017) were lower than those in Gupta
31.52 kW/m2, bottom tube,(Gupta, 2005)
(2005), which was possibly because in-line bundle layout exhibits
0.4 31.52 kW/m2, top tube,(Gupta, 2005)
better heat transfer performance than the staggered layout, consis-
0 90 180 270 360 tent with the comparison results in Liu and Liao (2006).
Angular position (degree) Swain et al. (2017) applied variable heat flux along the height of
a horizontal tube bundle to study the corresponding effects on heat
Fig. 9. The circumferential variations of the HTCs on a tube in horizontal bundles in transfer. Six different heat flux arrangements were applied to the
Lakhera et al. (2014) (P/D = 1.4) and Gupta (2005) (P/D = 1.5). bundle, including 3 RUNs with increasing heat flux from top to bot-
tom and another 3 RUNs with decreasing heat flux from top to bot-
tom (Table 4). When the lower tubes owned the higher heat flux
on the bottom tube due to the intense heat transfer and bubble
than the upper tubes in RUN 1–3, the high heat flux enhanced
motion on the top tube. The results reported in Lakhera et al.
the bubble nucleation on the lower tubes and consequently
(2014) were measured on the third tube from the bottom, and
strengthened the bundle effect. As shown in Figs. 11 and 12, in
the HTC range on this tube is between the HTC ranges on the top
RUN 1–3, the superheats of the upper tubes are lower than those
tube and the bottom tube in Gupta (2005).
under uniform heat flux conditions and RUN 4–6 and the HTCs
Swain et al. (2017) conducted a pool boiling experiment under a
on the uppers tubes are higher than those under the other cases,
similar heat flux range as Lakhera et al. (2014). The results showed
which is in coincidence with the results drawn from the two tan-
that the bundle effect played a significant role, as shown in Fig. 8.
dem tubes system in Kang (2015b). On the contrary, the increasing
The HTCs in Swain et al. (2017) were lower than the results from
heat flux from top to bottom results in the superheats on the lower
Lakhera et al. (2014), which is possibly due to the different bundle
tubes higher than those under uniform heat flux conditions and
RUN 4–6. Therefore, the decreasing heat flux arrangement from
top to bottom owns lower wall superheat range over the entire
bundle. Since a fewer number of bubbles will be generated on
the lower tubes under lower heat flux, the decreasing heat flux
arrangement from top to bottom contributes less to the bundle
effect, as shown in Fig. 12.
The variable heat fluxes (Table 4) were applied to the study on
flow boiling in Swain and Das (2017). The authors pointed out that
the heat flux would not remain constant in actual situations of the
two-phase shell and tube heat exchanger because of the shell side
cold liquid motion and bubbles striking to the upper tubes. Similar
to the pool boiling results, a noticeable bundle effect was observed
under both uniform and variable heat flux conditions. Bundle-
averaged HTCs were found to be higher under decreasing heat
fluxes from bottom to top rows than uniform heat flux condition
due to the enhanced bundle effect by strong bubble nucleation
on the lower tubes. Moreover, the runs with increasing heat fluxes
from bottom to top rows owned lower bundle averaged HTCs com-
pared to the HTC values under the corresponding uniform heat
fluxes, because of vapor blanketing effect caused by the coagula-
tion of bubbles from bottom tubes, which deteriorated the bundle
Fig. 10. Heat transfer coefficients along the height of plain tube bundles with data effect. The local superheat on each tube generally increases with
from Gupta (2005) (P/D = 1.5) and Swain et al. (2017) (P/D = 1.25). the increasing mass flux since the large fluid velocity interrupts

Table 4
Variable heat fluxes arrangement along the row height in Swain et al. (2017). (kW/m2).

Row 1 Row 2 Row 3 Row 4 Row 5 Ave. of the RUN Nearest uniform heat flux
RUN I 12.28 15.58 19.27 23.23 27.69 19.61 19.27
RUN2 15.58 19.27 23.23 27.69 32.63 23.68 23.23
RUN 3 19.27 23.23 27.69 32.63 37.83 28.13 27.69
RUN 4 27.69 23.23 19.27 15.58 12.28 19.61 19.27
RUN 5 32.63 27.69 23.23 19.27 15.58 23.68 23.23
RUN 6 37.83 32.63 27.69 23.23 19.27 28.13 27.69

Row 1 is the top row, and row 5 is the bottom row.

12 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

16 12
RUN 1 row 1 (top) RUN 4
RUN 2 row 2
14 RUN 3
row 3 12.28 kW/m 2
RUN 4
10 row 4
12 RUN 5
row 5 (bottom)
Wall superheat (K)

RUN 6

Wall superheat (K)

uniform 15.58 kW/m 2
10
8 19.27 kW/m 2

6 23.23 kW/m 2
6

4
27.69 kW/m 2
2 4

1.0x104 1.5x104 2.0x104 2.5x104 3.0x104 3.5x104 4.0x104

0 20 40 60 80 100 120 140 160 180 200
2
Heat flux (W/m ) 2
Mass flux (kg/m s)
Fig. 11. Variation of wall superheats in a horizontal bundle under multiple heat flux
Fig. 13. Variations of wall superheat with mass flux in RUN4 for P/D = 1.25. Adapted
conditions (P/D = 1.25). Adapted from Swain et al. (2017). The red symbols and line
from Swain and Das (2017).
indicate the bundle average heat transfer coefficients under uniform heat flux. (For
interpretation of the references to colour in this figure legend, the reader is referred
to the web version of this article.)
10
row 1 (bottom)
row 2
104 row 3
row 4
P/D=1.25
Wall superheat (K)
Heat transfer coefficient (W/m2K)

8x103 row 5 (top)

6x103

4x103

1
3
2x10
RUN 4
RUN 1 q=10.34 kW/m 2
RUN 5
RUN 2 RUN 6
0 2 4 6 8 10
RUN 3 uniform
Mass flux (kg/m 2s)
104 2x104 3x104 4x104
Fig. 14. Variations of wall superheat with mass flux. Adapted from Gupta (2005).
Heat flux (W/m 2)

Fig. 12. Variation of heat transfer coefficients in a horizontal bundle under multiple
heat flux conditions (P/D = 1.25). Adapted from Swain et al. (2017). The red symbols In the flow boiling study reported in Zhang et al. (2018), bundle
and line indicate the bundle average heat transfer coefficients under uniform heat averaged HTCs increased with the increasing heat flux in all the
flux. (For interpretation of the references to colour in this figure legend, the reader tested conditions and the HTCs on the top row were about twice
is referred to the web version of this article.)
as large as those on the bottom row. Besides, it was observed that
the increase in local vapor quality slightly increased the local HTC.
thermal boundary layer and the bubble growth thus impedes the The averaged HTC also increased with the increasing vapor quality,
heat transfer, as shown in Fig. 13. especially after the transition point, as shown in Fig. 15. Moreover,
Gupta (2005) found that at low mass flux (0–10 kg/m2s), the at low heat fluxes, the transition point, from the nucleate boiling
effect of mass flux on HTC is considerable at low heat fluxes and regime to convective boiling regime, shifts to left as the mass flux
nearly negligible at high heat fluxes. At low heat fluxes, mass flux increasing. After the transition, the averaged HTCs increased with
is more influential on the lower tubes than on the upper tubes, as the increasing mass flux at the lower heat flux of 20 kW/m2, and
shown in Fig. 14. Moreover, the increase of mass flux enhances the then this mass flux effect became insignificant as the heat flux
heat transfer on the lower tubes but deteriorates the heat transfer increases. This phenomenon was consistent with that obtained in
on the upper tubes. This is because the increase of inlet liquid mass Aprin et al. (2011) of iso-butane and did not contradict the results
flux impedes the bubbles rising from the lower tubes and imping- of Swain and Das (2017) and Gupta (2005), because of the zero
ing on the upper tubes, and thus suppresses the bundle effect. At inlet vapor quality in the latter two.
low heat fluxes, the nucleate boiling is weak, and the increase of Abbas and Ayub (2017) studied ammonia flooded flow boiling
mass flux can enhance the convection on the lower tubes so that on a four-pass horizontal bundle with three tubes in each pass.
the HTCs on the lower tubes increase with the increasing mass flux. The heat source was provided by in-tube hot water/glycol solution
However, at higher mass fluxes and higher heat fluxes, as shown in flow. The trends of total heat transfer coefficients were observed to
Fig. 13, larger mass velocity generally hinder the bundle boiling be increased with the increasing heat flux and saturated tempera-
heat transfer. ture, which was in coincidence with the results of Zheng et al.
 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278 13

9000 cannot overcome the heat transfer reduced by the decreased heat
15 kg/m 2s q=20 kW/m 2 flux. Therefore, the heat transfer coefficient decreases along the
8800 30 kg/m 2s bundle height.
Heat transfer coefficents (W/m2K)

50 kg/m 2s Abbas and Ayub (2017) also altered the tube side fluid inlet
8600 from the bottom nozzle to the top nozzle. The total heat transfer
performance showed no noticeable difference under a reversed
8400 tube side flow direction. However, the local HTC performance
under reversed in-tube flow was not presented by the authors.
8200 The authors developed a correlation for shell side tube bundle boil-
ing of ammonia through best possible data fitting as follows:
8000
0:1
0 0 0:90:4pr 0:6
hTP ¼ 70q p0:55
r ðlogpr Þ ð7Þ
7800
The authors compared their experimental data with some gen-
7600 eral pool boiling correlations, such as Cooper (1984), Gorenflo
(1993), Shah (2007), Ribatski and Jabardo (2003), Mostinski
7400 (1963), Stephan and Abdelsalam (1980) and Fernández-Seara
0 .0 1 0 .1 1 et al. (2016). A wide disparity was observed except the prediction
(xin+xout)/2 by Mostinski (1963) correlation, which indicated that correlations
formulated for water and common refrigerants are not suitable for
Fig. 15. Variation of averaged HTC with vapor quality under different mass fluxes. ammonia.
Adapted from Zhang et al. (2018).
Aiming to reduce the refrigerant charge during the evaporation,
Ayub et al. (2017) and Abbas et al. (2017b) further studied the exit
superheat effect on the ammonia expansion evaporation and
(2001) and Zeng et al. (2001). The heat flux decreased from the bot-
observed that the increase of exit superheat caused a diminishing
tom pass to the top pass as the in-tube fluid flowed up and cooled
effect of boiling and the dominance of a single-phase sensible
down. The HTC of individual pass decreased with the bundle height
regime. At higher exit superheat, a dryout performance was
significantly, as shown in Fig. 16. The authors emphasized that this
observed on the top pass, and the thermal energy was only served
‘diminishing heat flux’ phenomenon does not contradict the con-
to superheat ammonia vapor. The bundle effect was less significant
cept of ‘bundle effect’ which is typically observed under constant
under lower heat fluxes, and lower vapor exit superheats. Abbas
wall temperature or heat flux, because the in-tube fluid flow suf-
et al. (2017a) further found that at higher saturation temperature,
fered temperature decreasing during the heat transfer process. In
the inlet quality influence was considerable, and the average bun-
Zeng et al. (2001), the sprayed ammonia was heated by in-tube
dle HTC decreased with the increasing inlet quality.
hot fluid flow, but the heat flux did not vary along the bundle
height so that the obvious bundle effect was observed at higher
saturation temperatures. In Swain and Das (2017), a decreasing 3.3. Effects of bundle layout on bundle boiling
trend of heat flux from the bottom to the top of the tube bundle
was applied to RUN 1–3, which is similar to the heat flux distribu- The heat transfer in tube bundle boiling is highly dependent on
tion in Abbas and Ayub (2017). However, obvious bundle effect the bundle geometry, i.e., tube diameter and length, pool/shell size,
was observed in Swain and Das (2017), and the diminishing heat tube shape, pitch to diameter ratio, tube number and bundle lay-
flux trend even enhances the bundle effect. This is possibly out. Pitch to diameter ratio, which is usually in the range of 1.2–
because, in Abbas and Ayub (2017), the saturation temperature is 2 in general tube bundles, is a primary geometric parameter, even
low, and nucleate boiling is weak. The heat transfer enhancement though the specific effects on individual bundles were found to be
on upper tubes induced by the rising bubbles from the lower tubes different from the literature (Liu and Qiu, 2004, 2006; Ribatski
et al., 2008). Some previous studies found that boiling heat transfer
was enhanced in some compact tube bundles (P/d < 1.2) at low
10
heat fluxes but the bundle effect could be diminished at the same
9
Pass 1 (bottom) time (Swain and Das, 2014). Therefore, there is an optimal value of
Heat transfer coefficientc (kW/m2K)

Pass 2
tube spacing for which the maximum HTC can be obtained for a
8 Pass 3
Pass 4 (top) certain tube bundle and operating condition.
7 In the pool boiling study with uniform heat flux along the bun-
dle height in Swain et al. (2017), as shown in Fig. 17, a large pitch
6 to diameter ratio of 1.95 obviously deteriorated the bundle aver-
aged heat transfer coefficient, except the result under the lowest
5
heat flux. The HTCs of P/D = 1.6 were generally higher than the
4 HTCs of P/D = 1.25. The local HTC on each tube of RUN 6 (Table 4)
is depicted in Fig. 18. Similarly, the bundle with a large P/D of 1.95
3 had the lowest HTC on each tube. However, different from the bun-
2 dle averaged HTC results in Fig. 17, the local HTCs with P/D of 1.25
were slightly higher than those with P/D of 1.6, indicating when
1 lower heat flux was imposed on the lower tubes, small tube spac-
Tsat=-8 °C
ing can enhance the bundle heat transfer by boosting bubble inter-
0
0 10 20 30 40 50 60 70
action and natural convection. When a large number of bubbles
were generated on the lower tubes under higher heat flux, small
Heat flux (kW/m2)
tube pitch may inhibit the bubble growth and turbulence evolution
Fig. 16. Variation of averaged HTC with heat flux for each pass. Adapted from Abbas and thus deteriorate the bundle heat transfer. There should be an
and Ayub (2017). optimum tube pitch value for each different heat flux arrangement.
14 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

4.5x103 5.0
P/D=1.25 q"low=90 kW/m 2, q"top=10 kW/m 2
4.5
Heat transfer coefficient (kW/m2K)

4x103 P/D=1.6 q"low=90 kW/m 2, q"top=30 kW/m 2

P/D=1.95
4.0 q"low=90 kW/m 2, q"top=90 kW/m 2
3.5x103 q"low=30 kW/m 2, q"top=10 kW/m 2
3.5
q"low=30 kW/m 2, q"top=30 kW/m 2

Bundle effect
3.0 q"low=30 kW/m 2, q"top=90 kW/m 2
3x103
2.5

2.0
2.5x103
1.5

1.0
2x103
104 1.5x104 2x104 2.5x104 3x104 0.5
2
Heat flux (kW/m ) 0.0
1 2 3 4 5 6
Fig. 17. Variation of heat transfer coefficients in a horizontal bundle under different
heat flux and three pitch to diameter ratios. Adapted from Swain et al. (2017). P/D

Fig. 19. Variation of bundle effect with P/D under lower tube heat fluxes of
30 kW/m2 and 90 kW/m2. Adapted from Kang (2015b).
20
RUN 6
P/D=1.25 probably the reason why larger P/D of 1.95 owned higher bundle
Heat transfer coefficient (kW/m K)
2

10 P/D=1.6 averaged HTC under the lowest heat flux in Fig. 17 in Swain et al.
P/D=1.95 (2017). However, with the further increase of the tube spacing,
the bundle effects decreased again, because the large bubbles will
break up or shrink, and the turbulence intensity started to decay
before arriving at the influential region of the upper tube. Eventu-
ally, the influence of the bubble flow from the lower tube was neg-
ligible, and only heat flux dominates the heat transfer on the upper
tube, bundle effects converged to 1.
Lakhera et al. (2014) observed that the local heat transfer coef-
ficient of a tube in the bundle increased with the decreasing tube
pitch (P/D = 1.4, 1.7 and 2.0). They proposed a best-fitting correla-
tion (Eq. (8)) of pool boiling heat transfer enhancement for plain
1 and coated tube bundles from the experimental data as follows:
1 2 3 4 5
Row Number hnpb;local
p 0:297
¼ 2:288½1=ð1  eÞ0:275 ð8Þ
hbottomtube D
Fig. 18. Variation of heat transfer coefficients for tube rows under three pitch to
diameter ratios. Adapted from Swain et al. (2017). Kang (2015a, 2016) studied the influences of inclination angle
and elevation angle on the pool boiling of two tandem tubes, as
Kang (2015b) studied the bundle effect under a wide range of shown in Fig. 20. The increase of elevation angle (Kang, 2015a)
pitch to diameter ratio from 1.5 to 6, as shown in Fig. 19. The bun- from the horizontal position to the vertical position increases the
dle effect here refers to the Memory et al. (1994) definition, as bundle effect, especially when the elevation angle was greater than
mentioned in Section 3.2. The bundle effects under six heat flux 45°. This result validates that the pool boiling inside an in-line bun-
conditions all decreased as the P/D increases from 1.5 to 2 because dle has better heat transfer performance than that inside a stag-
the increase of tube spacing de-escalated the intensity of the nat- gered bundle under the same operating condition. The increase
ural convection. While after that, with the increase of tube pitch, in inclination angle (Kang, 2016) from horizontal to vertical
the bundle effects experienced a slight increase when the upper decreases the bundle effect due to the decrease of the influential
tube heat flux was larger than 30 kW/m2 and a remarkable area of the bubble flow and convection from the lower tube, partic-
increase when the upper tube heat flux was 10 kW/m2, but all ularly when the inclination angle was smaller than 30°.
eventually decreased again. The possible reason was the increase
of the tube pitch weakened the natural convection induced by bub-
ble dynamics and buoyancy effect but provided more space for
bubble growth and turbulence development which also con-
tributed to the turbulence intensity. The second mechanism was
responsible for the slight increase of the bundle effect after its ini-
tial decline, but with the further increase of P/D, the first mecha-
nism prevailed again. When the upper tube had low heat flux
and the lower tube has high heat flux, the bubble nucleation on
the lower tube was enhanced and the larger tube spacing allowed
the adequate growth of a large number of bubbles and the corre-
sponding turbulence development, so that the bundle effects expe- Fig. 20. The inclination angle and elevation angle of two tandem tubes. Adapted
rience a notable increase with the increasing P/D. This was from Kang (2015a, 2016).
 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278 15

3.4. Pressure effects on bundle boiling 25

10 kW/m2 (LL LP) 40 kW/m2 (LL LP)
As summarized by Feldmann and Luke (2008), the nucleate 25 kW/m2 (LL LP) 55 kW/m2 (LL LP)
20
boiling heat transfer coefficient generally increased with the test 65 kW/m2 (LL LP)
pressure, including the sub-atmospheric pressure condition. How-
15
ever, the experimental evidence of bundle boiling under sub-

Wall superheat (K)

atmospheric pressure is still minimal. The heat transfer behavior
could be very different from it under atmospheric pressure because 10
the static pressure induced by the liquid height can be of the same
order of magnitude as the local saturation pressure which results 5
in the decreasing of saturation temperature and subcooling with
the pool height.
0
Giraud et al. (2015) observed that unlike spherical bubbles of
40 kW/m2 (HP)
millimeter size formed at atmospheric pressure, the bubbles gen-
-5 10 kW/m2 (HP) 50 kW/m2 (HP)
erated under the sub-atmospheric pressure of 0.85–15 kPa own
20 kW/m2 (HP) 60 kW/m2 (HP)
much larger size in centimeters and a mushroom-shape. The hemi- 30 kW/m2 (HP) 70 kW/m2 (HP)
spherical growth of bubble occurs during an initial period which is -10
dominated by the inertia force. Then under the effect of the buoy- 0 1 2 3 4 5 6 7
ancy force, the radius of the dry area decreases, and the bubble Row number (from bottom to top)
starts to rise from the heating surface. Directly after the bubble
departure, a secondary vapor column is rapidly formed on the left Fig. 21. Variation of wall superheat on individual tubes in the central column of a
bundle under pressure of 850 mbar (HP) and 50 mbar (LL LP). Adapted from McNeil
dry spot and penetrates the flattened base of the departed primary
et al. (2015).
bubble, which results in the mushroom-shaped bubbles. The flat-
tened spheroid shape of the primary bubble is formed because of
the non-homogeneous temperature distribution, as discussed in model separately, indicating the heat transfer oscillated between
Madejski (1966) and Giraud et al. (2015). Giraud et al. (2015) also these two regimes. While one set of the data broadly agreed with
found a new ‘‘cyclic boiling regime” for pool boiling at sub- the isolated tube model, the other was between the two prediction
atmospheric pressure, which was marked by significant wall tem- lines at 40 kW/m2, probably because boiling is involved in the flow
perature fluctuation. In each cycle, the wall temperature increases oscillation but cannot be well predicted by the equilibrium model.
at the beginning to store the thermal energy. After the initial nucle- McNeil et al. (2017) further studied the boiling heat transfer of a
ation, the wall temperature drops sharply and then keeps decreas- glycerol-water binary mixture at 50 mbar. The binary mixture
ing as the bubble growth. After the bubble departure, the wall owns a lower onset heat flux compared to water, but basically,
temperature increases again during the succeeding waiting time. the same onset wall superheat. Moreover, the heat transfer perfor-
In the studies on compact bundle boiling at 20, 50, 100 kPa in mance of the glycerol-water mixture is poorer than the water’s,
Liu and Liao, (2006) and Liao and Liu, (2007), bundle HTC was and additional heat transfer resistance is induced due to the effects
observed to be increased with the increase of test pressure and of mass transfer at the liquid–vapor interface.
no noticeable difference of the heat transfer behavior from the When bundle boiling occurs at a near-critical pressure, due to
atmospheric pressure condition was reported. the high heat flux imposed, a departure from the nucleate boiling
McNeil et al. (2015) conducted three sets of tests on naturally (DNB) and dryout may happen, which will exhibit an abrupt rise
convective boiling of water at sub-atmospheric pressures: low pool of heated wall temperature. So, at a near-critical pressure, the heat
water level (0.8 m) at 850 mbar (HP series), low pool water lever transfer coefficients will not directly increase with the increasing
(0.8 m) at 50 mbar (LL LP series), and high pool water level (2 m) pressure. Wang et al. (2017) investigated upward flow boiling in
at 50 mbar (HL LP series). At 850 mbar, boiling occurred at all heat the central sub-channel of an inline 2  2 vertical rod bundle at
fluxes, while at 50 mbar, boiling happened only at higher heat subcritical pressures of 11–19 MPa to simulate the two-phase heat
fluxes. The HP and LL LP data were reasonably depicted by an equi- transfer in a fuel bundle of the supercritical water-cooled reactor
librium one-dimensional model which demonstrated an interac- (SCWR). As shown in Fig. 23, in the saturated boiling region with
tive behavior between tubes and a subcooled flow boiling steam quality between 0 and 1, the wall temperature almost kept
pattern. As shown in Fig. 21, the wall superheat is larger and has constant with the increase of heat flux at the lower steam quality,
a broader range under the lower pressure. However, no noticeable then a sharp increase occurs at higher heat fluxes and steam qual-
bundle effect was observed in the HP results, and only the super- ity of around 0.28, which indicated the happening of DNB and the
heat data on the central column under higher heat fluxes exhibited corresponding heat transfer deterioration. As shown in Fig. 24, the
a slight bundle effect in LL LP case. The possible reason was that increase of the pressure significantly enhanced the heat transfer
convection weighs more in the heat transfer at the lower pressure both before and after DNB, because of the decrease in latent heat
than at the higher pressure, even though it is the dominant factor and liquid surface tension under higher pressure. The rise of the
of the heat transfer in both HP and LL LP cases. So, the bundle effect mass flux also reduced the HTC deterioration after DNB. Moreover,
could benefit more from the stronger convection in LL LP case after under small q/G ratio, wall temperature at subcritical pressure was
the onset of nucleate boiling. The HL LP data were best described lower than at supercritical pressures, while an opposite trend was
by an isolated tube model, indicating heat transfer was dominated found under a large q/G ratio. Chen et al. (2019) observed similar
by natural convection or nucleation. As shown in Fig. 22, wall results in their study on water flow boiling in a three-rod bundle
superheat fluctuation was observed in the HL LP case under the at near-critical pressure.
heat flux of 25 kW/m2 and 40 kW/m2, but the author did not dis-
cuss if it was cyclic as found in Giraud et al. (2015). According to
both isolated tube and equilibrium theories, boiling cannot be 3.5. Bundle effect in vertical and special-shaped bundles
onset at 25 kW/m2 but can be partially triggered at 40 kW/m2.
The two data sets obtained at 25 kW/m2 ten minutes apart gener- Vertical tube bundles were widely used in a passive heat
ally agreed with the isolated tube model and the equilibrium removal system in a nuclear water reactor, as shown in Fig. 25.
16 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

25 kW/m2, Column 1, set a

20 25 kW/m2, Column 2, set a
25 kW/m2, Column 3, set a
25 kW/m2, Column 1, set b
10 25 kW/m2, Column 2, set b
Wall superheat (K) 25 kW/m2, Column 3, set b
25 kW/m2, equilibrium model
25 kW/m2, isolated tube model
0
40 kW/m2, Column 1, set a
40 kW/m2, Column 2, set a
40 kW/m2, Column 3, set a
-10 40 kW/m2, Column 1, set b
40 kW/m2, Column 2, set b
40 kW/m2, Column 3, set b
40 kW/m2, equilibrium model
-20
40 kW/m2, isolated tube model

0 1 2 3 4 5 6 7
Row number (from bottom to top)

Fig. 22. Wall superheat on individual tubes in different data sets. Adapted from McNeil et al. (2015).

Chung et al. (2015) studied the subcooled pool boiling heat transfer
550 in a confined vertical tube bundle. As shown in Fig. 26, the lower
2
200 kW/m
water subcooling enhanced the heat transfer coefficients at all
400 kW/m2
500 monitored locations, and the HTCs increased along the bundle
600 kW/m2
Bulk tempearture height. Moreover, the heat transfer coefficient of the central tube
was higher than the peripheral tubes owing to an active interaction
450
Wall temperature (K)

of the fluid among the tubes, especially under higher subcooling.

Generally, the averaged bundle HTC was slightly higher than the
400 single tube HTC under the same conditions. Compared to the satu-
rated pool boiling study of Gupta et al. (2010) using a similar bun-
350 dle structure, as shown in Fig. 27, the bundle enhancement ratios
(h/hbottom) of subcooled boiling under higher heat fluxes in
0 x 1 Chung et al. (2015) were lower than the results of saturated boiling
300
in Gupta et al. (2010). Lu et al. (2019) observed that in the flooded
boiling in a 9x9 vertical bundle, the tube wall temperature gradu-
250 P=15 MPa ally decreases along the height and the water temperature
G=1000 kg/m2s increases along the bundle height, which indicating an obvious
200 bundle effect.
800 1200 1600 2000 2400 2800 3200 Zhang et al. (2017a) studied flow boiling in a 7-tube vertical
Bulk enthalpy (kJ/kg) bundle at 50 kW/m2. It was observed that the heat transfer coeffi-
cient increases significantly with the increasing total mass velocity
Fig. 23. Variation of wall superheat with bulk enthalpy at different heat fluxes. at low flow rates, which indicated forced convection dominated
Adapted from Wang et al. (2017). the two-phase flow boiling heat transfer. This phenomenon was
very different from the results obtained over horizontal bundles
70 under the same heat flux (Zhang et al., 2018). Besides, under a cer-
q=400 kW/m2 11 MPa tain mass velocity, the HTCs increased with the increasing steam
60 G=1000 kg/m2s 15 MPa quality at lower steam qualities less than 0.1 and exhibited the
Heat transfer coefficient (kW/m2K)

19 MPa opposite trend at higher steam qualities. This behavior was also
different from the horizontal bundle results and can be attributed
50
to the domination of forced convective heat transfer rather than
nucleate boiling due to the weakened bubble motion effect, as dis-
40 cussed in Kang (2016). Zhang et al. (2017a) developed a superposi-
tion HTC correlation for the flow boiling over a vertical bundle,
30 with Weisman (1959) correlation presenting the forced convection
HTC (hCVL) and Forster and Zuber, (1955) correlation presenting the
20
nucleate boiling HTC (hnpb). The new suppression factor (S) and an
enhancement factor (F) correlations were developed, as illustrated
in Eqs. (9) and (10). The pitch to diameter ratio is included in
10
Weisman (1959) correlation, but the bundle height was not consid-
ered in their correlation.
0  0:2
800 1200 1600 2000 2400 2800 3200 1
F ¼ 1 þ 32000Bo1:18 þ 5:5 ð9Þ
Bulk enthalpy (kJ/kg) X tt

Fig. 24. Variation of heat transfer coefficient with bulk enthalpy at different 1
pressures. Adapted from Wang et al. (2017).
S ¼ ½1 þ 0:000001  ðF  ReL Þ1:26  ð10Þ
 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278 17

Fig. 27. Bundle enhancement ratios (h/hbottom) along the height of vertical tube
bundles with data from Gupta et al. (2010) and Chung et al. (2015).

11000
Fig. 25. Passive containment cooling system in Economic Simplified Boiling Water Row 1 (bottom) Row 5 Row 13
Reactor. From Sawyer and Boardman (2015). 10000 Row 21 Row 25
G=30 kg/m2s
9000
HTC (W/m2K)

8000
2.8
Bundle -center (70°C, 10kW)
Normalized Heat transfer coefficient (-)

Bundle -side (70°C, 10kW) 7000

2.4
Single tube (70°C, 10kW)
Bundle -center (30°C, 4kW) 6000
2.0 Bundle -side (30°C, 4kW)
Single tube (30°C, 4kW) 5000
1.6
4000
1.2
Red symbols: q=20 kW/m2 Black symbols: q=55 kW/m2
3000
0.8 0.00 0.05 0.10 0.15 0.20 0.25
xin
0.4
Fig. 28. Local HTCs on different rows for q = 20 and 55 kW/m2. Adapted from Zhang
et al. (2019).
0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Distance from bottom (m) ber of vapor bubbles and insufficient liquid on the upper tubes in
Fig. 26. Comparison of normalized heat transfer coefficients for a single tube and
the fully-developed flow boiling. The authors employed the
tube a bundle along the bundle/tube height under different cooling water Zhang et al. (2017a) correlation and Shah (2007) correlation to cal-
temperature (30 and 70 ) and different power supply (4 kW and 10 kW). Adapted culate the HTCs when the bundle was placed vertically and hori-
from Chung et al. (2015). (The HTCs were normalized by the inlet value of the zontally, respectively. From the calculation results, at low heat
central tube with 70 cooling water and 10 kW heater power, as represented by the
fluxes, the influence of inclination angle on the HTC is insignificant.
first black solid symbol.)
While at high heat fluxes, the HTCs in the inclined bundle are lower
than those in the horizontal and vertical bundles. The authors pro-
posed a new Chen type HTC correlation, but its applicability is
Zhang et al. (2019) further investigated the flow boiling heat restricted to the experimental conditions of their work.
transfer in an inclined bundle with the tube inclination angle of The V-shape tube bundle is widely applied in passive heat
45⁰. The bundle effect is significant in the lower part of the bundle, exchangers, and unlike tandem tubes, the pitch between the two
from the bottom row to the eleventh row. Then the local HTCs parts of a V-shape tube varies with the tube length, as shown in
maintained relatively stable from the eleventh to the top row (25 Fig. 29. Kang (2017, 2018) observed the bundle effect of a V-
rows in total), especially at the higher heat fluxes, as shown in shape tube array is slightly higher than the value of parallel tubes
Fig. 28. Moreover, the bundle average HTC was observed to in the author’s previous studies (2015a, b; 2016). The increase in
decrease with the increasing heat flux at both low and high mass included angle enhanced the bundle effect, especially when the
velocities. The authors attributed this phenomenon to a large num- upper tube was at low heat flux and the lower tube was at high
18 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

In Zhang et al. (2017b), due to the different heat transfer behav-

iors of the different sections of the C-shape bundle, the heat trans-
fer coefficient correlations were developed for the vertical and
horizontal sections separately. For the vertical part of the bundle,
an empirical correlation was proposed as follows:
 0:34
00 0:95 H
hbV ¼ 0:1ðq Þ ð11Þ
D
The Rohsenow (1951) pool boiling correlation can obtain a rel-
atively applicable prediction of the average HTC in the lower hor-
izontal bundle. The Chen (1966) forced convective boiling
correlation was recommended to evaluate the heat transfer behav-
ior on the upper horizontal bundle.
Jeon et al. (2015, 2016) found that in the nucleate pool boiling,
the heat transfer mechanisms on the upper part and lower part of a
Fig. 29. The included angle (a) and rotation angle (u) of a V-shape tube array. U-tube are different. The lower part of the U-tube is prevailed by
Adapted from Kang (2017, 2018). nucleate pool boiling, while the upper part of the U-tube is domi-
nated by forced convective boiling. Therefore, different HTC corre-
lations were applied to the different parts of the U-tube. Based on
heat flux because the increased tube pitch facilitated the formation the heat transfer mechanism, Jeon et al. (2015) developed a new
of large bubbles and turbulence development as discussed earlier. boiling model with a new subcooled pool boiling correlation and
When the rotation angle of the V-shape array varied from 0° to a new forced convective boiling correlation to the lower and upper
180°, the HTC was considerably influenced by this variation. The parts of the U-tube, respectively. The proposed boiling model pre-
maximum HTC was obtained at a rotation angle of 90°. The ratio dicted the experimental heat transfer coefficients better than the
of maximum HTC to minimum HTC under different rotation angles default nucleate boiling model by Chen (1966) in MARS. Jeon
reached about 2 when the heat flux of the upper tube was et al. (2016) developed a heat transfer model package for the hor-
60 kW/m2. izontal U-shaped HX submerged in a pool by improving the hori-
Zhang et al. (2017b) investigated saturated pool boiling in a zontal in-tube condensation model and developing the out-tube
C-shape bundle to explore the heat transfer characteristics in Pas- natural convective and nucleate boiling model. The proposed
sive Residual Heat Removal Heat Exchanger (PRHR HX) of the model provides an improved prediction of HX performance (con-
third-generation water reactor AP1000. As shown in Fig. 30, the densation, natural convection and nucleate boiling, and heat
HTCs were significantly influenced by the heat flux and the corre- removal rate of the HX) compared to the default model in MARS.
sponding bubble behavior, as well as the bundle height and orien- The Cornwell and Houston (1994) correlation was adopted to pre-
tation. The HTCs increased along the height of the vertical section dict the heat transfer on the lower part of the U-tube but modified
of the bundle, and at a specific H/D ratio, the increasing heat flux by a suppression factor (S) to include the effect of water subcooling
enhanced the heat transfer. The heat transfer performance of the degree, as shown in Eq. (12).
two horizontal sections was weaker than the vertical section under  0:167
similar thermal and flow conditions. Furthermore, the HTC of the T sub
S¼ 1þ ð12Þ
upper horizontal section of the bundle was higher than that of T sup
the lower horizontal section of the bundle because the fluid and
A superposition correlation was applied to predict the heat
bubbles from the lower part induced additional mixing effects
transfer on the upper part of the U-tube. The convective suppres-
and forced convective velocity to this region.
sion was disregarded and the same nucleate boiling expression
as the lower part was used for hnpb with S only counting the sub-
cooling effect. The enhancement factor (F) was calculated by
Polley (1980), owing to its consideration for the void fraction
effect. The liquid phase convective heat transfer term was deter-
mined by averaging the HTCs of both the cross and parallel flow,
as expressed in Eq. (13).
2 2 0:5
hCVL ¼ ðhCVL;cross þ hCVL;parallel Þ ð13Þ

The Whitaker (1972) correlation and Dittus and Boelter (1930)

correlation were used for calculating the HTCs of the cross flow and
parallel flow, respectively.
The above investigations reveal that bubble departure diameter
and frequency were both enhanced by the tube bundle comparing
to a single isolated tube. Bundle effect is significant at lower heat
fluxes (q < 60 kW/m2) and when the heat fluxes on the lower tubes
are larger than those on the upper tubes. The local HTCs on the
upper tubes are less sensitive to the variation of heat flux. The
decreasing heat flux distribution along the bundle height is bene-
ficial to the heat transfer and can be considered to be applied in
the industrial heat exchangers. At low heat flux, very low mass flux
(G < 10 kg/m2s) and zero inlet vapor quality, the increase of mass
Fig. 30. Heat transfer coefficients distribution on the C-shape tube bundle along the flux deteriorates bundle effect but enhances the total HTC. The
bundle height under different heat fluxes. Adapted from Zhang et al. (2017b). bundle HTC decreases with the further increase of the mass flux
 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278 19

(G > 10 kg/m2s). When heat flux is high, the effect of mass flux is layer. Therefore, the increase of surface wettability (decrease of
insignificant. contact angle) could decrease the nucleate boiling heat transfer.
The influences of P/D on bundle effect were different under dif- However, higher wettability can enhance the CHF due to the capil-
ferent conditions, but in the P/D range of 1.2–2, the bundle effect lary wicking ability (Takata et al., 2005; Chu et al. 2013). In addi-
was generally inhibited by increasing P/D. The optimal P/D of a tion, some hydrophilic surfaces with a porous layer can enhance
bundle also depends on the heat flux distribution along the bundle the bubble nucleation and rewetting, so that the CHF could be
height. Generally, the heat transfer performance of an in-line hor- delayed (O’Hanley et al., 2013).
izontal bundle is better than that in a staggered bundle or vertical Due to rapid progress in surface modification and fabrication
bundle. technology in recent years, a large number of enhanced heating
In addition, when the boiling heating source was provided by surfaces have been developed. Frequently used surface modifica-
in-tube hot fluid-flow instead of electrical heaters, the heat trans- tion methods include mechanical machining, surface coating,
fer coefficient will not monotonically increase along the bundle chemical process, and micro/nano electromechanical system tech-
height but depend on the in-tube heat transfer and may even exhi- nology (Kim et al., 2015; Leong et al., 2017), as shown in Fig. 31.
bit a diminishing trend, such as in Abbas and Ayub (2017). In Van Through modifying the surface structures, the disjoining and cap-
Rooyen et al. (2012), the decreasing temperature of the in-tube illary effects near the triple contact line are changed, which will
heating water along the flow path exhibited a quadratic distribu- significantly influence the bubble dynamics and microlayer evapo-
tion and resulted in a linear decrease in local heat flux along the ration. Therefore, the heat transfer is affected by the modified
water flow path. Shah (2017) pointed out bundle average HTCs heated surfaces.
are independent to the heating mode which means using uniform Most studies referring enhanced surfaces focused on the heat
wall heat fluxes is capable of predicting the real HTCs under vari- transfer on a small plate, inside or outside a single tube or in a
able wall heat fluxes caused by in-tube hot fluid flow. However, micro-scale channel, while only a few investigations concerned
under such conditions, the local two-phase flow and heat transfer surface enhancement combining with the bundle effect together.
behaviors could differ a lot in the different positions in a bundle, The present review only considers boiling heat transfer perfor-
and the average HTC is not capable of reflecting the variable heat mance of the tube bundle with the enhanced surfaces. Thus, the
transfer mechanisms. Therefore, the influence of the in-tube fluid surface fabrication approaches and surface microstructures will
flow should be considered in the HTC prediction. Moreover, in a not be discussed in detail. The experimental studies selected and
specially shaped bundle (V, U, and C shape), the interaction reviewed in this section are summarized in Table 5.
between the two-phase flow in different parts of the bundle should Some microstructures on the boiling tube surfaces can achieve
be counted, and different HTC correlations should be applied to dif- excellent heat transfer enhancement in comparison with smooth
ferent parts. tube surfaces. Boiling surface superheat was observed entirely
independent of the imposed heat flux (Mitrovic and Ustinov,
2006). Researches ascribed this phenomenon to the considerable
4. Boiling heat transfer over tube bundles with enhanced
length of the three-phase line (TPL) within the microstructure.
surfaces
Ustinov et al. (2011) also observed this phenomenon in their study
on R134a and FC3284 pool boiling. The lower tube strengthened
Surface characteristics of heated surfaces are among the most
the heat transfer on the upper tube by inducing turbulence in
influential factors in the boiling heat transfer in a tube bundle.
the two-phase bubbly flow, which was consistent with the plain
Parameters such as active nucleation site density, bubble departure
bundle results. However, different from plain bundle heat transfer,
diameter and frequency all tightly bound to surface characteristics,
the wall superheat increased with the increasing pressure. The
such as surface roughness, wettability, and porosity. Higher surface
authors attributed this phenomenon to the suppression effect of
roughness may enhance the active nucleation site density and thus
pressure on the critical vapor bubble radius, and the rough surface
improve the nucleate boiling heat transfer (Webb, 1981; Kim et al.,
cannot support the reduced sized bubbles at high pressure. The
2016). Wang and Dhir (1993) observed that higher wettability
microstructure properties were the most important influential fac-
decreases the number of active nucleate cavities. Phan et al.
tor in boiling heat transfer enhancement, and a higher pin density
(2009) found that a larger contact angle increases bubbled depar-
achieved a higher efficiency.
ture frequency and decreases the thickness of the liquid micro-

Fig. 31. Images of the enhanced surfaces. (a) SEM image of the micropin-enhanced surfaces from Ustinov et al. (2011). (b) Plasma coated surface from Swain and Das (2018).
 20
Table 5
Summary of experimental conditions and tube bundle characteristics for studies with enhanced tube bundles.

Author Boiling Fluids Tube bundle Tube surface Operating conditions Heating method Max. HTC (kW/m2K)
mode characteristics characteristics
Ustinov et al. (2011) Pool R134a, FC- 2 tandem horizontal Micropins by electrolytic 5–9 bar (R134a), 0.5–1.5 bar (FC-3284), Electrically heated ~66, (R134a, 5 bar, 125 KW/m2)
boiling 3284 tubes, D = 18 mm, deposition of copper q00 = 5–125 kW/m2
P/D = 1.5
Kukulka and Smith (2014) Flow n- 9  11 staggered Vipertex 1EHT 304L by Remid = 2010–20400, EMTD = 8.6–65.7 °C Steam flow Maximum overall HTC

S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

boiling pentane, horizontal, material surface modification enhancement ratio 200%
p-xylene D = 15.875 mm,
and water P/D = 1.25
Van Rooyen and Thome Flow R134a, 20 staggered horizontal Turbo-B5 (Wolverine) and Tsat = 5-15 °C, G = 4–40 kg/m2s, Hot water Turbo-B5: ~30 (R134a), ~23
(2013, 2014) boiling R236fa tubes, D = ~19 mm, Gewa-B5 (Wieland) q00 = 15–70 kW/m2, x = 0.1–0.9 circuit + enthalpy (R236fa)Gewa-B5: ~35 (R134a),
P/D = 1.167 profile method ~20
(R236fa)
Gorgy and Eckels (2013, Flow R134a, 20 staggered horizontal TBII with a double sides Tsat = 4.44 °C (R134a), 14.44 °C (R123), Hot water R134a: ~30
2016, 2019) boiling R123 tubes, D = 19.05 mm, reentrant enhanced structure G = 15–55 kg/m2s, q00 = 5–60 kW/m2, circuit + enthalpy R123: ~25
P/D = 1.167 x = 0.1–0.7 profile method
Swain and Das (2018, Flow Water 5  3 staggered plasma sprayed copper 1 bar, G = 20.25–192.9 kg/m2s, Electrically heated ~8.7 (65.94 kW/m2)
2019); Swain et al. boiling horizontal, D = 20 mm, q00 = 12.28–65.94 kW/m2
(2018) P/D = 1.25, 1.6, 1.95
Ji et al. (2016) Flow R134a 6  3 staggered external fins Tsat = 6 °C, C = 0.07–0.2 kg/m s, Hot water + Wilson plot ~26 (Tsat = 6 °C, 40 kW/m2,
boiling horizontal, uv = 0–3.1 m/s, q00 = 20–180 kW/m2 u = 2.1 m/s)
D = 19.05 mm,
Pv/D = 1.17, Pl/D = 1.04
Zhao et al. (2018a) Flow R134a 10 staggered horizontal external fins Tsat = 6 °C, 10 °C, 16 °C, C = 0.035–0.07 kg/m s, Hot water + Wilson plot ~21 (Tsat = 6 °C, 40 kW/m2,
boiling tubes, D = 19.05 mm, uv = 0–2.4 m/s, q00 = 20–60 kW/m2 u = 1.5 m/s)
Pv/D = 1.17, Pl/D = 1.04
Zhao et al. (2018b) Flow R134a 7  3 staggered external fins Tsat = 6 °C, C = 0.035–0.1 kg/m s, Hot water + Wilson plot ~24 (Tsat = 6 °C, 30 kW/m2,
boiling horizontal, uv = 0–2.7 m/s, q00 = 20–40 kW/m2 u = 1 m/s)
D = 19.05 mm,
Pv/D = 1.17, Pl/D = 1.04
Zhao et al. (2018c) Flow R134a 7  3 staggered external fins Tsat = 6 °C, 10 °C, 16 °C, q00 = 20–80 kW/m2, Hot water + Wilson plot ~26 (Tsat = 16 °C, 20 kW/m2)
boiling horizontal, Re = 254–3700
D = 19.05 mm,
Pv/D = 1.17, Pl/D = 1.04
 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278 21

Swain and Das (2018) and Swain et al. (2018) investigated sat- coated tube bundle (Eq. (15)) respectively, which matched experi-
urated flow boiling over a plasma-sprayed copper-coated horizon- mental data within a maximum error of ±10%.
tal tube bundle under uniform and variable heat fluxes. The bundle
 0:55
averaged HTC increased with the increasing heat flux, but the P
Nub ¼ 0:74   ð1 þ Peb Þ0:66 Fr 0:045 ð14Þ
increase rate reduced at the high heat fluxes. The bundle averaged D
HTC decreased with the increasing mass flux, especially at higher
mass fluxes. Also, the bundle with smaller P/D owned higher aver-  0:033
P
aged HTC. Bundle effect is still significant in the coated bundle and Nubcoated ¼ 1:71   ð1 þ Peb Þ0:425 Fr0:043  Cn0:043 ð15Þ
enhanced by the decreasing heat flux distribution from the bottom D
to the top. Compared to the smooth tube bundle, a significant heat
Cn ¼ edp =t p is a coating parameter in terms of particle diameter (dp)
transfer enhancement was observed due to the larger active nucle-
used for the porous layer, the porosity of the layer (e), and the thick-
ation site density induced by the coated surface. The wall super-
ness of the coating layer (tp).
heats of an enhanced bundle are much lower and have smaller
Swain and Das (2019) further found that the surface enhance-
ranges, as shown in Figs. 32 and 33. The surface enhancement fac-
ment factor is highest at the bottom row and reduces along the
tor was defined by the authors as the ratio of bundle average HTC
bundle height, which reveals that the upper coated tubes do not
of the coated bundle to that of the plain bundle. High enhancement
perform at their full potential. This is because of the inhibition
factors were obtained under high pitch to diameter ratio, high
effect of the vapor clouding on the upper rows, which is caused
mass flux, and low heat flux. The authors proposed correlations
by the growth and coagulation of vapor bubbles generated from
for flow boiling over plain tube bundle (Eq. (14)) and porous-
the lower tubes as moving upwards in the bundle. This inhibition
effect will become more significant on the coated tube bundle at
high heat fluxes due to the enhanced bubble formation on the
104 coated surface. While this inhibition effect is alleviated at higher
mass fluxes owing to the broken-up of vapor clouding, as shown
8x103 in Fig. 34. The authors then only used a limited number of coated
Heat transfer coefficient (W/m 2K)

tubes in the bundle, i.e., upper two rows coated, lower two rows
coated and lower three rows coated, in order to reduce the cost
6x103
of the coated heat exchangers. Higher bundle HTC was observed
Plain, P/D=1.25,
in the bundles with the lower rows using coated tubes. Moreover,
G=20.25 kg/m2s the averaged HTCs of the mixed tube bundle are comparable to the
Plain, P/D=1.6,
4x103 averaged HTC of the fully coated bundle, as shown in Fig. 34. The
G=77.16 kg/m2s optimal number of rows that to be installed with coated tubed
Plain, P/D=1.95, depends on the surface characteristics, bundle configuration, and
G=144.6 kg/m2s operating conditions.
Coated, P/D=1.25, G=20.25 kg/m2s The commercial tube bundles with mechanical machining sur-
Coated, P/D=1.6, G=77.16 kg/m2s faces are more commonly used in industrial heat exchangers.
3
2x10 Coated, P/D=1.95, G=144.6 kg/m2s Kukulka and Smith (2014) examined the flow boiling heat transfer
in a staggered horizontal bundle with double-side mechanically
2x104 4x104 6x104 8x104
Heat flux (W/m2) roughened Vipertex 1EHT enhanced tubes. The 1EHT enhanced
surfaces created more nucleation sites and obtained an up to 97%
Fig. 32. Comparison of bundle average heat transfer coefficients of plain and coated overall enhancement for n-pentane and enhancement of about
tube bundles under uniform heat fluxes at different P/D and mass fluxes. Adapted 200% for water. The enhancement ratio of p-xylene was approxi-
from Swain and Das (2018).
mate 24%, which experienced single-phase convection.

16 9500
G=20.25 kg.m2s, P/D=1.25, coated All rows, 65.94 kW/m2
P/D=1.6
G=77.16 kg.m2s, P/D=1.6, coated
9000 3 bottom rows, 65.94 kW/m2
14 G=192.9 kg.m2s, P/D=1.95, coated
G=20.25 kg.m2s, P/D=1.25, plain 2 bottom rows, 65.94 kW/m2
G=77.16 kg.m2s, P/D=1.6, plain 8500
12 G=192.9 kg.m2s, P/D=1.95, plain
Wall superheat (K)

HTC (W/m2K)

8000
10
7500

8
7000

6
6500
All rows, 32.48 kW/m2
3 bottom rows, 32.48 kW/m2
4 RUN 6 6000
2 bottom rows, 32.48 kW/m2

2.0x104 2.5x104 3.0x104 3.5x104 4.0x104 0 20 40 60 80 100 120 140 160 180 200

Heat flux (W/m2) Mass flux (kg/m2s)

Fig. 33. Comparison of local wall heat superheats between coated and plain tube Fig. 34. Comparison of bundle average HTC for different bundles. Adapted from
bundles. Adapted from Swain et al. (2018). Swain et al. (2019).
22 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

Gorgy and Eckels (2013) studied the saturated convective boil- DpL ¼ f 2/ 2G2 NR =q ð19Þ
ing of R134a and R123 in an enhanced horizontal bundle composed
of finned TBII tubes. The authors concluded that heat flux is the NR is the number of tube rows. In the right side of the correlation,
dominant factor in the enhanced tube bundle boiling over mass except the leading coefficient, there are four terms representing
flux or vapor quality. Gorgy and Eckels (2016, 2019) further geometry, modified Helmholtz instability, liquid Weber number,
observed that the bundle HTCs increases rapidly at low heat flux and interfacial shear, respectively.
in comparison to pool boiling probably because of the earlier bub- Van Rooyen and Thome (2014) then found that forced convec-
ble nucleation caused by the flow boiling. The smaller P/D (1.167) tion has little impact on the bundle heat transfer coefficient, and
bundle shows lower performance compared with the other two (P/ the heat transfer performance is similar to pool boiling that heat
D = 1.33 and 1.50), as shown in Fig. 35. The bundle HTCs increase flux plays the dominant role. Moreover, the effects of tube position
with the increasing heat flux at lower heat fluxes then decrease at and flow pattern in the bundle on heat transfer performance are
higher heat fluxes. Besides, for R134a (Gorgy and Eckels, 2016), at negligible. The Turbo-B5 tube exhibits less heat flux dependence.
lower heat fluxes of 10–20 kW/m2, the HTCs increase with the The authors introduced a bundle factor KBB (KBB = hB/hnpb) to pre-
increasing vapor quality, while at higher heat fluxes, an opposite sent the bundle heat transfer enhancement. A new nucleate boiling
trend was observed. Regarding R123 (Gorgy and Eckels, 2019), HTC correlation was proposed with surface characteristics depen-
dry-out was observed at higher vapor qualities for the bundle with dent coefficients, as illustrated in Eqs. (20)–(21). The row effect
the smallest P/D (1.167). New HTC correlations were developed for and flow pattern related effects were not considered because of
the studied enhanced bundle by modifying the nucleate pool boil- their relatively low effects.
ing HTC using a multiplication factor (F), as follows:  TSF b !
hnpb Dext: lH 1
hB ¼ Fhnpb ð16Þ ¼ TSF a 1 ð20Þ
kL da ðDT þ 1Þ7:2
P a
F ¼ ð Þ  ðb  2  ðc  eÞ2 Þ ð17Þ  1=3 2=3
D lH jAj ðqL hfg Þ
¼ ð21Þ
where, a, b, and c are coefficients that depend on heat flux and da DT 1=2
ðkL lL Þ T sat1=6

working fluid.
Van Rooyen and Thome (2013, 2014) studied flow boiling of where TSFa = 2711 (Turbo-B5), 967 (Gewa-B5); TSFb = 0.48 (Turbo-
R134a and R236fa over Turbo-B5 and Gewa-B5 two types of B5), 1.06 (Gewa-B5). The Hamaker constant |A| is 8.6e1021 Pa.m3.
enhanced tube bundles. The authors found that frictional pressure The onset of dryout happened suddenly at a very high vapor quality
drop is generally a function of flow rate and vapor quality but not a of 98% near the thermodynamic limiting condition for both tube
function of tube type, temperature or refrigerants in the tested types, exhibiting a significant improvement compared to the plain
range. In a diabatic condition, the frictional pressure dropped bundle value of 90%. The authors developed a new film Reynolds
increases with the mean vapor quality before the onset of dryout, number expression for the onset of dryout of convective bundle
which occurs as a sudden drop in frictional pressure drop at higher boiling as follows:
mean vapor quality. Because of the difficulty to obtain accurate _  xdry Þ
4mð1
local conditions in diabatic testing, a new prediction model of pres- Reonset ¼ ð22Þ
2Ntubes LlL
sure drop with a phenomenological representation of underlying
mechanisms was proposed based only on the adiabatic data as Ji et al. (2016) studied the impacts of the countercurrent vapor
follows: flow on falling film evaporation of R134a in a horizontal bundle
" #1:2844 !0:15921 with six finned copper tubes in the central column. Compared to
dgap d  d2 ðqL  qG Þgd2 l 0:14487 the pool boiling, the HTCs of the falling film evaporation were
f 2/ ¼ 165 WeL 0:84625 G
2
dgap r lL found to be about 38.4–72.3% higher and less sensitive to heat flux.
ð18Þ Significant bundle effect was observed, especially at lower Reynold
number and lower heat fluxes. After the introduction of additional
upward vapor flow, a slight heat transfer enhancement was
observed especially for the top two tubes and at higher vapor
30000 velocities. The HTC variation with the increasing vapor velocity
2
P/D=1.167, G=15 kg/m s R134a (Fig. 36) is attributed to the film thickness change or sliding effects
2
caused by the ascending vapor flow. The authors speculated the
Heat transfer coefficient (W/m 2K)

P/D=1.33, G=15 kg/m s

25000 2
P/D=1.5, G=15 kg/m s significant dependence of heat transfer behavior on specific tube
positions was because of the liquid maldistribution and dryout
outside the tube surface.
20000 Zhao et al. (2018a) then studied the effects of crosscurrent
vapor flow on the falling film evaporation in a 2-3-2-3 triangularly
arranged tube bundle. Because of the smaller tube bundle depth,
15000 the bundle effect is low (Fig. 36). With the increase of vapor veloc-
ity, the HTC increases first and then declines. Very stable film flow
P/D=1.167, G=25 kg/m s
2 was observed when no extra vapor flow was acting, while the flow
10000 2 became turbulent once the vapor flow was induced. Moreover, the
P/D=1.33, G=25 kg/m s
P/D=1.5, G=25 kg/m s
2 amount of liquid entrainment and displacement increases with the
increasing vapor velocity. Zhao et al. (2018b,c) further investigated
5000
the effects of additional concurrent vapor flow upon falling film
0 5 10 15 20 25 30 35 40 45 50
evaporation. More significant tube bundle effect than countercur-
Heat flux (kW/m2)
rent vapor flow cases was observed (Fig. 37). The vapor stream
Fig. 35. Heat transfer coefficients at different P/D and mass fluxes. Adapted from enhances the heat transfer of the upper tubes while weakens heat
Gorgy and Eckels (2016). transfer of lower tubes.
 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278 23

40 and Van Rooyen and Thome (2014). Gupte and Webb (1995a)
35 Crosscurrent flow: applied fin pitch parameters to predict the boiling HTC in a finned
Heat transfer coefficient (kW/m 2K)

30 Upper: No.1 No.3 No.5 tube bundle. From the perspective of the generality, the application
Lower: No.2 No.4 No.6 of empirical coefficients corresponding to surface characteristics
25
seems a better way to establish heat transfer coefficient correla-
20 tions for bundle boiling with enhanced surfaces. The number of
studies on boiling over enhanced tube bundle is still limited, and
15
the experimental database including more surface categories and
more fluids, especially for local heat transfer performance, are nec-
essary to obtain a comprehensive view.
10
5. Conclusions
Countercurrent flow:
No.1 (top) No.4 This work presents a comprehensive review of recently pub-
No.2 No.5
Tsat=6 oC
lished investigations on the heat transfer during pre-CHF boiling
No.3 No.6 (bottom) q=40 kW/m2 over plain and enhanced tube bundles which is significant to the
5
0.0 0.5 1.0 1.5 2.0 2.5 industrial design of shell and tube heat exchanger. The main con-
clusions drawn from the present literature review are as follows:
u (m/s)

Fig. 36. Heat transfer coefficient distribution of falling film evaporation of R134a 1. The objective approaches measuring local void fraction using
with countercurrent vapor flow (Ji et al., 2016) (falling film flow rate of 0.08 kg/m∙s) time-series signals are suggested to determine the two-phase
and crosscurrent vapor flow (falling film flow rate of 0.07 kg/m∙s) (Zhao et al., flow patterns in tube bundles. The transitions between flow
2018a). patterns seem independent of the liquid superficial velocity in
boiling two-phase flow. The bubble nucleation effects on the
flow regime transition will become insignificant with the
increasing vapor void fraction. For vertical bundles, some
40
authors pointed out that there is no slug bubble because of
the corresponding surface instability.
Heat transfer coefficient (kW/m 2K)

2. The bubble departure diameter and frequency are both

enhanced by the tube bundle compared to a single isolated
tube. Bundle effect is significant at lower heat fluxes
10 (q < 60 kW/m2) and when the heat fluxes on the lower tubes
are larger than those on the upper tubes. The local HTCs on
the upper tubes are less sensitive to the variation of heat flux.
The decreasing heat flux distribution along the bundle height
is beneficial to the heat transfer and can be considered to be
applied in the industrial heat exchangers.
No.1 (top) 3. At low heat flux and zero inlet vapor quality, a low mass flux
No.2 No.5 (G < 10 kg/m2s) is beneficial to the total HTC but detrimental
Tsat=6 oC
No.3 No.6 to the bundle effect. A higher mass flux (G > 10 kg/m2s) deteri-
No.4 No.7 q=40 kW/m2 orates the bundle HTC. When heat flux is high, the effect of
1
0.0 0.5 1.0 1.5 2.0 2.5 3.0 mass flux is insignificant. When the inlet vapor quality is non-
zero, the local tube HTC slightly increases with the increasing
u (m/s)
local vapor quality, while the bundle average HTC is enhanced
Fig. 37. Heat transfer coefficient distribution of falling film evaporation of R134a by the increasing average vapor quality after the transition
with countercurrent vapor flow (Zhao et al., 2018a) (red symbols and lines, falling point.
film flow rate of 0.08 kg/m∙s) and concurrent vapor flow (black symbols and lines, 4. In the P/D range of 1.2–2, the bundle effect is generally inhib-
falling film flow rate of 0.07 kg/m∙s) (Zhao et al., 2018c). (For interpretation of the
ited by increasing P/D, especially under larger heat fluxes. The
references to colour in this figure legend, the reader is referred to the web version of
this article.) optimal P/D of a bundle also depends on the heat flux distribu-
tion along the bundle height. Obvious bundle effect can be
observed in vertical bundles and inclinded bundles. The
From the studies of boiling heat transfer on enhanced tube bun- increase of bundle inclination angle from horizontal to vertical
dles discussed above, the heat transfer performance was domi- decreases the bundle effect, while the increase in elevation
nated by the enhanced surface characteristics and the applied angle from the horizontal position to the vertical position
heat flux. The performance of tube bundles with enhanced surfaces increases the bundle effect. Generally, the heat transfer perfor-
is generally better than the plain tube bundles. The surface mance of an in-line horizontal bundle is better than that in a
enhancement is higher on the lower tubes and reduces along the staggered bundle or vertical bundle. Complex bundle effect
bundle height. Bundle effect is still prominent, especially under was found in special shape bundles, such as V-shape, C-shape,
relatively low heat fluxes. A mixed bundle with enhanced tubes and U-shape bundle, which suggests applying different HTC
only in the lower part can achieve the same heat transfer perfor- correlations to different regions in a bundle.
mance as a fully enhanced bundle based on the combined effect 5. At sub-atmospheric pressure, the bundle effect is undistin-
of surface enhancement and bundle effect. Some researchers devel- guished, and the liquid height has a significant effect on heat
oped different correlations for different surfaces such as Thome transfer behavior. At near-critical pressure, the increase of the
and Robinson (2006), while others used empirical coefficients cor- pressure considerably enhanced the heat transfer both before
responding to surface characteristics such as Swain and Das (2018) and after DNB.
24 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

6. The heat transfer performance of boiling heat transfer on Aprin, L., Mercier, P., Tadrist, L., 2007. Experimental analysis of local void fractions
measurements for boiling hydrocarbons in complex geometry. Int. J. Multiph.
enhanced tube bundles is dominated by the enhanced surface
Flow 33 (4), 371–393.
characteristics and the applied heat flux. A general enhance- Aprin, L., Mercier, P., Tadrist, L., 2011. Local heat transfer analysis for boiling of
ment effect of the enhanced surfaces was observed in boiling hydrocarbons in complex geometries: a new approach for heat transfer
across tube bundles. Bundle effect is still prominent, especially prediction in staggered tube bundle. Int. J. Heat Mass Transf. 54 (19–20),
4203–4219.
under relatively low heat fluxes. The surface enhancement is Ayub, Z.H., Abbas, A., Ayub, A.H., Khan, T.S., Chattha, J.A., 2017. Shell side direct
higher on the lower tubes and reduces along the bundle height. expansion evaporation of ammonia on a plain tube bundle with exit superheat
A mixed bundle with enhanced tubes only in the lower part can effect. Int. J. Refrig 76, 126–135.
Browne, M.W., Bansal, P.K., 1999. Heat transfer characteristics of boiling
achieve the same heat transfer performance as a fully enhanced phenomenon in flooded refrigerant evaporators. Appl. Therm. Eng. 19 (6),
bundle. This type of mixed bundle can reduce the cost and is 595–624.
suggested to be applied to the industrial heat exchangers. Burnside, B.M., Shire, N.F., 2005. Heat transfer in flow boiling over a bundle of
horizontal tubes. Chem. Eng. Res. Des. 83 (5), 527–538.
7. Flow regimes dependent correlations are suggested to be imple- Casciaro, S., Thome, J.R., 2001. Thermal performance of flooded evaporators, part 1:
mented in the local HTC prediction. Shah (2017) correlation Review of boiling heat transfer studies/Discussion. ASHRAE Trans. 107, 903.
owns broad applicability and can provide a reliable prediction Chen, J.C., 1966. Correlation for boiling heat transfer to saturated fluids in
convective flow. Ind. Eng. Chem. Process Des. Dev. 5 (3), 322–329.
of bundle averaged HTC. The application of empirical coeffi- Chen, S., Xiao, Y., Gu, H., 2019. Experimental study on boiling heat transfer in a
cients representing surface characteristics seems a better way three-rod bundle at near-critical pressure. Ann. Nucl. Energy 131, 196–209.
to establish heat transfer coefficient correlations for bundle Chu, K.H., Soo Joung, Y., Enright, R., Buie, C.R., Wang, E.N., 2013. Hierarchically
structured surfaces for boiling critical heat flux enhancement. Appl. Phys. Lett.
boiling with enhanced surfaces.
102, (15) 151602.
Chung, Y.J., Park, H.S., Lee, W.J., Kim, K.K., 2015. Heat transfer in a cooling water pool
with tube bundles under natural circulation. Ann. Nucl. Energy 77, 402–407.
6. Recommendations for future studies Ciloglu, D., Bolukbasi, A., 2015. A comprehensive review on pool boiling of
nanofluids. Appl. Therm. Eng. 84, 45–63.
The review suggests that more experimental studies on boiling Cole, R., Rohsenow, W.M., 1969. Correlation of bubble departure diameters for
boiling of saturated liquids. Chem. Eng. Prog. Symp. Ser. 65 (92), 211–213.
two-phase flow over various bundle geometries and more test flu- Collier, J.G., Thome, J.R., 1994. Convective boiling and condensation. Clarendon
ids (hydrocarbons and refrigerants) are required to establish a pro- Press.
found understanding of flow boiling regimes transition over tube Cooper, M. G., 1984. Heat flow rates in saturated nucleate pool boiling-a wide-
ranging examination using reduced properties. In Advances in heat transfer
bundles. Experimental and analytical studies on the development (Vol. 16, pp. 157-239). Elsevier..
of local HTC correlations with broad applicability are required for Cornwell, K., 1990a. The role of sliding bubbles in boiling on tube bundles. In Proc
the reliable design of heat exchangers. The influences of in-tube 9th International Heat Transfer Conference, Vol. 3, pp. 455–460..
Cornwell, K., 1990b. The influence of bubbly flow on boiling from a tube in a bundle.
hot fluid flow on the external boiling heat transfer over bundles
Int. J. Heat Mass Transf. 33 (12), 2579–2584.
still need more experimental evidence, and the local heat transfer Cornwell, K., Houston, S.D., 1994. Nucleate pool boiling on horizontal tubes: a
behavior should be focused on to extend the understanding of the convection-based correlation. Int. J. Heat Mass Transf. 37, 303–309.
inherent mechanisms and establish a more reliable prediction Cornwell, K., Schüller, R.B., 1982. A study of boiling outside a tube bundle using
high-speed photography. Int. J. Heat Mass Transf. 25 (5), 683–690.
method. The mixed bundle can also be applied in vertical and spe- Dowlati, R., Kawaji, M., Chan, A.M., 1990. Pitch-to-diameter effect on two-phase
cial shape bundles which are commonly used in nuclear installa- flow across an in-line tube bundle. AIChE J. 36 (5), 765–772.
tions to reduce the equipment cost. The number of studies on Dowlati, R., Chan, A.M.C., Kawaji, M., 1992. Hydrodynamics of two-phase flow
across horizontal in-line and staggered rod bundles. J. Fluids Eng. 114 (3), 450–
boiling over enhanced tube bundle is still limited, and the experi- 456.
mental database including more surface categories and more flu- Dowlati, R., Kawaji, M., Chan, A.M.C., 1996. Two-phase crossflow and boiling heat
ids, especially for local heat transfer performance, are necessary transfer in horizontal tube bundles. J. Heat Transfer 118 (1), 124–131.
Feldmann, H., Luke, A., 2008. Nucleate boiling in water for different pressures. Int.
to obtain a comprehensive view. Refrig. Air Cond. Conf. Paper, 982.
Fernández-Seara, J., Pardiñas, Á.Á., Diz, R., 2016. Heat transfer enhancement of
ammonia pool boiling with an integral-fin tube. Int. J. Refrig 69, 175–185.
Declaration of Competing Interest Forster, H.K., Zuber, N., 1955. Dynamics of vapor bubbles and boiling heat transfer.
AIChE J. 1 (4), 531–535.
The authors declare that they have no known competing finan- Fritz, W., 1935. Berechnung des maximalvolumes von dampfblasen. Physik. Zeitschr
36, 379–384.
cial interests or personal relationships that could have appeared Furuya, M., Takiguchi, H., Okawa, R., Arai, T., 2019. Three dimensional void
to influence the work reported in this paper. distribution measurement of salt-water pool-boiling in 5 5 bundle geometry
with X-ray CT system. Ann. Nucl. Energy 129, 207–213.
Giraud, F., Rullière, R., Toublanc, C., Clausse, M., Bonjour, J., 2015. Experimental
Acknowledgments evidence of a new regime for boiling of water at subatmospheric pressure. Exp.
Therm Fluid Sci. 60, 45–53.
Goel, P., Nayak, A.K., Das, M.K., Joshi, J.B., 2018. Bubble departure characteristics in a
The financial support from the Fundamental Research Funds for horizontal tube bundle under crossflow conditions. Int. J. Multiph. Flow 100,
the Central Universities of China, China (No. 45000-18841210), the 143–154.
Gorenflo, D., 1993. Pool Boiling, VDI–Heat. Atlas, VDI–Verlag.
Hong Kong Early Career Scheme Grant, Hong Kong (No.
Gorenflo, D., Baumhögger, E., Herres, G., Kotthoff, S., 2014. Prediction
CityU21202114) and CityU Start-up and Equipment Grants, Hong methods for pool boiling heat transfer: A state-of-the-art review. Int. J. Refrig
Kong (No. 7200343 and No. 9610289) is highly appreciated. 43, 203–226.
Gorgy, E., Eckels, S., 2012. Local heat transfer coefficient for pool boiling of R-134a
and R-123 on smooth and enhanced tubes. Int. J. Heat Mass Transf. 55 (11–12),
References 3021–3028.
Gorgy, E., Eckels, S., 2013. Convective boiling of R-134a and R-123 on an enhanced
tube bundle with standard pitch, RP-1316. HVAC&R Res. 19 (2), 193–206.
Abbas, A., Ayub, Z.H., 2017. Experimental study of ammonia flooded boiling on a
Gorgy, E., Eckels, S., 2016. Convective boiling of R-134a on enhanced-tube bundles.
triangular pitch plain tube bundle. Appl. Therm. Eng. 121, 484–491.
Int. J. Refrig 68, 145–160.
Abbas, A., Ayub, Z.H., Ayub, A.H., Chattha, J.A., 2017a. Shell side direct expansion
Gorgy, E., Eckels, S., 2019. Convective boiling of R-123 on enhanced-tube bundles.
evaporation of ammonia on a plain tube bundle with inlet quality effect in the
International Journal of Heat and Mass Transfer 134, 752–767.
presence of exit superheat. Int. J. Refrig 82, 11–21.
Gupta, A., 2005. Enhancement of boiling heat transfer in a 5 3 tube bundle. Int. J.
Abbas, A., Ayub, Z.H., Ayub, A.H., Khan, T.S., Chattha, J.A., 2017b. Shell side plain tube
Heat Mass Transf. 48 (18), 3763–3772.
bundle performance of a multi-pass direct expansion evaporation of ammonia
Gupta, A., Kumar, R., Kumar, V., 2010. Nucleate pool boiling heat transfer over a
at various degrees of exit superheat. Int. J. Refrig 78, 70–82.
bundle of vertical tubes. Int. Commun. Heat Mass Transfer 37 (2), 178–181.
Abbas, A., Ayub, Z.H., Khan, T.S., Ayub, A.H., Chattha, J.A., 2017c. A review of
Gupte, N.S., Webb, R.L., 1994. Convective vaporization of pure refrigerants in
correlations for outside boiling of ammonia on single tube and bundles. Heat
enhanced and integral-fin tube banks. J. Enhanced Heat Transfer 1 (4).
Transfer Eng., 1–12.
 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278 25

Gupte, N.S., Webb, R.L., 1995a. Shell-side boiling in flooded refrigerant evaporators Liu, Z.H., Qiu, Y.H., 2006. Boiling heat transfer enhancement of water on tubes in
part I: Integral finned tubes. HVAC&R Res. 1 (1), 35–47. compact in-line bundles. Heat Mass Transf. 42 (3), 248–254.
Gupte, N.S., Webb, R.L., 1995b. Convective vaporization data for pure refrigerants in Lockhart, R.W., Martinelli, R.C., 1949. Proposed correlation of data for isothermal
tube banks Part II: enhanced tubes. HVAC&R Res. 1 (1), 48–60. two-phase, two-component flow in pipes. Chem. Eng. Prog. 45 (1), 39–48.
Hahne, E., Spindler, K.K., Chen, Q., Windisch, R., 1990. Local void fraction Lu, D., Yu, Z., Zhong, Y., Wang, H., Zhang, Y., Cao, Q., Gao, S., 2019. Experimental
measurements in finned tube bundles. International Heat Transfer Conference investigation on boiling heat transfer characteristics of the spent fuel bundle
Digital Library. Begel House Inc.. under flooded condition. Nucl. Eng. Des. 344, 168–173.
Hahne, E., Chen, Q.R., Windisch, R., 1991. Pool boiling heat transfer on finned Mao, K., Hibiki, T., 2017. Flow regime transition criteria for upward two-phase
tubes—an experimental and theoretical study. Int. J. Heat Mass Transf. 34 (8), cross-flow in horizontal tube bundles. Appl. Therm. Eng. 112, 1533–1546.
2071–2079. Memory, S.B., Chilman, S.V., Marto, P.J., 1994. Nucleate pool boiling of a TURBO-B
Hahne, E., Müller, J., 1983. Boiling on a finned tube and a finned tube bundle. Int. J. bundle in R-113. J. Heat Transfer 116 (3), 670–678.
Heat Mass Transf. 26 (6), 849–859. McNeil, D.A., Burnside, B.M., Elsaye, E.A., Salem, S.M., Baker, S., 2017. Shell-side
Hsieh, S.S., Huang, G.Z., Tsai, H.H., 2003. Nucleate pool boiling characteristics from boiling of a glycerol-water mixture at low sub-atmospheric pressures. Appl.
coated tube bundles in saturated R-134a. Int. J. Heat Mass Transf. 46 (7), 1223– Therm. Eng. 115, 1438–1450.
1239. McNeil, D.A., Burnside, B.M., Rylatt, D.I., Elsaye, E.A., Baker, S., 2015. Shell-side
Hsu, J.T., Lin, C.S., Jensen, M.K., 1993. Boiling heat transfer mechanisms in a boiling of water at sub-atmospheric pressures. Int. J. Heat Mass Transf. 85, 488–
horizontal tube bundle. Exp. Heat Transf. 6 (3), 259–271. 504.
Ishibashi, E., 2001. Enhanced boiling heat transfer of water/salt mixtures in the Mitrovic, J., Ustinov, A., 2006. Nucleate boiling heat transfer on a tube provided with
restricted space of the compact tube bundle. Heat Transfer Eng. 22 (3), 4–10. a novel microstructure. J. Enhanced Heat Transfer 13 (50), 261–278.
Jensen, M. K., & Hsu, J. T., 1987. A parametric study of boiling heat transfer in a tube Mohanty, R.L., Das, M.K., 2017. A critical review on bubble dynamics
bundle. In Proc. 2nd ASME/JSME Thermal Engineering Joint Conf (Vol. 3, pp. parameters influencing boiling heat transfer. Renew. Sustain. Energy Rev. 78,
132–140).. 466–494.
Jeon, S.S., Hong, S.J., Cho, H.K., Park, G.C., 2015. Prediction of nucleate boiling heat Mostinski, I.L., 1963. Application of the rule of corresponding states for calculation
transfer on horizontal U-shaped heat exchanger submerged in a pool of water of heat transfer and critical heat flux. Teploenergetika 4 (4), 66–71.
using MARS code. Nucl. Eng. Des. 295, 317–337. Murakawa, H., Baba, M., Miyazaki, T., Sugimoto, K., Asano, H., Ito, D., 2018. Local
Jeon, S.S., Hong, S.J., Cho, H.K., Park, G.C., 2016. Development of heat transfer model void fraction and heat transfer characteristics around tubes in two-phase flows
package for horizontal u-shaped heat exchanger submerged in pool of passive across horizontal in-line and staggered tube bundles. Nucl. Eng. Des. 334, 66–
safety system. Nucl. Technol. 196 (2), 303–318. 74.
Ji, W.T., Zhao, C.Y., Zhang, D.C., Yoshioka, S., He, Y.L., Tao, W.Q., 2016. Effect of vapor Noghrehkar, G.R., Kawaji, M., Chan, A.M.C., 1999. Investigation of two-phase flow
flow on the falling film evaporation of R134a outside a horizontal tube bundle. regimes in tube bundles under cross-flow conditions. Int. J. Multiph. Flow 25
Int. J. Heat Mass Transf. 92, 1171–1181. (5), 857–874.
Kaichiro, M., Ishii, M., 1984. Flow regime transition criteria for upward two-phase Nukiyama, S., 1966. The maximum and minimum values of the heat Q transmitted
flow in vertical tubes. Int. J. Heat Mass Transf. 27 (5), 723–737. from metal to boiling water under atmospheric pressure. Int. J. Heat Mass
Kang, M.G., 2015a. Effects of elevation angle on pool boiling heat transfer of tandem Transf. 9 (12), 1419–1433.
tubes. Int. J. Heat Mass Transf. 85, 918–923. O’Hanley, H., Coyle, C., Buongiorno, J., McKrell, T., Hu, L.-W., Rubner, M., Cohen, R.,
Kang, M.G., 2015b. Pool boiling heat transfer on tandem tubes in vertical alignment. 2013. Separate effects of surface roughness, wettability, and porosity on the
Int. J. Heat Mass Transf. 87, 138–144. boiling critical heat flux. Appl. Phys. Lett. 103, 024102.
Kang, M.G., 2016. Pool boiling heat transfer from an inclined tube bundle. Int. J. Heat Phan, H.T., Caney, N., Marty, P., Colasson, S., Gavillet, J., 2009. Surface wettability
Mass Transf. 101, 445–451. control by nanocoating: the effects on pool boiling heat transfer and nucleation
Kang, M.G., 2017. Effects of included angle on pool boiling heat transfer of V-shape mechanism. Int. J. Heat Mass Transf. 52 (23–24), 5459–5471.
tubes in vertical alignment. Int. J. Heat Mass Transf. 108, 901–906. Polley, G.T., 1980. Reboilers. Developments in Heat Exchanger Technology, Part, p. 1.
Kang, M.G., 2018. Variation of pool boiling heat transfer due to rotation angle of V- Ribatski, G., Jabardo, J.M.S., 2003. Experimental study of nucleate boiling of
shape tube array. Int. J. Heat Mass Transf. 125, 788–791. halocarbon refrigerants on cylindrical surfaces. Int. J. Heat Mass Transf. 46
Kanizawa, F.T., Ribatski, G., 2016. Two-phase flow patterns across triangular tube (23), 4439–4451.
bundles for air-water upward flow. Int. J. Multiph. Flow 80, 43–56. Ribatski, G., Jabardo, J.M.S., Da Silva, E.F., 2008. Modeling and experimental study of
Kanizawa, F.T., Ribatski, G., 2017a. Void fraction and pressure drop during external nucleate boiling on a vertical array of horizontal plain tubes. Exp. Therm Fluid
upward two-phase crossflow in tube bundles–Part I: Experimental Sci. 32 (8), 1530–1537.
investigation. Int. J. Heat Fluid Flow 65, 200–209. Ribatski, G., Thome, J.R., 2005. Dynamics of two-phase flow across horizontal tube
Kanizawa, F.T., Ribatski, G., 2017b. Void fraction and pressure drop during external bundles-a review. Rev. Eng. Term. 4 (2).
upward two-phase cross flow in tube bundles–Part II: Predictive methods. Int. J. Ribatski, G., Thome, J.R., 2007. Two-phase flow and heat transfer across horizontal
Heat Fluid Flow 65, 210–219. tube bundles-a review. Heat Transfer Eng. 28 (6), 508–524.
Kim, N.H., Cho, J.P., Youn, B., 2002. Forced convective boiling of pure refrigerants in a Robinson, D.M., Thome, J.R., 2004a. Local bundle boiling heat transfer coefficients on
bundle of enhanced tubes having pores and connecting gaps. Int. J. Heat Mass a plain tube bundle (RP-1089). HVAC&R Res. 10 (1), 33–51.
Transf. 45 (12), 2449–2463. Robinson, D.M., Thome, J.R., 2004b. Local bundle boiling heat transfer
Kim, J., Jun, S., Laksnarain, R., You, S.M., 2016. Effect of surface roughness on pool coefficients on an integral finned tube bundle (RP-1089). HVAC&R Res. 10 (3),
boiling heat transfer at a heated surface having moderate wettability. 331–344.
International Journal of Heat and Mass Transfer 101, 992–1002. Robinson, D.M., Thome, J.R., 2004c. Local bundle boiling heat transfer coefficients on
Kim, D.E., Yu, D.I., Jerng, D.W., Kim, M.H., Ahn, H.S., 2015. Review of boiling heat a turbo-BII HP tube bundle (RP-1089). HVAC&R Res. 10 (4), 441–457.
transfer enhancement on micro/nanostructured surfaces. Exp. Therm Fluid Sci. Rohsenow, W.M., 1951. A Method of Correlating Heat Transfer Data for Surface
66, 173–196. Boiling of Liquids. MIT Division of Industrial Cooperation, Cambridge, Mass.
Kocamustafaogullari, G., Ishii, M., 1983. Interfacial area and nucleation site density Schäfer, D., Tamme, R., Müller-Steinhagen, H., 2007. The effect of novel plasma-
in boiling systems. Int. J. Heat Mass Transf. 26 (9), 1377–1387. coated compact tube bundles on pool boiling. Heat Transfer Eng. 28 (1), 19–24.
Kukulka, D.J., Smith, R., 2014. Heat transfer evaluation of an enhanced heat transfer Shah, M.M., 2017. A correlation for heat transfer during boiling on bundles of
tube bundle. Energy 75, 97–103. horizontal plain and enhanced tubes. Int. J. Refrig 78, 47–59.
Lakhera, V.J., Gupta, A., Kumar, R., 2012. Enhanced boiling outside 8 3 plain and Shah, M.M., 2007. A general correlation for heat transfer during saturated boiling
coated tube bundles. Heat Transfer Eng. 33 (9), 828–834. with flow across tube bundles. HVAC&R Res. 13 (5), 749–768.
Lakhera, V.J., Akhilesh, G., Ravi, K., 2014. Investigations on pool boiling outside Stephan, K., Abdelsalam, M., 1980. Heat-transfer correlations for natural convection
horizontal tube bundles. International Conference on Heat Transfer, Fluid boiling. Int. J. Heat Mass Transf. 23 (1), 73–87.
Mechanics and Thermodynamics. Swain, A., Das, M.K., 2014. A review on saturated boiling of liquids on tube bundles.
Leong, K.C., Ho, J.Y., Wong, K.K., 2017. A critical review of pool and flow boiling heat Heat Mass Transf. 50 (5), 617–637.
transfer of dielectric fluids on enhanced surfaces. Appl. Therm. Eng. 112, 999– Swain, A., Das, M.K., 2017. Flow boiling of distilled water over plain tube bundle
1019. with uniform and varying heat flux along the height of the tube bundle. Exp.
Liao, L., Liu, Z.H., 2007. Enhanced boiling heat transfer of the compact staggered Therm Fluid Sci. 82, 222–230.
tube bundles under sub-atmospheric pressures. Heat Transfer Eng. 28 (5), 444– Swain, A., Mohanty, R.L., Das, M.K., 2017. Pool boiling of distilled water over tube
450. bundle with variable heat flux. Heat Mass Transf. 53 (8), 2487–2495.
Liu, H., Hibiki, T., 2017. Flow regime transition criteria for upward two-phase flow Swain, A., Das, M.K., 2018. Performance of porous coated 5 3 staggered horizontal
in vertical rod bundles. Int. J. Heat Mass Transf. 108, 423–433. tube bundle under flow boiling. Appl. Therm. Eng. 128, 444–452.
Liu, Z.H., Liao, L., 2006. Enhancement boiling heat transfer study of a newly compact Swain, A., Mohanty, R.L., Das, M.K., 2018. Flow boiling under variable heat flux along
in-line bundle evaporator under reduced pressure conditions. Chem. Eng. the height over coated tube bundle. Exp. Therm Fluid Sci. 97, 89–93.
Technol.: Ind. Chem.-Plant Equip.-Process Eng.-Biotechnol. 29 (3), 408–413. Swain, A., Das, M.K., 2019. Flow boiling over tube bundles with combination of plain
Liu, Z.H., Qiu, Y.H., 2002. Enhanced boiling heat transfer in restricted spaces of a and coated tubes. J. Thermophys Heat Transfer 33 (2), 559–567.
compact tube bundle with enhanced tubes. Appl. Therm. Eng. 22 (17), 1931– Takata, Y., Hidaka, S., Cao, J.M., Nakamura, T., Yamamoto, H., Masuda, M., Ito, T.,
1941. 2005. Effect of surface wettability on boiling and evaporation. Energy 30 (2–4),
Liu, Z.H., Qiu, Y.H., 2004. Boiling heat transfer enhancement of water/salt mixtures 209–220.
on roll-worked enhanced tubes in compact staggered tube bundles. Chem. Eng. Taylor, C.E., Pettigrew, M.J., 2001. Effect of flow regime and void fraction on tube
Technol.: Ind. Chem.-Plant Equip.-Process Eng.-Biotechnol. 27 (11), 1187–1194. bundle vibration. J. Pressure Vessel Technol. 123 (4), 407–413.
26 S. Ren, W. Zhou / Annals of Nuclear Energy 139 (2020) 107278

Thome, J.R., Robinson, D.M., 2006. Prediction of local bundle boiling heat transfer Xu, G.P., Tso, C.P., Tou, K.W., 1998. Hydrodynamics of two-phase flow in vertical up
coefficients: pure refrigerant boiling on plain, low fin, and turbo-BII HP tube and down-flow across a horizontal tube bundle. Int. J. Multiph. Flow 24 (8),
bundles. Heat Transfer Eng. 27 (10), 20–29. 1317–1342.
Ulbrich, R., Mewes, D., 1994. Vertical, upward gas-liquid two-phase flow across a Zeng, X., Chyu, M.C., Ayub, Z.H., 2001. Experimental investigation on ammonia
tube bundle. Int. J. Multiph. Flow 20 (2), 249–272. spray evaporator with triangular-pitch plain-tube bundle, Part II: evaporator
Ustinov, A., Ustinov, V., Mitrovic, J., 2011. Pool boiling heat transfer of tandem tubes performance. Int. J. Heat Mass Transf. 44 (11), 2081–2092.
provided with the novel microstructures. Int. J. Heat Fluid Flow 32 (4), 777–784. Zhang, K., Hou, Y.D., Tian, W.X., Fan, Y.Q., Su, G.H., Qiu, S.Z., 2017a. Experimental
Van Rooyen, E., 2011. Boiling on a Tube Bundle: Heat Transfer, Pressure Drop and investigations on single-phase convection and steam-water two-phase flow
Flow Patterns Ph.D. diss. École Polytechnique Fédérale de Lausanne, Lausanne, boiling in a vertical rod bundle. Exp. Therm Fluid Sci. 80, 147–154.
Switzerland. Zhang, Y., Lu, D., Wang, Z., Fu, X., Cao, Q., Yang, Y., 2017b. Experimental investigation
Van Rooyen, E., Agostini, F., Borhani, N., Thome, J.R., 2012. Boiling on a tube bundle: on pool-boiling of C-shape heat exchanger bundle used in PRHR HX. Appl.
part II—heat transfer and pressure drop. Heat Transfer Eng. 33 (11), 930–946. Therm. Eng. 114, 186–195.
Van Rooyen, E., Thome, J.R., 2013. Pressure drop data and prediction method for Zhang, K., Hou, Y.D., Tian, W.X., Zhang, Y.P., Su, G.H., Qiu, S.Z., 2018. Experimental
enhanced external boiling tube bundles with R-134a and R-236fa. Int. J. Refrig investigation on steam-water two-phase flow boiling heat transfer in a
36 (6), 1669–1680. staggered horizontal rod bundle under cross-flow condition. Exp. Therm Fluid
Van Rooyen, E., Thome, J.R., 2014. Flow boiling data and prediction method for Sci. 96, 192–204.
enhanced boiling tubes and tube bundles with R-134a and R-236fa including a Zhang, K., Hou, Y.D., Tian, W.X., Zhang, Y.P., Su, G.H., Qiu, S.Z., 2019. Experimental
comparison with falling film evaporation. Int. J. Refrig 41, 60–71. investigations on single-phase convection and two-phase flow boiling heat
Venkateswararao, P., Semiat, R., Dukler, A.E., 1982. Flow pattern transition for gas- transfer in an inclined rod bundle. Appl. Therm. Eng. 148, 340–351.
liquid flow in a vertical rod bundle. Int. J. Multiph. Flow 8 (5), 509–524. Zhao, C.Y., Ji, W.T., Jin, P.H., Tao, W.Q., 2018a. Cross vapor stream effect on falling
Wang, C.H., Dhir, V.K., 1993. Effect of surface wettability on active nucleation site film evaporation in horizontal tube bundle using R134a. Heat Transfer Eng. 39
density during pool boiling of water on a vertical surface. J. Heat Transfer 115 (7–8), 724–737.
(3), 659–669. Zhao, C.Y., Ji, W.T., Jin, P.H., Yoshioka, S., Tao, W.Q., 2018b. Effect of downward vapor
Wang, H., Wang, W., Bi, Q., 2017. Experimental investigation on boiling heat stream on falling film evaporation of R134a in a tube bundle. Int. J. Refrig 89,
transfer of high-pressure water in a SCWR sub-channel. Int. J. Heat Mass Transf. 112–121.
105, 799–810. Zhao, C.Y., Ji, W.T., Jin, P.H., Zhong, Y.J., Tao, W.Q., 2018c. Experimental study of the
Webb, R.L., 1981. The evolution of enhanced surface geometries for nucleate local and average falling film evaporation coefficients in a horizontal enhanced
boiling. Heat Transfer Eng. 2 (3–4), 46–69. tube bundle using R134a. Appl. Therm. Eng. 129, 502–511.
Webb, R.L., Chien, L.H., 1994. Correlation of convective vaporization on banks of Zheng, J.X., Jin, G.P., Chyu, M.C., Ayub, Z.H., 2001. Flooded boiling of ammonia with
plain tubes using refrigerants. Heat Transfer Eng. 15 (3), 57–69. miscible oil outside a horizontal plain tube. HVAC&R Research 7 (2), 185–204.
Weisman, J., 1959. Heat transfer to water flowing parallel to tube bundles. Nucl. Sci. Zhou, Y., Hou, Y., Li, H., Sun, B., Yang, D., 2015. Flow pattern map and multi-scale
Eng. 6 (1), 78–79. entropy analysis in 3 3 rod bundle channel. Ann. Nucl. Energy 80, 144–150.
Williams, C.L., Peterson Jr, A.C., 1978. Two-phase flow patterns with high-pressure Žukauskas, A., 1972. Heat transfer from tubes in crossflow. In: Advances in heat
water in a heated four-rod bundle. Nucl. Sci. Eng. 68 (2), 155–169. transfer. Elsevier, pp. 93–160.