Dissertation Tenzer

Heat transfer during transient spray cooling:
An experimental and analytical study
Vom Fachbereich Maschinenbau

der Technischen Universität Darmstadt
zur
Erlangung des Grades eines Doktor-Ingenieurs (Dr.-Ing.)
genehmigte
Dissertation
vorgelegt von
Fabian Michael Tenzer, M.Sc.
aus Heppenheim
Berichterstatter: Prof. Dr.-Ing. Cameron Tropea

1. Mitberichterstatter: Apl. Prof. Dr. Ilia V. Roisman
2. Mitberichterstatter: Prof. Dr.-Ing. Eckehard Specht
Tag der Einreichung: 20.01.2020
Tag der mündlichen Prüfung: 26.03.2020
Darmstadt 2020
D 17
Tenzer, Fabian Michael:
Heat transfer during transient spray cooling: An experimental and analytical
study
Darmstadt, Technische Universität Darmstadt
Jahr der Veröffentlichung der Dissertation auf TUprints: 2020
Tag der mündlichen Prüfung: 26.03.2020
Bitte zitieren Sie dieses Dokument als:

URN: urn:nbn:de:tuda-tuprints-113442
URL: https://tuprints.ulb.tu-darmstadt.de/id/eprint/11344
Dieses Dokument wird bereitgestellt von TU Prints, E-Publishing-Service

der Technischen Universität Darmstadt http://tuprints.ulb.tu-darmstadt.de
tuprints@ulb.tu-darmstadt.de
Veröffentlicht unter CC BY-SA 4.0 International:

Namensnennung - Weitergabe unter gleichen Bedingungen
https://creativecommons.org/licenses/by-sa/4.0/
Für Tom
Abstract
Spray cooling features a very high, homogeneously distributed cooling per-

formance. Therefore it is used in various industrial applications, like cooling
of high powered electronics, for quenching during metalworking or cooling of
tools during hot forging. The cooling efficiency is influenced by a large number
of parameters: drop diameter and velocity, mass flux, surface temperature,
spray fluid and temperature, surface material and conditions, etc. The entire
process is extremely complex and to date only few physical models exist
describing the heat transfer rates as a function of these parameters. Instead,
the heat flux is mainly predicted using empirical correlations, which are often
not suited for conditions other than those from which they were derived -
universality of these correlations is lacking. The present study strives to
replace these empirical correlations with theories based on physics.
In this study the transient spray cooling of a hot thick target is experimen-
tally and analytically investigated. The locally resolved temporal evolution of
the heat flux and surface temperature of an initially homogeneously heated
substrate is measured during continuous spray impact. This experimental
quantification captures the influence of various spray features (droplet diame-
ter, velocity and mass flux), spray impact angle, spray fluid temperature, wall
thermal properties and wall surface roughness, whereby the mass flux is found
to be the dominating factor in determining cooling performance. Furthermore,
visual observations of spray impact at high surfaces temperatures beyond the
Leidenfrost point identifies no closed liquid film separated from the surface
by a vapor layer, as it is often imagined for these conditions. Instead, the
spray impact is governed by a superposition of single drop impacts at a dry
wall. Therefore, there is no risk of unintentionally flooding the surface, thus
limiting the heat flux due to an inordinately high mass flux. Spraying a
surface at an oblique angle, as well as spraying with hot fluid decreases the
heat flux. Increasing the surface roughness or using substrates which have a
high thermal conductivity results in higher cooling performances.
i
Theoretical models to predict the heat flux during spray cooling are devel-
oped and validated with experimental results. A model for the film boiling
regime accounts for spray and wall properties and predicts the temporal evo-
lution of surface temperature and heat flux. It agrees well with experimental
data.
A theory for the nucleate boiling regime indicates that the heat flux is
limited by the thermal inertia of the substrate material and is not a function
of the spray properties.
Furthermore, the Leidenfrost point is found to be nearly independent of
the spray properties. Instead, it is strongly influenced by the material of
the substrate and fluid: A high thermal effusivity leads to a low Leidenfrost
temperature and vice versa. This influence is captured in a newly developed
theoretical prediction.
ii
Kurzfassung
Die Sprühkühlung zeichnet sich durch eine sehr hohe und gleichmäßig verteilte
Kühlleistung aus. Deshalb wird sie in einer großen Bandbreite von industriellen
Prozessen eingesetzt, wie beispielsweise bei der Kühlung von Hochleistungs-
elektronik, zum Abschrecken während der Metallherstellung oder zur Kühlung
von Schmiedewerkzeugen in der Massivumformung. Die Effizienz der Kühlung
wird von einer Vielzahl von Größen beeinflusst. Dazu zählen unter anderem:
Tropfendurchmesser und -geschwindigkeit, Beaufschlagungsdichte, Oberflä-
chentemperatur, das Spraymedium sowie dessen Temperatur, Material und
Beschaffenheit der Oberfläche und viele mehr. Der gesamte Prozess ist extrem
komplex und bis jetzt existieren nur sehr wenige physikalische Modelle, die
diesen beschreiben. Stattdessen müssen zur Vorhersage des Wärmestroms
empirische Korrelationen verwendet werden, die auf Grund von fehlender
Universalität nicht immer zur entsprechenden Anwendung passen.
In dieser Arbeit wird die transiente Sprühkühlung eines heißen, dicken

Körpers experimentell und analytisch untersucht. Hierbei wird die zeitliche
Entwicklung der Wärmestromdichte und der Oberflächentemperatur während
der kontinuierlichen Sprühkühlung des zu Beginn homogen erwärmten Kör-
pers gemessen. Diese experimentelle Quantifizierung umfasst den Einfluss
verschiedener Arten des Sprays (Tropfendurchmesser, Geschwindigkeit und
Beaufschlagungsdichte), des Winkels des Sprayaufpralls, der Wassertempera-
tur des Sprays, der thermischen Eigenschaften des gekühlten Materials und
der Rauigkeit der Oberfläche. Die Ergebnisse zeigen, dass die Beaufschlagungs-
dichte die maßgebende Einflussgröße für eine hohe Kühlleistung ist. Weiterhin
zeigen die visuellen Beobachtungen keinen geschlossenen Wasserfilm, der durch
einen Dampffilm von der Oberfläche getrennt ist. Somit ergibt sich in den
meisten Anwendungen kein Risiko, durch eine zu große Beaufschlagungsdichte
die Oberfläche zu überschwemmen und damit den Wärmestrom zu limitieren.
iii
Die Kühlung einer schrägen Oberfläche sowie das Kühlen mit heißem Wasser
reduziert die Kühlleistung, während Oberflächenrauigkeit und Verwendung
eines Substrats mit hoher Wärmeleitfähigkeit die Wärmestromdichte erhöhen.
Weiterhin wurden während dieser Arbeit theoretische Modelle zur Vorhersa-
ge der Wärmestromdichte entwickelt und mit den experimentellen Ergebnissen
validiert. Ein Modell für das „film boiling regime“ berücksichtigt den Einfluss
verschiedener Sprays und Substrateigenschaften. Es beschreibt die zeitliche
Entwicklung der Wärmestromdichte und Oberflächentemperatur und stimmt
sehr gut mit den experimentellen Ergebnissen überein.
Weiterhin zeigt eine Theorie für das „nucleate boiling regime“, dass die
Wärmestromdichte durch die thermische Trägheit des Substrats limitiert wird
und nahezu unabhängig von den Sprayeigenschaften ist.
Der Leidenfrostpunkt ist ebenfalls unabhängig von den Sprayeigenschaften.
Stattdessen wird er stark vom Material des Substrats und Fluids beeinflusst:
Ein großer Wärmeeindringkoeffizient erzeugt eine geringe Leidenfrosttempera-
tur und umgekehrt. Dieser Einfluss wird durch die entwickelte Theorie gut
vorhergesagt.
iv
Danksagung
Zuallererst möchte ich mich bei Prof. Dr.-Ing. Cameron Tropea für die Anstel-
lung am Fachgebiet „Strömungslehre und Aerodynamik“ und die Möglichkeit
diese Dissertation anzufertigen bedanken. Ich konnte mich jederzeit auf seine
Zusagen und einen strukturellen Rahmen verlassen, der wenige bis keine
Einschränkungen bedeutete. Weiterhin geht mein ganz besonderer Dank an
Apl. Prof. Dr. Ilia V. Roisman, der mir beim Verständnis der physikalischen
Vorgänge während der Sprühkühlung und deren Modellierung essenziell gehol-
fen hat. Die freudvollen Diskussionen endeten häufig mit neuen Erkenntnissen
und manchmal nur mit noch mehr Fragen. Bei Prof. Dr.-Ing. Eckehard Specht
möchte ich mich vielmals für die Übernahme des Korreferats dieser Arbeit
bedanken.
Weiterhin möchte ich mich bei der Deutschen Forschungsgemeinschaft
(DFG) bedanken, die diese Forschung im Rahmen des Sonderforschungsbe-
reichs SFB-TRR 75 erst finanziell möglich gemacht hat. Ohne die tat- und
sachkräftige Förderung durch den Industrieverband Massivumformung e.V.
wäre dies ebenfalls nicht möglich gewesen. Vielen Dank hierfür und die vielen
Projekttreffen der Patengruppe, die immer wieder aufs Neue den Bezug zur
industriellen Anwendung sichergestellt haben.
Ich danke weiterhin den Studenten Julian Hofmann, Till Kaupe, Minh
Khang Pham und Sebastian Wolter, die durch ihre Mühe und ihren Ehrgeiz
diese Arbeit wesentlich unterstützt haben. Ein ganz besonderer Dank geht
an die Griesheimer Jungs, die durch zahlreiche Diskussionen, gemeinsame
Kochaktionen zur Mittagspause, allerlei Unfug während und nach der Arbeit
und ein ganz besonderes Arbeitsklima das Arbeiten in der Exklave lebens-
wert, äußerst produktiv und sozial verträglich gemacht haben. Dazu zählen:
Andreas Bauer, Alexander Beck, Jan Breitenbach, Sebastian Brulin, Johannes
Feldmann, Julian Hofmann, Till Kaupe, Johannes Kissing, Maximilian Kuhn-
henn, Max Luh, Klaus Schiffmann, Benedikt Schmidt, Bernhard Simon und
Martin Stenger. Danke für die immer wiederkehrende gegenseitige Motivation!
v
Weiterhin danke ich dem Werkstatt-Team in Griesheim um Timm Geelhaar,
Joachim Heyl, Ilona Kaufhold und Martin Weiß für die Unterstützung beim
Bau der Versuchsanlage.
Obwohl zuletzt genannt, geht mein allergrößter Dank an meine Familie
und Freunde. Ich möchte meinen Eltern Georg und Marina Tenzer und
meiner Schwester Anne Tenzer für die nicht endende Unterstützung und
den immer währenden Rückhalt danken. Ihr habt durch eure Erziehung und
Verbundenheit den Grundstein für alles gelegt. Mein größter Dank geht an
Meike, die mich unzählige Male bestärkt hat weiterzumachen sowie sich ebenso
oft bahnbrechenden Erkenntnisse über Sprühkühlung anhören durfte, was sie
sicherlich immer brennend interessiert hat. Ohne dich wäre mir vieles deutlich
schwerer gefallen.
Danke,
Fabian
vi
Contents
Abstract i
Kurzfassung iii
Danksagung v
1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Fundamentals of spray cooling . . . . . . . . . . . . . . . . . 4
1.3 Objective and outline of this thesis . . . . . . . . . . . . . . . 10
2 Experimental setup and methods 13

2.1 Overview of the setup . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Spraying test rig . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2.1 Phase Doppler technique . . . . . . . . . . . . . . . . 16
2.2.2 Patternator . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.3 Results of experimental spray characterization . . . . 18
2.3 Cooling test rig . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.3.1 Mechanical and thermodynamic construction . . . . . 24
2.3.2 Measurement technique for heat flux and surface tem-
perature . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3.3 Observation system . . . . . . . . . . . . . . . . . . . 29
3 Typical results of transient spray cooling 31

3.1 Measurements of the heat flux . . . . . . . . . . . . . . . . . . 31
3.2 Visualization of the spray impact . . . . . . . . . . . . . . . . 34
3.3 Influence of various parameters . . . . . . . . . . . . . . . . . 40
3.3.1 Spray impact parameters . . . . . . . . . . . . . . . . 42
3.3.2 Spray impact angle . . . . . . . . . . . . . . . . . . . . 45
3.3.3 Initial spray water temperature . . . . . . . . . . . . . 47
3.3.4 Initial substrate temperature . . . . . . . . . . . . . . 54
vii
Contents
3.3.5 Wall thermal properties . . . . . . . . . . . . . . . . . 55

3.3.6 Wall surface roughness . . . . . . . . . . . . . . . . . . 58
4 Heat transfer in film boiling regime 61

4.1 Analysis of heat transfer during spray cooling . . . . . . . . . 61
4.2 Modelling of temperature and heat flux evolution . . . . . . . 62
4.3 Model validation . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.4 Application of the model . . . . . . . . . . . . . . . . . . . . . 71
4.5 Transitional spray cooling regime below the Leidenfrost point 74
5 Heat transfer in nucleate boiling regime 77

5.1 Modelling heat flux as a remote asymptotic solution . . . . . 77
5.2 Model validation . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.3 Heat flux and its upper boundary . . . . . . . . . . . . . . . . 82
5.4 Application of the model . . . . . . . . . . . . . . . . . . . . . 83
6 Leidenfrost point 87
6.1 Influences on the Leidenfrost point . . . . . . . . . . . . . . . 87
6.2 Comparison with literature . . . . . . . . . . . . . . . . . . . 91
6.3 Theoretical prediction of the Leidenfrost point . . . . . . . . 94
7 Conclusion and outlook 103
Nomenclature 107
Bibliography 111
List of Figures 123
List of Tables 127
viii
1 Introduction
The present chapter provides a short introduction into the topic of spray
cooling. The first part gives a motivation, why research and progress in the
field of spray cooling is of interest for a broad community. In the subsequent
section an overview of the literature indicates questions which remain open.
The present chapter ends with the objectives and outline of this thesis. Parts
of this chapter have already been published in Tenzer et al. (2019b) and
appeared in Tenzer et al. (2020).
1.1 Motivation
Spray cooling is a process capable of achieving a very high, nearly uniform
heat flux, and therefore high cooling performance. A comparison of the heat
transfer coefficients of different cooling technologies is shown in Fig. 1.1. Of all
shown technologies spray cooling clearly has the highest cooling performance.
Therefore it is used in various industrial applications, like cooling of micro-
chips and other high powered electronics or electrical parts (Mudawar, 2001;
Bar-Cohen et al., 2006; Ebadian & Lin, 2011), cooling of metal products in
metallurgy during quenching processes, in metalworking (Chen & Tseng, 1992),
cooling of tools for hot forging (Pola et al., 2013), of solar panels (Nižetić
et al., 2016; Sargunanathan et al., 2016) and in many other technological
processes.
These completely different applications indicate the complex nature and
versatility of the term spray cooling. A rough classification regarding the
temperature range, range of spray properties and temporal evolution helps
to understand the different features of each application. For example, spray
cooling of high power electronics is performed at rather low surface temper-
atures. The mass flux as well as the sprayed surface is small. The process
is stationary, therefore the surface temperature remains nearly constant. In
contrast, spray cooling for quenching during steel-making exhibits a com-
1
1 Introduction
Figure 1.1: Heat transfer coefficient of different cooling technologies. (Adapted

from Sienski et al. (1996), with permission of IEEE. © 1996 IEEE.)
pletely different behavior. The surface temperatures are very high. The mass
flux and the sprayed surface are large. The process is transient, meaning the
surface temperature continuously decreases with increasing time and therefore
different hydrodynamic and thermodynamic phenomena occur over time at
the surface.
The motivation for this project is focused on cooling during hot forging.
Here metal blanks are heated up to 1260 ◦ C and pressed between forging dies
to reach the desired geometry. This process is used for the production of
various parts that require high strength. Exemplary parts are crankshafts,
gears and other parts used in industrial machinery and consumer products.
Since the heated parts stay in contact with the forging dies, the temperature
of the dies also rise; therefore, to prevent overheating and damage they need
to be cooled. This is most often achieved by spray cooling. By shortening the
necessary cooling time, the entire process time for the production of each part
can be reduced. This directly leads to an increased productivity and economic
efficiency. One goal of the present work is to help to overcome the bottleneck
of spray cooling by achieving a better understanding of the physical processes
which determine the cooling rate and by developing reliable models describing
these processes, eventually leading to shortened cycle times in the cooling
phase.
2
1.1 Motivation
Figure 1.2: Exemplary temporal evolution of the temperature 1 mm below the

surface of a workpiece during a hot forging process.
To gain a first impression of the temperatures and timescales during hot

forging a preliminary experiment during a hot forging process was performed.
Cylindrical steel blanks (60 mm diameter, 90 mm height) having a temperature
of 1230 ◦ C were compressed to disks (110 mm diameter, 30 mm height). The
corresponding temperature inside the tool was acquired using a thermocouple,
placed 1 mm below the surface.
In Fig. 1.2 the temperature reading from this experiment is plotted as a

function of time. The corresponding process steps at the surface are indicated
in the diagram. We can clearly identify each of the steps from the temperature
response inside the target. The highest temperature of approximately 350 ◦ C
is reached at the end of the pressing throw. Although the temperature reading
is not exactly the surface temperature, it is representative. This preliminary
example experiment leads to a general impression of the temperature range
this study should cover. Therefore, the test facility developed during this
study is designed to operate at surface temperatures of up to 500 ◦ C, as will
be shown in Section 2.3.
3
1 Introduction
1.2 Fundamentals of spray cooling

The phenomena of spray impingement onto a very hot substrate is significantly
influenced by the wall temperature since the flow generated by each drop
impact is influenced by various micro-scale thermodynamic effects, governed
by intensive evaporation. The hydrodynamics and heat transfer during single
drop impact onto a heated wall have been extensively investigated (Chandra &
Avedisian, 1991; Bernardin et al., 1997; Bertola, 2015; Staat et al., 2015). The
regimes of single drop impact observed in the experiments include single phase
cooling, nucleate boiling, transition regime, thermal atomization (Roisman
et al., 2018) and film boiling, the latter case occuring if the wall temperature
rises above the Leidenfrost condition.
For the investigation of heat transfer during spray cooling it is often rea-
sonable to plot the heat flux as a function of the surface temperature, as
illustrated in Fig. 1.3. By representing the data in this way, the typical boiling
regimes, which are known from pool boiling experiments (Nukiyama, 1966),
become apparent: At high surface temperatures in the film boiling regime the
heat flux decreases slightly with deceasing surface temperature. At a certain
surface temperature the heat flux reaches its minimum. This point is called
the Leidenfrost point. Accordingly, the equivalent surface temperature is
called Leidenfrost temperature. At surface temperatures above the Leidenfrost
point, a droplet that is placed on the surface has no contact to the surface,
but is separated from it by a vapor layer (Leidenfrost, 1966). At temperatures
below the Leidenfrost point the heat flux strongly increases with decreasing
surface temperature. This regime is called the transition boiling regime. It
ends at the critical heat flux which is the maximum of the curve. In the
following nucleate boiling regime the heat flux again continuously decreases
with deceasing surface temperature.
A detailed review of the current state of the art concerning the prediction
of heat transfer during spray cooling is dispensed with here since there are
many comprehensive reviews available in the literature: Liang & Mudawar
(2017b,c); Cheng et al. (2016); Kim (2007); Breitenbach et al. (2018b). Many
studies deal with the influence of various parameters on the performance
of spray cooling. These studies are mostly focused on the determination of
the critical (maximum) heat flux and on obtaining the boiling curves which
describe the dependence of the heat flux on the substrate temperature. An
4
Figure 1.3: Exemplary evolution of the heat flux as a function of the surface
temperature during spray cooling.
example for empirical correlations can be found in Mudawar & Deiters (1994).
Among the governing parameters are spray properties, like droplet velocity,
droplet diameter, mass flux, or liquid properties (Puschmann & Specht, 2004;
Wendelstorf et al., 2008; Yang et al., 1996; Chen et al., 2002; Estes & Mudawar,
1995). Moreover, experiments of Cebo-Rudnicka et al. (2016) with different
target materials demonstrated that the heat transfer is influenced also by the
thermal conductivity of the surface.
Most of the models for the heat flux and for the critical heat flux are
completely empirical. The main goal of the present study is to develop a
predictive theoretical model for fast transient cooling of a very hot thick
substrate by spray impact. The model should be based on the identification
of the main influencing physical parameters. These influencing parameters
are different for the film boiling regime and for the nucleate boiling regime.
The following subsections address some selected issues and influencing
parameters during spray cooling which are covered to a lesser extent in the
review literature cited above.
The Leidenfrost point The following section gives an overview of the

state of the art regarding the understanding of the Leidenfrost point and its
influencing parameters.
5
1 Introduction
The physics of the transition from the nucleate boiling regime to the film
boiling regime at the Leidenfrost point is not yet completely known in that
the Leidenfrost temperature cannot be reliably predicted. Several theoretical
models have been developed based on the hydrodynamic stability analysis of
the vapor/liquid interface Jerome (1960); Zuber (1958); Kakac & Bon (2008)
or thermocapillary stability Aursand et al. (2018). Some authors assume that
the Leidenfrost temperature is determined by the foam limit Spiegler et al.
(1963); Wang et al. (2019b), which is the maximum temperature to which
a liquid can be superheated, or by the limiting minimum vapor thickness
Cai et al. (2020) when it becomes comparable with the surface roughness.
However, the influence of the surface roughness is not yet clearly delineated
and requires further investigations.
During single drop impacts: Numerous studies deal with the Leidenfrost
point during single drop impact (Quéré, 2013; Biance et al., 2003; Castanet
et al., 2015; Tran et al., 2012), showing a difference between the static
Leidenfrost temperature TLs of a sessile droplet and the dynamic Leidenfrost
temperature TLd of an impacting droplet. In these studies the dynamic
Leidenfrost point is often observed as the transition between a wet and dry
rebound. A review on the static Leidenfrost temperature can be found in
Bernardin & Mudawar (1999).
Although, there is an ongoing discussion about the influencing parameters
on this point (Liang & Mudawar, 2017a), there exists some agreement that
the dynamic Leidenfrost point depends on the Weber number. Therefore,
common correlations for the dynamic Leidenfrost temperature in the form of
TLd = TLs + C1 WeC2 (1.1)
can be found (Bertola, 2015; Yao & Cai, 1988). Here the dynamic Leidenfrost
temperature is linked to the static one by an influence of the Weber number.
Although Eq. (1.1) is used in many studies, the values found for the empirical
constants C1 and C2 span a large range, depending on which study is quoted.
The studies have in common that C2 < 0.5. This leads to a somewhat diffuse
picture of the physics involved in the Leidenfrost point. Nevertheless, since
C2 < 0.5, Eq. (1.1) indicates that for a sufficiently large Weber number the
6
second term in Eq. (1.1) will converge to a constant. Therefore the influence
of the Weber number becomes small and the Leidenfrost temperature is nearly
constant for sufficiently large Weber numbers.
Wang et al. (2019b) performed experiments with various substrate materials
for single drop impacts having a Weber number in the range We = 30 − 120.
The results indicate no influence of the Weber number. Instead different
substrate materials lead to different Leidenfrost temperatures. The authors
conclude that a higher thermal effusivity leads to a higher Leidenfrost tem-
perature. They further compared their experimental data to a Leidenfrost
point model proposed by Bjornard & Griffith (1977) which is based on the
homogeneous nucleation theory. They find a good agreement between their
experimental data and the theory.
In sprays: The studies of Sozbir et al. (2010) and Sozbir et al. (2003) indi-
cate an higher Leidenfrost temperature for higher mass fluxes. Furthermore,
their study shows an increasing Leidenfrost temperature for increased air
velocity of a pneumatic atomizer. The same influence of the mass flux is
reported in Hoogendoorn & den Hond (1974). Here the range of Leidenfrost
temperatures is 350 to 900 ◦ C. The authors also conclude that outcomes
concerning the Leidenfrost point of single droplets cannot be transferred to
the Leidenfrost point in sprays. The same influence of rising Leidenfrost
temperature with rising mass flux is confirmed by Gottfried et al. (1966);
Al-Ahmadi & Yao (2008). Al-Ahmadi & Yao (2008) report a insignificant
influence of the droplet diameter and velocity. They correlated their data for
ṁ = 1.5 − 30 kg/m2 s with the expression
TL = 536.8 ṁ0.116 . (1.2)
Here TL is the Leidenfrost temperature and ṁ the mass flux. Additionally,

the authors show that the Leidenfrost temperature depends on the substrate
material. Their experiments performed with stainless steel have much higher
Leidenfrost temperature than those with copper or aluminum.
Based on a dimensional analysis Yao & Cox (2002) derived a modified
Reynolds and Weber number using spray relevant parameters to represent
the inertia of the entire spray instead of that of a single droplet. These spray
7
1 Introduction
Figure 1.4: Dependence of the Leidenfrost temperature on the spray Weber number
and application of correlation Eq. (1.5). (Adapted from Yao & Cox (2002), with
permission of Taylor & Francis. © 2002 Taylor & Francis.)
Reynolds and spray Weber numbers are defined as
ṁD
ReS = and (1.3)
η
ṁ2 D
WeS = , (1.4)
ρσ
where D, η, ρ and σ are a mean diameter, dynamic viscosity, density and

surface tension. The authors identify a positive correlation between spray
Weber number WeS and Leidenfrost temperature. The same effect is reported
by Labergue et al. (2015). Yao & Cox (2002) used their own experimental
data in combination with results from various other studies and derived the
following correlation:
TL = 1400We0.13
S (1.5)
The correlation as well as the underlying data is shown in Fig. 1.4.

Bernardin & Mudawar (2004) developed a model for the Leidenfrost tem-
perature of impacting droplets and sprays which accounts for the droplet
velocity U ,
TL = 162 + 24.3U 0.64 . (1.6)
8
The model is applied to experimental data of Klinzing et al. (1992) and

Bernardin & Mudawar (1996) and indicates only an influence of the droplet
velocity.
In summary, the literature dealing with the Leidenfrost point for sprays is
rather sparse. For single drop impacts there is some agreement on an influence
of the Weber number on the Leidenfrost temperature. Although, there is some
agreement that the Leidenfrost point during spray impact mainly depends
on the mass flux, the overall picture is not clear. Concerning the Leidenfrost
point in sprays, there currently exist no physical models or explanations, but
only a few empirical correlations, which themselves exhibit large scatter.
Influence of the initial spray water temperature Spray cooling in-

volves single drop impact. Depending on the regime and the number flux
of spray droplets, the heat flux can even be described by a superposition of
single drop impacts (Breitenbach et al., 2017b). Therefore it is worth starting
the review of the existing literature with studies dealing with single drops.
In Hiroyasu et al. (1974) the influence of the initial drop temperature
on the drop lifetime of a benzene drop is investigated for different surface
temperatures. The lifetime becomes smaller at all surface temperatures for
the warmer drop.
The findings in Bernardin & Mudawar (1999) indicate no effect of the
liquid subcooling on the Leidenfrost temperature of a sessile drop. This is
explained by the very fast heating to saturation temperature of the small
amount of liquid contained in the drop. This outcome is in accordance with
other studies.
In Xu & Gadala (2006) the influence of the water temperature during jet
cooling is investigated. Hot water results in a lower heat flux at all surface
temperatures and also in a lower Leidenfrost temperature. The film boiling
regime is only present in the case of low water temperatures. The authors
conclude that the film boiling regime only exists for high water temperatures.
The outcomes of Nimi et al. (2012) show a strong shift of the Leidenfrost
point during spray cooling towards lower temperatures for higher initial water
temperatures.
9
1 Introduction
We can summarize: First, the fluid temperature strongly influences the heat
flux during spray and jet cooling. Second, it has an impact on the Leidenfrost
point for spray and jet cooling. Third, there is no influence on the Leidenfrost
point of a single drop.
1.3 Objective and outline of this thesis
Spray cooling is used in many different industrial applications because of

its high cooling capability. In recent years various applications of spray
cooling demand an even higher and more predictive cooling performance. For
example: In the forging industry, a higher cooling performance leads to a
shorter cooling time, shorter cycle times and therefore higher productivity.
High power electronics could be more powerful when equipped with better
cooling devices. These trends require a deeper understanding of the physics
of spray cooling and reliable models to design and predict the cooling process.
To date, the overall physical picture is not complete and existing predictive
tools are highly empirical and not universal.
The goal of the present thesis is to improve the knowledge of spray cooling
and develop models predicting the heat flux. For this purpose measurements
of heat transfer and its dependence on various parameters like spray or wall
properties are performed. The results are linked with visualizations of the
hydrodynamics during spray impact. This leads to a new understanding of
certain regimes of spray cooling and to new predictive models.
Chapter 2 describes the experimental setup and measurement techniques
used for the experimental investigations.
Chapter 3 starts with an illustration of the typical outcomes of transient
spray cooling. First the heat flux and surface temperature are investigated
and connected to observations of the spray impact. This is followed by
remarks about the influence of various parameters, like spray impact or wall
parameters.
In Chapter 4 the heat transfer in the film boiling regime is analyzed. A
predictive model for the heat transfer, which accounts for various spray and
wall properties, is developed and validated.
10
1.3 Objective and outline of this thesis
In Chapter 5 a predictive model for the heat transfer in the nucleate

boiling regime is developed. The model acts as an upper boundary of the
heat flux and is validated against various influencing parameters.
Chapter 6 deals with the Leidenfrost point and its influencing parameters.
The experimental data is analyzed and compared to existing data from
literature. The main influence is characterized and correlated to predict what
conditions determine the Leidenfrost point.
Chapter 7 concludes this thesis and offers an outlook for topics to be
addressed in future studies.
11
12
2 Experimental setup and methods
This chapter describes the different experimental setups and measurement
techniques used for the acquisition of the experimental data in this study,
with some exemplary results shown for illustration.
Parts of this chapter have already been published in Tenzer et al. (2018,
2019a,b). In addition Wolter (2018) supported the construction of the experi-
mental setup and performed first measurements.
2.1 Overview of the setup

In this study two different setups are used, as shown in Fig. 2.1. The test
rig on the left side is used for performing the cooling experiments and is
therefore called the cooling setup. It consists of the heated target equipped
with thermocouples connected to a data acquisition system. This forms
the measurement system for the acquisition of the surface temperature and
the heat flux at the surface where the spray impact takes place. A high
speed camera equipped with a long distance microscope or other lens in
combination with a back light illumination is used for visual observation of the
hydrodynamics during spray impact. The spray is produced by conventional
atomizers driven by a water supply system.
The test rig on the right side is called spraying setup and features all
instrumentation for a precise characterization of the spray. The atomizer
and water supply system is exactly the same as in the previously described
cooling setup. A phase Doppler (PD) measurement system consisting of a PD
receiver, PD transmitter and data acquisition is used to measure the droplet
diameters and velocities. A custom built patternator measures the local mass
flux.
During post processing, data from both test rigs are combined to correlate
the local heat transfer, accompanied by the visual observation of the hydro-
dynamics during spray impact, with the local spray properties. By changing
13
Figure 2.1: Schematic of experimental setup. Left: Heat flux measurements with
thermocouples and visualization with HS-camera. Right: Spray characterization
with phase Doppler measurement system and patternator. (Adapted from Tenzer
et al. (2019b), with permission of Cambridge University Press. © 2019 Cambridge
University Press.)
the type of atomizer, the supply pressure of the atomizer and the distance
between the atomizer and the heated target, different kinds of sprays can
be produced. Geometric constraints limit the distance between the atomizer
and target to 50 − 350 mm. The entire setup is completely automated and
controlled by a PC running Labview, which is also used for acquiring and
recording the data.
A more detailed description of this equipment follows in the subsequent
sections of this chapter.
2.2 Spraying test rig

A detailed sketch of all components involved in the process of spray generation
is shown in Fig. 2.2. Purified water from a reverse osmosis device is stored in
a tank (1). In cases of experiments with elevated spray water temperature,
the water is heated by an immersion heater (2) connected to a temperature
controller (3). In other cases the water remains at ambient temperature.
A gear pump (4) is used to supply the atomizer. A pressure (5) and a
temperature sensor (6) acquire the corresponding quantities of the water. The
14
Figure 2.2: Detailed sketch of the atomizer and water supply system used for spray
generation.
atomizer (9) is separated from the supply line by a check valve (8). The supply
line is connected to the recirculation line through a directional valve (7). A
movable shutter (10) is located directly below the orifice of the atomizer. This
device is able to collect all spray that exits the atomizer. It is quickly pushed
into or out of the spray stream by a pneumatic cylinder. The shutter is opened
only when the spray has reached its steady state. This avoids larger fluid
ligaments, which are typical for the unsteady starting phase of an atomizer,
from reaching the heated surface or patternator.
The setup allows the water to circulate through the entire system when the
directional valve is in open position. The check valve remains closed since the
pressure upstream is lower than its opening pressure and therefore no water
leaves the atomizer. During this circulation phase the circulating water heats
all components of the system to the desired water temperature to ensure a
constant spray water temperature during spraying. When the directional
valve is closed the pressure upstream of the check valve rises until finally the
check valve opens and the water leaves the atomizer as a spray. The check
valve creates a pressure loss of 0.2 bar which is registered during operation.
However, since the same spraying system is used for all experiments the
pressure drop does not have to be considered further. The desired atomizer
operation pressure is kept constant using a controller, which controls the
speed of the gear pump based on the pressure sensor data.
15
Table 2.1: Overview of atomizers used for the experiments and their operational
parameters.
Atomizer Bore diame- Spray Operation Mass flow
ter, mm angle pressure, bar rate, kg/h
Lechler 490.403 1.25 45◦ 1.5 − 10 51 − 135
Lechler 490.603 2 45◦ 2−5 186 − 266
Spraying Systems 0.79 30◦ 1.5 − 10 46 − 102
3002.5
Spraying Systems 0.79 30◦ 2 − 10 28 − 64
3001.4
The atomizers are mainly one component, pressure swirl, full-cone type,
which are available from various companies. For this work four different atom-
izers are chosen, with the aim of obtaining a broad span of spray parameters.
The designation and typical operational parameters of the four atomizers are
summarized in Table 2.1. The bore diameter and spray angle are taken from
the manufacturer’s data sheet. The operation pressure and the mass flow
rates were acquired during the phase Doppler measurements. For this purpose
a Coriolis mass flow meter (Optimass 7400 C from Krohne) was installed
upstream of the atomizer. Due to the limited capacity of the gear pump, the
pressure range for the largest atomizer is limited.
Preliminary experiments were performed using a pneumatic atomizer, which
uses a secondary air stream to enhance the atomization process. The desig-
nation of that atomizer is Lechler 136.115. Generally a pneumatic atomizer
produces a spray having smaller and faster droplets and a lower mass flux.
Since only a few experiments were performed with this atomizer, a detailed
description of its spray properties have been omitted below.
2.2.1 Phase Doppler technique

The local drop diameter and two-component drop velocity vector is acquired
using a phase Doppler measurement system operated in the dual-mode configu-
ration. The system is built by DantecDynamics and consists of a transmitting
and a receiving optic and a processing unit. A detailed description of this
measurement system is given in Albrecht et al. (2013); Tropea (2011); Tropea
16
Table 2.2: Parameters of the phase Doppler measurement system.

Parameter Setting
Laser type DPSS laser
Laser power 18 − 300 mW
Wavelength 1 532 mm
Wavelength 2 561 mm
Beam spacing 60 mm
Transmitter focal length 500 mm
Receiving system 112 mm Dual PDA
Receiver focal length 500 mm
Scattering angle 28◦
Refractive index water 1.33
et al. (1996). The relevant system parameters are summarized in Table 2.2.
Laser power and other operational parameters were adjusted according to
Araneo & Tropea (2000) to ensure the best measurement results and validation
rates. The interchangeable receiving aperture mask was selected depending on
the droplet size range of each atomizer and operating pressure. More details
about the setup and experimental procedure can be found in Wolter (2018).
2.2.2 Patternator
A measurement device for the acquisition of the local mass flux is called a
patternator (Lefebvre, 1988). During this work two different custom made
mechanical patternators were used. Both devices use tubes to collect the
amount of spray water m that hits a part of an surface area A in a given time
t. Since the collection area A is small compared to the overall sprayed surface,
the mass flux ṁ used in this work is a local quantity. It can be calculated by
m
ṁ = . (2.1)
At
The difference of the two patternators employed in this study lies in the
method with which the amount of collected water was measured. The first
device measures the mass of the collected liquid by weighing. The water is
gathered by 17 small tubes of inner diameter of 4 mm, placed in a row with a
17
distance of 6 mm from one another. These tubes feed the water to containers
which are then weighed. A more detailed description can be found in Wolter
(2018).
Due to the necessity of weighing each container, these measurements take
a long time. To overcome this drawback an automated patternator was
developed. The collection surface was increased to 6 × 6 mm2 which is still
small enough to capture any spray inhomogeneities. The weighing is replaced
by a capacitive measurement of the fluid level in each of the cells. Details
of the working principle and the construction can be found in Kaupe (2017)
and Hofmann (2019). With the addition of an automated drain system the
measurement time is greatly reduced.
The measurement axis of the patternator was aligned with the position at
which the temperature measurements were performed, ensuring that both
quantities are acquired at the same position.
2.2.3 Results of experimental spray characterization

In the following section only exemplary results are discussed in detail, simply
to show the typical behavior of the spray. Since the remaining results exhibit
the same trends, they are only summarized in an overview.
Exemplary results of droplet diameter and velocity Phase Doppler

data is acquired without the target on horizontal planes at distances to the
nozzle from 25 to 300 mm. Further decreasing the distance is not possible
due to a strong decrease of the validation rate, since the breakup process is
not completed and many droplets are non-spherical, prohibiting measurement
using the phase Doppler technique. Above 300 mm there is no significant
change of the spray properties.
Figure 2.3a shows the mean droplet diameter D10 at 100 mm distance from
the nozzle and at certain positions along the radial axis, where 0 corresponds
to the geometrical center of the nozzle. The experiments were performed using
the Lechler 490.403 atomizer. As expected, the droplet diameter reaches
a minimum in the center of the nozzle and increases slightly towards the
periphery. With increasing pressure the mean droplet diameter decreases,
which is also expected.
18
(a) Mean droplet diameter D10 . (b) Mean droplet velocity U .
Figure 2.3: Mean droplet diameter and velocity as a function of radial position.
Position r = 0 mm corresponds to the geometrical center below the orifice.
(Adapted from Tenzer et al. (2019b), with permission of Cambridge University
Press. © 2019 Cambridge University Press.)
In Fig. 2.3b the mean droplet velocity in the main flow direction U is shown
at the same positions and distances as considered before. In the center the
velocity attains a maximum and decreases at the outer positions. An increased
pressure results in an increased droplet velocity.
Influence of the target as a displacement body. Since the phase

Doppler measurements were performed in absence of the target, the question
arises whether these free stream results are comparable to those in presence of
the heated target. For that reason additional measurements were performed
in presence of a bluff body having the same dimensions as the heated target.
The measurement plane was located 10 mm above the surface of the target.
During data processing we included only those droplets having a mean velocity
in the downward direction, to exclude those droplets that had already im-
pacted at the surface and rebounded. In an exemplary case the integral spray
characteristics are for free stream conditions: D10 = 59 µm and U = 9.6 m/s.
Due to the presence of the bluff body the values changed to: D10 = 81 µm and
U = 9.7 m/s. Especially the small droplets are affected by the displacement
of the target, resulting in the higher D10 , since only a smaller portion of them
reach the surface. However, because the largest influence on the heat flux
19
Figure 2.4: Mass flux as a function of the radial position. Position r = 0 mm

corresponds to the geometrical center below the orifice. (Adapted from Tenzer
University Press.)
is the mass flux and the previously mentioned differences are rather small,
we conclude that it is reasonable to use the droplet diameter and velocity
acquired under free stream conditions as input data for the present work.
Details of the experiments and results regarding the influence of the target
can be found in Wolter (2018). Perhaps even more significant is however the
fact that the measurement of the mass flux using the patternator does mimic
exactly the conditions prevailing with the heated target.
Exemplary results of mass flux In Fig. 2.4 the mass flux at different
radial positions is shown for different operating pressures of the atomizer
at 100 mm distance to the nozzle. The measurements are performed with
the patternator. The plot for 2 bar demonstrates the expected trend of a
full-cone atomizer: the highest mass flux appears in the center with a strong
decrease towards the outside. This behaviour changes with increasing pressure.
Normally one would expect the shape to remain nearly constant with a shift
towards higher mass fluxes. From the measurements we see instead a decrease
of the mass flux in the center and an increase at the outer region. The
entire spray pattern changes with rising pressure and the atomizer appears to
behave more like a hollow cone atomizer. This means that the main spray is
concentrated in a ring of the radius 40 to 50 mm. This behaviour is typical of
a pressure swirl atomizer.
20
Figure 2.5: Droplet velocity versus diameter for various atomizer in the center
region below the atomizer. Operation pressure and distance also varied but not
indicated in the figure.
It is interesting that the mass flux at the center depends only weakly on
the injection pressure. For the nozzles used in the experiments the pressure
influences mainly the mass flux distribution in the outer ring of the spray
cross-section.
Overview of the spray parameter range By changing the type of at-

omizer, operating pressure and distance between the atomizer and target,
different kinds of sprays can be produced. The resulting parameter range,
which can be achieved by the previously mentioned operation conditions, is
shown in Fig. 2.5. Here the mean droplet velocity is plotted versus mean
droplet diameter for the central area below the atomizer. The parameters span
the range D10 = 38 − 114 µm and U = 5 − 26 m/s. Within these boundaries
the parameters scatter highly.
The local mass flux spans the range ṁ = 0.5 − 29.5 kg/m2 s. An isolated
trend for the dependence of the local mass flux on operation conditions cannot
be identified, as shown in Fig. 2.4. Nevertheless, decreasing the distance
between the atomizer and target obviously increases the mass flux.
21
Since the outcomes of isothermal drop wall interaction are governed by the
droplet velocity and diameter, it is reasonable to introduce the dimensionless
parameters, Reynolds and Weber number, as follows:
U D10
Re = (2.2)
νf
ρf D10 U 2
We = (2.3)
σf
Here νf is the kinematic viscosity, ρf the density and σf the surface tension of
the liquid. The range of parameters used for this experimental study can also
be described as Re = 403 − 1732 and We = 26 − 549.
For the sake of completeness, it is mentioned that the spray parameters of the
pneumatic atomizer span the ranges: ṁ = 0.2 − 3.3 kg/m2 s, D10 = 12 − 24 µm
and U = 12.7 − 23.4 m/s or Re = 200 − 434 and We = 32 − 125.
These ranges of spray parameters are also used for the cooling experiments.
2.3 Cooling test rig

In Fig. 2.6 the heated target and its peripheral equipment are shown as it is
used during the cooling experiments. The target (inner surface) is installed
in a watertight housing which also contains insulation material. An outer
hollow cylinder surrounds the housing and is connected at the bottom to a
ventilation system (not shown here). Thus, the slit between outer cylinder and
housing acts as a suction channel which extracts the overspray and secondary
droplets. The extraction system is necessary to ensure optical access to the
heated surface and to protect the surrounding equipment from contamination
with water.
The heated target The heated surface of the spray impact target is the top
end of a circular cylinder (diameter dT = 100 mm and height hT = 53.2 mm).
A detailed drawing is found in Fig. 2.8. During the course of this study
different targets having the same geometry were used. They were built from
various materials and have different surface conditions. Table 2.3 summarizes
the different targets and their material properties. Here ρ, cp and λ are the
density, specific heat capacity and thermal conductivity respectively. Each
22
Figure 2.6: Heated target equipped with watertight housing and ventilation slit.
Table 2.3: Overview of targets and material parameters used of cooling experiments.
Name Material Surface ρ, kg/m3 cp , J/kgK λ, W/mK
condi-
tion
Stainless steel 1.4841 smooth 7900 542 16
Stainless steel 1.4841 rough 7900 542 16
Hot-work 1.2365 smooth 7800 510 33
tool steel
Nickel 2.4068 smooth 8900 500 63
value corresponds to the calculated mean between 100 and 500 ◦ C. The smooth
surface condition corresponds to a mirror polished surface made by lapping
and polishing. The average roughness of the these surfaces is < 0.03 µm. The
rough surface was produced by sandblasting, resulting in an average roughness
of ≈ 10 µm.
Both stainless steel and nickel exhibit good resistance against corrosion and
oxidation. The majority of experiments were performed with stainless steel.
This target endured more than 300 cooling experiments. Since the surface
suffered from thermo-shock, changes of the surface, like less brightness, are
visible. Despite this, the repeatability between the first and last experiments
is good.
23
In contrast, the target built of hot-work tool steel exhibits much more
corrosion. Therefore only a few exemplary experiments were performed to fill
the gap of thermal properties between stainless steel and nickel.
2.3.1 Mechanical and thermodynamic construction

The main challenges during the construction of the heated target are achieving
good thermal insulation and maintaining a watertight housing to ensure that a
complete flooding of the target is possible without any damage to components.
These difficulties arise from the following two problems:
• Typically insulation materials are porous and therefore not suitable for
contact with water.
• The dimensions of the heated target enlarge approximately 1 mm be-

tween ambient temperature and operation temperature due to thermal
expansion. Any typical sealing material having the flexibility to com-
pensate this enlargement does not withstand the high temperatures of
about 500 ◦ C. On the other hand, a rigid connection between the parts
would break due to the force of the thermal expansion.
The above mentioned difficulties lead to a rather complicated construction,

as shown in Fig. 2.7:
A split copper disk (2) is screwed to the bottom of the target (1) by using
four threaded rods (4). For protection of the soft copper, a pressing plate (5)
built of stainless steel (1.4841) is used to equally distribute the connection
force. Four cartridge heaters (3) (500 W each) are installed inside the split
copper disk.
The top cover of the housing is a rigid plate (6). It is built of fiber cement
Kelutherm 700 and has a low thermal conductivity of 0.38 W/mK. This
plate is covered by a 0.3 mm thin stainless steel sheet (7), to protect the fiber
cement from water. At the outside the fiber cement plate and stainless steel
sheet are screwed to the outer cylinder (8) which is built of aluminum. An
aluminum ring (9) is installed to equally distribute the pressure of the screws.
Since there the temperature is low, conventional silicone sealing plates are
used to watertight the interfaces between the different components.
24
Figure 2.7: Heated target and additional components inside the watertight housing.
25
The target (1) rests with its collar on the top cover. A metallic seal (10) is
installed between the target collar and stainless steel cover to watertight this
interface. The seal is a metallic high temperature duct seal type “E-Ring”. It
has the benefit of not requiring high sealing forces. These seals are usually
custom built parts, except some stock models used in the aviation industry.
For simplicity this setup is built using such a stock seal (AS1895/7-450 ),
which can be purchased from aviation aftermarket suppliers.
High temperature disk springs (11) built by Vinsco Spring Limited are
installed at the bottom of the heated target to ensure a nearly constant sealing
force, independent of the thermal expansion. Theses disk springs act between
the threaded rods (4) and a compression structure. The structure consists of
a plate (12) built of Kelutherm 800 M and two hollow cylinders (13) built of
Kelutherm 700. A spacer (14) limits the compression of the seal to its range of
operation. Due to this construction, the insulation material is only subjected
to pressure, and the heated parts can freely expand and contract.
The free space inside the cylinder (8) is filled with Insulfrax S insulating
blanket for insulation propose. Additionally, the inside is filled with Argon
gas to protect all materials and especially the copper disk from oxidation.
Due to the seals, the housing is nearly gas tight and the pressure inside can
be raised to ≈ 0.2 bar above atmosphere. Therefore, failure of any seal or part
can be detected by an increased volume flow of the Argon gas. Thus, any
damage of the watertight housing is detected before water enters the inside
and leads to further destruction.
A temperature controller based on Labview is implemented in the measuring
PC and controls the desired temperature inside the copper disk by taking
temperature readings from a thermocouple inside the copper disk as an input
value and by adjusting the heating power.
2.3.2 Measurement technique for heat flux and surface

temperature
The aim of this work is to achieve a better understanding of transient spray
cooling. Since transient cooling requires a thick target as thermal mass which
can be continuously cooled, the possible measurement techniques for heat
flux are rather limited. For example, infrared imaging from the bottom is not
26
Figure 2.8: Sketch of the sectional view of the heated target showing the thermo-
couple positions. Dimensions are in mm.
possible due to the presence of the non-transparent target. Infrared imaging

from the top suffers from poor optical accessibility caused by the dense spray
and an undefined emission coefficient due to changing surface conditions.
The measurement technique used in this work is therefore based on solving
the inverse heat conduction problem (IHCP) using temperature readings
from inside of the target as input data. By solving the IHCP the boundary
conditions can be calculated from data taken inside of the domain. In our
case the boundary condition corresponds to the unknown heat flux and the
data from inside of the domain to temperature readings from the inside of the
target. The temperature is acquired by thermocouples embedded inside the
target. In Fig. 2.8 the position of the thermocouples inside the target is shown.
They are placed in two rows. The first is located 0.5 mm below the surface
to achieve a quick response time. The second is 20 mm below the surface
to make the IHCP independent of the boundary condition at the bottom.
The radial distance between each sensor in the first row is 3.5 mm to account
27
Figure 2.9: Sketch of the thermocouples having an open measuring tip which is
aligned with the shield.
for any radial distribution of the heat flux. The problem is assumed to be
two-dimensional, axisymmetric and having adiabatic boundary conditions at
the curved surface area.
The IHCP is solved using a procedure published by Monde et al. (2003).
Furthermore, Monde and coworkers at Saga University developed and made
available an inverse heat conduction analysis tool “Invers2D”. In the present
work, this tool is used for calculating the heat flux and surface temperature
from temperature readings of the thermocouples.
The procedure is described in detail in Woodfield et al. (2006). It solves
the problem in Laplace space. Therefore the two-dimensional heat conduction
equation is transformed into Laplace space. The measured temperature inside
the target is approximated in time as a series of half-power polynomials and in
space direction as Fourier-Bessel series. After solving the problem in Laplace
space, an inverse Laplace transform leads to the solution.
The thermocouples are type K, class 1, with 0.5 mm shield diameter. The
measuring tip is open and aligned with the shield, as shown in Fig. 2.9. This
results in a fast rise time.
The holes, inside which the thermocouples are placed, are produced using
the spark erosion technique. The resulting hole diameter is 0.6 mm. The
sensors are bonded inside the holes using a thermally high conductive adhesive
(Aremco Ceramabond 569 VFG) to ensure good thermal contact.
The temperatures are sampled at a sample rate of 95 Hz using National
Instruments NI 9212 thermocouple input modules attached to a National
Instruments cRio 9074.
28
Figure 2.10: Field of view for configuration with Tamron 180 mm macro optical
lens.
Figure 2.11: Field of view for configuration with Questar QM-100 long distance
microscope objective lens.
2.3.3 Observation system
Since there are only qualitative and no quantitative outcomes based on the
visualization, a long description of the visualization system is omitted and
only a brief summary of the hardware is given.
The observation system consists of a high-speed camera equipped with two
different lenses and a back light illumination source. The camera is either a
Phantom v12.1 or Phantom v2012. Depending on the desired field of view
a Tamron 180 mm macro optical lens (far field) or a Questar QM-100 long
distance microscope (near field) is used. An exemplary image of the surface
and scale for both configurations is shown in Figs. 2.10 and 2.11. In addition
to the different magnification the depth of view is much smaller in the case of
the long distance microscope. This helps to mask the droplets that are outside
of the focal plane and therefore prevent these droplets from obstructing the
optical accessibility.
29
The back light illumination consists of a light source and a diffusor plate.
In the case of the objective lens, a high power LED is used as light source.
For the configuration with the long distance microscope an infrared Cavilux
HF laser is necessary to achieve short illumination times to prevent motion
blur while maintaining high light intensity.
The visualization and heat flux measurements are temporally matched and
therefore visual observations can be directly associated with the instantaneous
local heat flux and target surface temperature.
30
3 Typical results of transient spray
cooling
This chapter describes the experimental results obtained in this study. First
some general outcomes of the heat transfer during spray cooling are shown
and afterwards these are connected to the visualization of the hydrodynamics
involved in spray impact. The subsequent parts show the influence of various
parameters on the heat transfer.
Unless otherwise stated, the underlying experiments of the following results
were performed using the stainless steel target having a smooth surface.
Parts of this chapter have already been published in Tenzer et al. (2018,
2019a,b). Furthermore Pham (2018) and Hofmann (2019) supported the
experimental investigation and performed some of the measurements.
3.1 Measurements of the heat flux

Contrary to many experimental studies of spray cooling in which the substrate
temperature is kept constant, in the present study transient heat transfer is
investigated, since this situation is relevant to many industrial applications
mentioned in Section 1.1. Therefore, the target is initially heated until it
achieves a given uniform initial temperature. The spray parameters are kept
constant during the entire experiment. The influence of any start-up phase
during the initial development of the spray is avoided by using a shutter in
front of the nozzle. The heat transfer measurements are started only after
the spray is fully developed, the target has been uniformly heated and the
shutter has been opened. At time t = 0 the heating of the target is switched
off, simultaneous with the opening of the spray shutter. At this instant the
substrate temperature starts to change due to the heat flux associated with
spray impact.
31
3 Typical results of transient spray cooling
Figure 3.1: Calculated temperature T inside the substrate as a function of the

depth z for different times t. The dashed lines indicate the position of the
thermocouples at z1 = 0.5 mm and z2 = 20 mm. (Adapted from Tenzer et al.
(2019b), with permission of Cambridge University Press. © 2019 Cambridge
University Press.)
The typical evolution of the temperature field inside the substrate, the
surface temperature and the heat flux are illustrated in Figs. 3.1 to 3.4. The
initial wall temperature is Tw0 = 450 ◦ C. The spray parameters for this case
are: ṁ = 2.9 kg/m2 s, D10 = 55 µm and U = 10.3 m/s.
The evolution of the temperature profiles T (z) is calculated by solving the
inverse heat conduction problem and is shown in Fig. 3.1 for different times
t. The z coordinate coincides with the spray axis while the position z = 0
corresponds to the wall surface, where the spray impact takes place. The
bottom of the heated plate corresponds to z = 53.2 mm.
The vertical dashed lines in Fig. 3.1 indicate the position of the thermocou-
ples at z1 = 0.5 mm and z2 = 20 mm. The temperature measurements of the
thermocouples are used as the input data for the solution of the inverse heat
conduction problem.
The corresponding time series of the surface temperature and heat flux are
plotted in Fig. 3.2. In Fig. 3.3 the heat flux is plotted as a function of the
surface temperature, which shows good agreement to the schematic diagram
in Fig. 1.3. The vertical dashed lines in Figs. 3.2 and 3.3 correspond to the
boundaries of the boiling regimes: film boiling, transition boiling and nucleate
boiling regime, which are described in more detail in Figs. 3.4 and 3.5.
32
3.1 Measurements of the heat flux
Figure 3.2: Typical temporal evolution of the surface temperature Ti (t) and in-
stantaneous local heat flux q̇(t). The vertical dashed lines indicate the boundaries
between film boiling, transition and nucleate boiling. (Adapted from Tenzer
University Press.)
Figure 3.3: Typical evolution of the instantaneous local heat flux q̇(t) as a function
of the surface temperature Ti (t). The vertical dashed lines indicate the boundaries
between film boiling, transition and nucleate boiling.
33
Figure 3.4: Spray cooling regimes at different surface temperatures Ti (t) around the
Leidenfrost point. a) Measured heat flux q̇ as a function of surface temperature Ti ;
b) visualized spray impact in the film boiling regime; c) at the Leidenfrost point,
characterized by the first appearance of liquid patches; d) the fast expansion
of the liquid spots. (Adapted from Tenzer et al. (2019b), with permission of
Cambridge University Press. © 2019 Cambridge University Press.)
3.2 Visualization of the spray impact

Near field observation In Fig. 3.4a the instantaneous local heat flux q̇(t)
in the central area of the target and spray is plotted as a function of the
surface temperature Ti (t) for the same experimental data as in Figs. 3.1 to 3.3,
whereby time t increases following the curve to the left.
The precision of the thermal measurements at the initial stage of cooling is
not high, because of the very high temperature gradients that occur in the first
instances of cooling. The rise time of the thermocouples is not short enough
and the thermal inertia of the material between the tip of the thermocouple
and the surface is too high to precisely capture these fast temperature changes.
This part of the plot, where the measurement precision is not well quantified,
is indicated here and in the following figures by a dotted curve.
34
In Fig. 3.4b, 3.4c and 3.4d images of spray impact and hydrodynamic
phenomena at the surface captured at different instants after the spray cooling
begins are shown. The time instants chosen for these images are marked on
the graph in Fig. 3.4a.
The hydrodynamic phenomena visualized in Fig. 3.4b, 3.4c and 3.4d are
each different, since they correspond to different drop and spray impact
thermodynamic regimes. In Fig. 3.4b each drop impact onto the wall leads
to its break up, formation of multiple secondary droplets (Roisman et al.,
2018) and rebound. The contact time is short and there is no remaining
wetting of the surface. As a result the heat flux is low, which is typical for
the film boiling regime. At the Leidenfrost point, illustrated in Fig. 3.4c, a
few impacting drops start to wet and spread on the surface - a part of the
surface is covered by initial liquid patches. At the next instant the area of the
wet patches increases, Fig. 3.4d, and the heat flux starts to rapidly increase.
This phenomena correspond to the transition boiling regime.
Similar phenomena in the film and transition regimes are observed in
Fig. 3.5b and 3.5c, respectively. The images are of higher contrast, since
the mass flux of the spray is smaller in the experiment shown in Fig. 3.5.
In Fig. 3.5a the measured heat flux is plotted as a function of the surface
temperature measured during continuous spraying. In the illustrated case the
spray properties are: ṁ = 0.9 kg/m2 s, D10 = 43 µm and U = 9.8 m/s. In this
example the target is initially heated to a surface temperature of 306 ◦ C.
Shortly before the point where the maximum heat flux is achieved a large
portion of the surface is wetted by a liquid water film, as shown in Fig. 3.5c.
Small nucleation bubbles form, grow and collapse. The heat flux increases
significantly, since the wetted area of the substrate increases rapidly. Heat
goes into the overheating of the liquid (sensible heat) and into the formation
of bubbles.
At the instant corresponding to the critical heat flux, Fig. 3.5d Ti = 170 ◦ C,
the surface area is almost completely wetted by liquid. The appearance of a
corona of an impacting drop is clear evidence that the drop impacts onto a
liquid film (Yarin et al., 2017).
35
Figure 3.5: Phenomena of spray impact regimes at different surface temperatures

Ti (t). a) Measured heat flux q̇ as a function of surface temperature Ti ; b) image of
the substrate exposed to spray impact in the film boiling regime; c) inception of
the transition boiling regime; d) fast expansion of the wetted area; e) apparently
completely wetted surface at the instant corresponding to the critical heat flux.
At larger times the substrate is completely covered by a thin boiling liquid

film. The film is continuously fed by fresh water from the spray and increases in
coverage and depth. The heat flux reduces with time. This regime corresponds
to fully developed nucleate boiling of spray cooling.
Far field observation By decreasing the magnification of the optical sys-

tem a larger field of view is achieved. Instead of identifying the individual drop
impacts a more macroscopic view of the hydrodynamics and wetting behavior
at the surface is possible. Figure 3.6 shows the evolution and movement of
the wetting front during spray cooling. The experiment was performed using
the Lechler 490.403 atomizer at 3 bar supply pressure and a distance between
atomizer and target of 100 mm. The appropriate radial distribution of the
mean droplet diameter, mean droplet velocity and mass flux can be found
36
in Figs. 2.3a, 2.3b and 2.4. All data and images are taken from the same
continuous experiment but at different instants in time. The time t, starting at
the moment when the shutter opens, is indicated in the caption of each figure.
The images at the top show pictures of the surface where the spray impact
takes place. The corresponding radial distribution of surface temperature
is indicated in the plot below. The circular markers at the radial positions
r = 3.5, 7, 10.5, ..., mm correspond to the positions of the thermocouples.
Since the radial positions of picture and plot are matched, a direct comparison
between hydrodynamics at the surface and corresponding surface temperature
is possible. The vertical dashed line in the picture at the top indicates the
area at the surface where the hydrodynamic behavior changes. It is detected
from visual observation of the conditions at the surface. The vertical dashed
line in the plot at the bottom is simply an elongation and thus corresponds to
the same radial position r. In the plot at the bottom, the horizontal dashed
line indicates the Leidenfrost temperature, which is determined from the
minimum heat flux in the area around r = 14 mm.
In Fig. 3.6a on the left-hand side of the dashed vertical line a part of
the liquid of the impacting droplets remains at the surface after the drop
has impacted. A persisting wetting of the surface is present and results in a
continuous increase of the amount of water at the surface. This is equivalent to
the previously mentioned nucleate boiling regime. The right-hand side of the
dashed vertical line corresponds to the film boiling regime: Impacting droplets
spread at the surface, atomize and form secondary droplets. However, there
is no remaining wetting at the surface. The surface remains dry. Therefore,
the vertical dashed line corresponds to the area where the Leidenfrost point is
located, which marks the transition between dry and wet droplet impact. The
previously discussed transition boiling regime cannot be explicitly identified
in this representation. Although not clearly identifiable, this regime also takes
place in the area of the dashed line together with the Leidenfrost point.
These observations at the surface correspond to the temperature distribution
in the plot at the bottom. On the right side of the vertical dashed line, in
the film boiling regime, the temperature is nearly constant. It approximately
corresponds to the calculated Leidenfrost temperature (horizontal dashed
line). On the left side of the vertical dashed line the temperature strongly
decreases. Therefore, the Leidenfrost point corresponds to the point at which
the temperature starts to decrease.
37
(a) t = 75.2 s
(b) t = 75.5 s
(c) t = 76 s
38
(d) t = 76.2 s
(e) t = 77.1 s
Figure 3.6: Position of the wetting front that follows the Leidenfrost point at
different instants of time t. Top image shows the hydrodynamics at the surface.
Bottom plot indicates the surface temperature distribution along the radial coor-
dinate r. The vertical dashed line corresponds to the approximate position of the
wetting front. The horizontal dashed line indicates the Leidenfrost temperature.
Figs. 3.6b to 3.6d show the same phenomena. Comparing these figures
with each other, the movement of the wetting front is visible. Starting at
r = 18 mm in Fig. 3.6a, the wetting front is at r = 17 mm in Fig. 3.6b, at
r = 14 mm in Fig. 3.6c and finally at r = 12 mm in Fig. 3.6d. The movement
of the wetting front is a result of the inhomogeneous distribution of the spray
properties. As shown in Section 2.2.3, the sprays in this investigation, which
were produced using high water supply pressure, tend to have a higher mass
39
flux at the outer edges of the spray. Therefore at the outer region (left-hand
part of the surface in the image) the cooling is more intensive due to the
higher mass flux. At the inside (right-hand part of the surface in the image)
the mass flux is lower resulting in a less intensive cooling. This results in a
wetting front that moves from the outside to the inside with increasing time.
A slightly different behavior can be observed in Fig. 3.6e, which is still the
same experiment, but at some instants later. Here a second wetted spot is
visible at the inside of the target (right-hand side of the dashed line), which
is again a result of the inhomogeneous spray distribution. At this instant the
wetting pattern at the surface directly corresponds to the radial distribution
of the mass flux, as shown in Fig. 2.4. In summary, we identify a moving
wetting front that starts at the outer region. It moves towards the inside and
meets at approximately r = 7 mm with a second wetting front that forms in
the center and moves towards the outside.
The previously shown temporal evolution of the wetting front at the surface
is of exemplary nature. Depending on the atomizer, its supply pressure or the
distance between atomizer and surface, different wetting patterns are possible.
For example in the case of a low atomizer supply pressure, the wetting starts
in the center and moves towards the outside, due to the highest mass flux
located in the center.
Figure 3.7 summarizes the different hydrodynamic phenomena that occur
during spray impact at a hot surface: Film boiling features single drop impacts
onto a dry surface. At the Leidenfrost point a small amount of liquid begins
to remain on the surface immediately after the drop impact. A first wetting
occurs. During the transition boiling regime the amount of liquid wetting the
surface increases. Finally, in the nucleate boiling regime a continuous liquid
film is formed, which covers the entire surface.
3.3 Influence of various parameters

In the following section the influence of various parameters on the heat
transfer is investigated. Some influences are discussed only briefly, since
they are addressed again in subsequent chapters. In contrast, others are
discussed in more detail, because they are regarded as a prerequisite for
further investigation.
40
Figure 3.7: Sketch of the hydrodynamics occurring during spray impact. For
clarity, the height of the heated target is much smaller than the one used in the
experiments. (Adapted from Hofmann (2019).)
41
3.3.1 Spray impact parameters

To better understand the influence of the main spray parameters on heat
transfer, measurements with different mass fluxes were performed. Figure 3.8a
shows the heat flux q̇ as a function of time t for the mass fluxes 28.1, 9.3,
2.9 and 1.5 kg/m2 s. The variation of other spray parameters remain in a
relatively narrower range (D10 = 43 − 52 µm and U = 11.2 − 17.4 m/s). The
initial substrate temperature is 450 ◦ C for all the tests. The experiments were
stopped when the first thermocouple reading reached 100 ◦ C. Increasing the
mass flux results in an increased heat flux at all times and boiling regimes.
The corresponding surface temperature Ti as a function of time t is shown in
Fig. 3.8b. Figure 3.8c shows the same data plotted as heat flux dependence
on surface temperature, where the temporal information does not come to
the fore.
The slopes of the curves are similar, especially in the nucleate boiling
regime, and differ mainly at high temperatures between the start of the
cooling experiments and the Leidenfrost point. In this region the heat flux is
much lower for sparse sprays (1.5 and 2.9 kg/m2 s) than for the more dense
sprays (9.3 and 28.1 kg/m2 s). Moreover, for ṁ = 9.3 and 28.1 kg/m2 s no film
boiling regime can be identified, since the time of the film boiling regime is
very short (blue curve in Fig. 3.8a). This can be explained by the limited
response time of the thermocouples, which prevents the detection of the
Leidenfrost point and film boiling regime in the case of this very fast cooling
process. For this reason only those experiments showing the linear trend of
the heat flux in the film boiling regime were considered for further analysis of
the film boiling regime and the Leidenfrost point.
As shown in Section 2.2.3 the spray parameters cannot be varied inde-
pendently from one another. For example, increasing the pressure results in
faster and smaller droplets, but also changes the mass flux. For this reason a
quantitative discussion of the individual dependence of the heat flux on droplet
diameter and velocity is avoided, referring instead to Chapter 4. There, the
modeling accounts for the influence of each individual parameter. However, it
is possible to deduce qualitative trends from Fig. 3.9.
The sprays underlying both experiments in Fig. 3.9 have both the same mass
flux (ṁ = 1.2 kg/m2 s) and the same droplet diameter (D10 = 47 µm). From
the different droplet velocities we can conclude that a higher droplet velocity
42
(a) Heat flux against time. (b) Surface temperature against time.
(c) Heat flux against surface temperature.
Figure 3.8: Influence of mass flux ṁ on the heat transfer. (Adapted from Tenzer
University Press.)
43
(a) Heat flux against surface temperature. (b) Heat flux against time.
Figure 3.9: Influence of the mean droplet velocity on the heat transfer.
(red curve) results in a higher heat flux, compared to the lower velocity (blue
curve). The Leidenfrost point is slightly different, but the overall trend of the
curves is comparable. The corresponding plot showing the dependence of the
heat flux on time is found in Fig. 3.9b.
Due to the previously mentioned interdependence of the spray parameters,
it was not possible to generate sprays with different droplet diameters, but
the same velocity and mass flux. Therefore an explicit presentation of the
individual influence of the droplet diameter is not possible.
Although for illustration propose only one set of experimental data is
shown here, generally the same outcomes are found in the other experiments.
Furthermore, the influence of both single drop quantities, droplet diameter
and velocity, is rather small compared to that of the mass flux.
Pneumatic atomizing nozzle The present study was conducted primarily

using one-component pressure driven atomizers. In contrast, pneumatic
atomizers are often used in industrial applications. Due to the compressed
air, which is used to support the atomization process, the impacting spray
is superimposed on an air stream. To exemplarily capture the influence of
an air-assisted spray, a few experiments using a pneumatic atomizing nozzle
were performed. These results do not represent a complete parameter study,
but provide an initial impression, whether the phenomena are comparable or
which new effects occur. A Lechler 136.115 atomizer is used. For the sake
44
(a) Heat flux versus surface temperature. (b) Heat flux versus time.
Figure 3.10: Influence of different spray produced by a pneumatic atomizer.
of completeness, it is mentioned that the experiments were performed with

the nickel target, whereas the discussion of the influence of different substrate
materials follows in Section 3.3.5.
In Fig. 3.10a the heat flux as a function of surface temperature and in
Fig. 3.10b of time is shown for different spray properties of a pneumatic atomiz-
ing nozzle. Generally the same overall trend can be observed, compared to the
previously described experiments performed with one-component atomizers:
The Leidenfrost temperatures spread across a comparable range. All boiling
regimes are visible. Again a higher mass flux results in a higher heat flux. The
experiments take a very long time of up to 600 s, due to very low mass fluxes
of some experiments. A quantitative comparison between one-component
pressure driven and pneumatic atomizer is not possible because of the large
difference in spray properties. Due to its working principle, pneumatic atom-
izers tend to have smaller and faster droplets than a one-component atomizer.
Therefore it is not possible to produce the same kind of spray with the two
atomizers.
3.3.2 Spray impact angle

In many spray cooling applications the spray does not impact normal to the
surface. Instead there is an inclination between substrate and spray main
direction. Figure 3.11 illustrates such an arrangement.
45
Figure 3.11: Sketch of oblique spray impact.
To capture the influence of an oblique spray impact, experiments with

different spray impact angles β (0◦ , 15◦ , 30◦ and 45◦ ) were performed. The
angle between target and main spray direction is defined in Fig. 3.11. The
exemplary results for one set of spray parameter is shown in Fig. 3.12. The
spray parameters are ṁ = 0.9 kg/m2 s, D10 = 43 µm and U = 10 m/s. Gen-
erally, all curves are very similar and correspond to the typical trend, as
already discussed in previous sections. A larger spray impact angle results in
a lower heat flux inside the film boiling regime. The Leidenfrost temperature
moves towards higher temperatures. This influence is nonlinear, meaning the
difference between 0◦ and 15◦ is smaller than between 30◦ and 45◦ . Because
the data of the radial distribution of the mass flux is only available for the
normal impact case, the shown data is evaluated in the central area below
the nozzle.
Since the target is rotated around an axis that is aligned with the geometrical
center, it can be assumed that the mass flux in the center remains nearly
constant. In this configuration the only changing parameter, when tilting
the target surface, is the wall-normal impact velocity of the droplets. This
velocity can be calculated by simple geometric considerations,
Uβ = cos(β)U, (3.1)
46
Figure 3.12: Influence of the spray impact angle β.
where U is the mean velocity measured by the phase Doppler system. There
is no reason for a changing droplet diameter and the mass flux is assumed to
be constant.
In Fig. 3.13 the scaled heat flux in the film boiling regime q̇β /q̇ is plotted
as a function of the scaled impact velocity Uβ /U for different experiments.
Here the indices β correspond to the different impact angles (0◦ , 15◦ , 30◦
and 45◦ ) and q̇ is the heat flux obtained during wall-normal impact (0◦ case).
The spray impact parameters of the different experiments are indicated in the
figure by different colors. All experimental data collapse onto one straight line
having a slope of unity. This indicates that the heat flux in the film boiling
regime of an oblique spray impact scales with the wall normal spray impact
velocity. Knowing the heat flux of the normal impact q̇, the heat flux for the
oblique impact can be calculated according to
Uβ
q̇β = q̇ = q̇ cos β. (3.2)
U
3.3.3 Initial spray water temperature

Since there is obviously an influence of the spray fluid temperature on the heat
flux, experiments with different spray fluid temperatures 20 ◦ C < Tf0 < 80 ◦ C
were performed.
47
Figure 3.13: Scaled heat flux in the film boiling regime q̇β /q̇ as a function of the
scaled impact velocity Uβ /U for different spray impact angles β. The different
colored lines represent different kinds of sprays.
Figure 3.14: Heat flux dependence on the surface temperature for different spray
water temperatures. The spray parameters are: ṁ = 0.9 kg/m2 s, D10 = 43 µm
and U = 9.9 m/s. (Adapted from Tenzer et al. (2019a).)
48
In Fig. 3.14 the heat flux q̇ is plotted as a function of the surface tempera-
ture Ti . Since the surface is continuously cooled and the heating is turned
off at the beginning of the spraying process, the curves in this figure can
be read as increasing time (t) from higher to lower surface temperatures.
The spray parameters are ṁ = 0.9 kg/m2 s, D10 = 43 µm and U = 9.9 m/s.
For all different spray water temperatures Tf0 the commonly known boiling
regimes can be identified: film boiling regime, where the heat flux linearly
slightly decreases, at high surface temperatures from the beginning of the
experiment (Ti = 450 ◦ C) until Leidenfrost point is reached (Ti ≈ 340 ◦ C).
After passing the Leidenfrost point the transition boiling regime follows, which
is characterized by a strong increase of the heat flux. It ends at the critical
heat flux point (maximum heat flux) at Ti = 215 ◦ C. After that point the
nucleate boiling regime follows, after which the heat flux again decreases.
We can conclude that an increase of the spray water temperature yields
a decrease of the heat flux at all surface temperatures and therefore for all
boiling regimes. Especially in the film boiling regime and at the critical heat
flux, the heat flux is obviously decreased by the higher water temperature. We
can further identify the non-linearity of the effect: the drop of the heat flux
between Tf0 = 30 and 50 ◦ C is higher than that between Tf0 = 50 and 79 ◦ C.
For further investigation, in Fig. 3.15 we expand the variation of the different
water temperatures and only show the film boiling regime and the beginning of
the transition boiling regime. The previously mentioned outcomes can easily
be confirmed. In addition, the non-linearity results in nearly no difference
at high water temperatures (between Tf0 = 69 and 79 ◦ C). The strongest
influence is at low water temperatures and continuously decreases towards
higher water temperatures.
Figure 3.15 also leads to a deeper insight into the Leidenfrost point: For
the lowest water temperature the Leidenfrost temperature is TL ≈ 345 ◦ C and
for the highest water temperature TL ≈ 350 ◦ C. The results of the other spray
water temperatures are in-between these boundaries. Keeping in mind that
the Leidenfrost point is not a distinct point, but more a region where the slope
of the curve changes, it is obviously difficult to determine the Leidenfrost
point. Therefore, and because of the deviation of only 5 ◦ C being very small
compared to the change of the water temperature (50 ◦ C), the influence of the
water temperature on the Leidenfrost point is also weak or even nonexistent.
49
water temperatures. Only film boiling regime, Leidenfrost point and beginning of
transition boiling regime are shown. Spray parameters are equivalent to Fig. 3.14.
(Adapted from Tenzer et al. (2019a).)
In Fig. 3.16 similar results for a different spray are shown. The inserts show
representative pictures of the surface at the corresponding surface temperature
and are taken from videos capturing the spray impact. This time the spray
properties are ṁ = 2.8 kg/m2 s, D10 = 54 µm and U = 10.6 m/s. Especially
the strong influence of the higher mass flux results in an overall higher
heat flux compared to Fig. 3.14. The general trends are comparable to the
previously mentioned observations: Higher spray water temperature results
in a lower heat flux. The Leidenfrost temperature is only weakly influenced,
TL ≈ 320 ◦ C.
From the insert on the right we can identify single drop impact onto a dry
wall as the leading spray impact mechanism during film boiling at temperatures
above Leidenfrost. The insert on the left indicates that at temperatures below
Leidenfrost, first small patches of liquid remain and wet the surface (marked
in red). From the videos we can identify the Leidenfrost point as the instant,
when for the first time a very small wetted spot remains shortly after the
impacting drop left the surface, which was already discussed in the previous
sections. Also, shown only for one spray water temperature, these observations
are independent of the water temperature. We can conclude that different
spray water temperatures do not change the hydrodynamics at the surface.
50
water temperatures. The inserts show the conditions at the surface at the
corresponding surface temperature. The shape of the remaining liquid patch
at the surface below Leidenfrost is marked in red. The spray parameters are:
ṁ = 2.8 kg/m2 s, D10 = 54 µm and U = 10.6 m/s. (Adapted from Tenzer et al.
(2019a).)
51
Compared to Fig. 3.14 the influence of the water temperature on the heat
flux, as well as the non-linearity of the effect, are stronger. In addition, the
Leidenfrost temperature is lower, TL ≈ 320 ◦ C in contrast to TL ≈ 345 ◦ C in
Fig. 3.16.
These results are in strong contrast to the findings in literature. Both
experimental studies of Nimi et al. (2012) and Xu & Gadala (2006) conclude
a strong decrease of the Leidenfrost temperature for hot water. In our
experiments this decrease is not observed, which is confirmed by data from both
measurement techniques (inverse heat conduction based on thermocouples
and visual observation). From the visual observation we identify single drop
impact in the film boiling regime. This leads to the comparison of our spray
experimental results to those experimental results of single drop impact:
The Leidenfrost temperature of a single drop is not influenced by the water
subcooling (Bernardin & Mudawar, 1999), which is in accordance to our
results. The Leidenfrost point in sprays and that of single drops share this
common feature.
The previously described different Leidenfrost temperature for both series of
experiments must be caused by the different sprays of the experiments, since
nothing else was changed. The visual observations indicate single drop impact
and no film building above Leidenfrost. Therefore we deduce no influence
of the mass flux on the Leidenfrost temperature. The remaining possible
influencing factors are the drop diameter and velocity. At this stage, from
our experiments shown in Figs. 3.14 to 3.16 we identify, that drops having a
larger diameter and being faster result in a lower Leidenfrost temperature.
This is in contrast to the findings of the influence of the drop velocity on the
Leidenfrost temperature of a single drop in the experimental studies of Celata
et al. (2006); Bernardin & Mudawar (2004); Testa & Nicotra (2009). Here a
higher drop velocity results in a higher drop Leidenfrost temperature.
Let us further try to understand the mechanisms leading to the influence
of the water temperature on the heat flux in the film boiling regime. The
impacting drop hits the surface and stays in contact with the wall for a very
short time. Immediately, heat is transferred from the hot wall to the liquid
of the drop by heat conduction. The sensible heat of the drop increases. A
thermal boundary layer develops. When the saturation temperature in the
liquid is reached at the interface between the drop and the wall, a vapor layer
forms and separates the drop from the wall. There is no contact between the
52
Figure 3.17: Heat flux as a function of the initial spray water temperature. The
data and spray parameters are equal to Fig. 3.15. (Adapted from Tenzer et al.
(2019a).)
drop and the wall anymore. The drop breaks up due to thermal atomization
(Roisman et al., 2018) or inertia. The sensible heat which is necessary to
heat the liquid to saturation temperature, so vapor is generated, is higher for
a cold water drop. This means more heat from the wall goes into sensible
heat of the drop. Therefore, less vapor is generated in the case of cold water
compared to the case of a higher initial water temperature. Less vapor results
in a smaller vapor layer thickness (Breitenbach et al., 2017b) and a smaller
thermal resistance. Therefore the heat flux is higher. Since spray impact in
the film boiling regime is a superposition of single drop impacts, the previously
discussed behavior of single drop impact can be directly transferred to spray
impact.
In Fig. 3.17 we replot the data from Fig. 3.15. This time the heat flux at
different surface temperatures Ti is plotted as a function of the spray water
temperature. For low water temperatures Tf0 < 70 ◦ C the influence is nearly
linear. For higher water temperatures the slope starts to flatten which leads
to non-linearity. The slope is comparable for all surface temperatures Ti .
From the experimental results it is obvious, that the sensible heat plays an
important role during spray impact at temperatures above Leidenfrost.
53
Figure 3.18: Influence of the initial wall temperature Tw0 for a small mass flux
of ṁ = 1.6 kg/m2 s. (Adapted from Tenzer et al. (2019b), with permission of
3.3.4 Initial substrate temperature
In Fig. 3.18 the heat flux and surface temperature are shown as a function
of time for various initial substrate temperatures: 350, 400 and 450 ◦ C. The
spray properties are ṁ = 1.6 kg/m2 s, D10 = 64 µm and U = 8 m/s. To
highlight the transient behavior the temporal axis is limited to t < 50 s. Some
minor influence of the initial temperature on the heat flux in the film boiling
regime can be identified. However, the overall trend of the curves remain
the same. The time shift is a direct result of the different initial substrate
temperatures.
Similar curves, but this time for a larger mass flux, are shown in Fig. 3.19.
The spray properties are ṁ = 9.3 kg/m2 s, D10 = 48 µm and U = 15.6 m/s.
Again the slopes are comparable but the heat flux at the critical heat flux is
higher for the higher initial substrate temperatures. Furthermore, no typical
film boiling regime is visible. This can again be explained by the small time
scales in the film boiling regime which are of the same order as the rise time
of the thermocouples.
54
Figure 3.19: Influence of the initial wall temperature Tw0 for a large mass flux
of ṁ = 9.3 kg/m2 s. (Adapted from Tenzer et al. (2019b), with permission of
3.3.5 Wall thermal properties

Since spray cooling is used in many different cooling applications, there is also
a large variety of materials that need to be cooled. This arises the question
in which manner the thermal properties of the wall influence the heat flux
during spray cooling. To capture this influence, experiments with different
wall materials were performed during this investigation. A detailed description
of the respective targets and their thermodynamic properties can be found in
Table 2.3.
In Fig. 3.20a the heat flux is plotted as a function of the surface temperature
for different materials. The spray parameters are held constant for all shown
experiments, ṁ = 3.1 kg/m2 s, D10 = 60 µm and U = 9.3 m/s. The general
trend of all curves is comparable and corresponds to the previously described
behavior of the other cooling experiments. By comparing the values of the
heat flux between each material, we can identify an influence of the substrate
material on the heat flux. In all boiling regimes the heat flux is lowest for
stainless steel, followed by hot-work tool steel and nickel. Especially in the
film boiling regime we can identify that the difference between stainless steel
and hot-work toll steel is smaller than between hot-work tool steel and nickel.
This indicates a positive correlation between thermal conductivity of the
substrate material and heat flux. The critical heat flux as well as the heat flux
55
Figure 3.20: Influence of the substrate material on the heat transfer for a constant
rather sparse spray.
in the film boiling regime follow the same trend. Comparing the Leidenfrost
temperature for the three different materials, a negative correlation between
this temperature and the thermal conductivity can be identified. This means
a higher thermal conductivity results in a lower Leidenfrost temperature. In
Fig. 3.20b the corresponding dependence of the heat flux on time is shown.
Here we can again identify the same general slopes for all curves. Since all
curves end when a surface temperature of 100 ◦ C is reached, the curves show
an influence of the substrate material on the cooling time. The experiment
performed with stainless steel takes a shorter time, followed by hot-work tool
steel and nickel. This temporal difference arises mainly during the film boiling
regime. The time between Leidenfrost point, critical heat flux and end of the
experiment is comparable for the different materials.
Fig. 3.21 shows a second exemplary set of experiments having the same
spray parameters but different substrate materials. This time the spray is
much more dense. The properties are ṁ = 11.9 kg/m2 s, D10 = 79 µm and
U = 10.2 m/s. The curve shows obviously different general trends. From
the curve measured with the stainless steel target (blue) no film boiling can
be distinguished. This curve does not exhibit the typical shape and boiling
regimes. As previously stated in Section 3.3.1, this is caused by the limited
temporal response of the measurement system in combination with the short
time until the critical heat flux is reached, as seen in Fig. 3.21b. The curve
56
Figure 3.21: Influence of the substrate material on the heat transfer for a constant
rather dense spray.
of hot-work tool steel (yellow) slightly resembles the typical trend we know
from previous experiments. However, the timescales are rather short, so that
the data inside the film boiling regime is not considered so reliable. The
experiment performed with nickel (red) clearly shows the typical behavior: We
can identify the film boiling, transition and nucleate boiling regime. The slope
in the film boiling regime is rather steep compared to previously discussed
results. This appears to be connected to the heat flux inside the film boiling
regime: A higher heat flux results in a steeper slope. Furthermore this slope
is highest at the beginning of the experiment, except for the part indicated by
dashed line, where the rise time prevents valid data. It flattens with increasing
time. Since the time in the film boiling regime is long (Fig. 3.21b), this data
is regarded as valid and can be used for further analysis.
Using nickel as target material, experiments with mass fluxes of up to

30 kg/m2 s can be performed, while still achieving reasonable results in the
film boiling regime. For stainless steel this range ends at approximately
8 kg/m2 s, depending on the other parameters. Therefore, for nickel a much
larger span of spray parameters can be used for investigating the influences
on heat transfer.
57
Figure 3.22: Influence of the surface roughness on the heat transfer, for a small
mass flux of 1 kg/m2 s.
3.3.6 Wall surface roughness
In many industrial processes surfaces spray cooling involves rough surfaces,

arising either from abrasive wear or for other reasons. To investigate the
influence of surface roughness on the heat transfer some preliminary experi-
ments on a sandblasted surface were performed. The surface material of both
targets used in these tests was stainless steel. Further details can be found in
Section 2.3.
58
A comparison between a smooth and rough target is shown in Fig. 3.22.

The spray properties of both experiments are identical: ṁ = 1 kg/m2 s,
D10 = 51 µm and U = 9.4 m/s. The heat flux of the rough target is higher
than that of the smooth target. Particularly notable is that for surface
temperatures above the Leidenfrost point, the heat flux for the rough surface
does not decrease with decreasing temperature, but increases.
In the nucleate boiling regime, after reaching the maximum heat flux, both
surfaces result in the same temperature dependence of heat flux (Fig. 3.22c).
For the rough surface a distinct Leidenfrost point is no longer distinguishable,
since there is no minimum of the heat flux following the film boiling regime.
The curve in this diagram suggests a lower Leidenfrost point due to the
change of the slope at Ti ≈ 335 ◦ C; however, since a minimum heat flux is not
achieved, no valid prediction is possible.
The corresponding temporal evolution in Figs. 3.22a and 3.22b show basi-
cally the same outcomes. The time spent in the film boiling regime is shorter
for the rough surface compared to the smooth one. This is in accordance with
the higher heat flux in the film boiling regime. The timescales spent in the
nucleate boiling regime are comparable.
For the sake of completeness Fig. 3.23 shows results for a slightly higher
mass flux of ṁ = 3.1 kg/m2 s. The other parameters are: D10 = 60 µm and
U = 9.3 m/s. For this case we cannot identify a film boiling regime for the
rough target. This is again due to the short timescales resulting from the
higher heat flux, which are then comparable to the rise time of the temperature
measurement system.
As a conclusion we can summarize that surface roughness strongly increases
the heat flux in the film boiling regime, whereas there is no impact on the
nucleate boiling regime. It is remarkable that the heat flux in the film boiling
does not decrease with decreasing surface temperature, which is well known
for smooth surfaces. Instead it exhibits the opposite behavior and increases
with decreasing surface temperature. It is not possible to develop a more
elaborate theory or explanation for this observation based only on the results
of these preliminary experiments. Instead, these results suggest a new topic
for further research.
59
Figure 3.23: Influence of the surface roughness on the heat transfer, for a small
mass flux of 3.1 kg/m2 s.
60
4 Heat transfer in film boiling regime
In this chapter a predictive model for the heat flux and surface temperature
in the film boiling regime is developed. Originally the model was derived for
the heated target built of stainless steel, which is discussed in the first part of
this chapter. In the second part the model is validated against experimental
data. Here we also take into account experimental data acquired with other
target materials, to show the universality of the theoretical model.
Parts of this chapter have already been published in Tenzer et al. (2019b).
4.1 Analysis of heat transfer during spray

cooling
Let us analyze heat transfer during very intensive transient spray cooling of an
initially uniformly heated substrate. Consider for simplicity one-dimensional
heat conduction in a semi-infinite solid substrate. This assumption is valid
√
for cases when the thickness of the thermal boundary layer, αt, is much
smaller than the thickness of the target, and the temperature gradients in
this boundary are much higher than the gradients associated with the spray
distribution in the radial direction. Here α = λ/ρcp is the thermal diffusivity,
where λ is the thermal conductivity, ρ is the density and cp is the heat capacity
of the target material.
In our case the material properties of the stainless steel substrate are:
λ = 16 W/mK, ρ = 7900 kg/m3 and cp = 542 J/kgK. The longest experiments
last about 200 s. This results in a thermal boundary layer thickness of about
30 mm which is comparable to half of the target height. Therefore the heat
conduction in the target can be considered as semi-infinite. Since the spray
parameters are nearly constant in the central area the assumption of a one-
dimensional problem is valid.
61
Consider also a coordinate system {z, t} fixed at the interface z = 0 of the

semi-infinite target, belonging to the interval 0 < z < ∞. The temperature
field T (z, t) in the target can be calculated by solving the one-dimensional
heat equation
∂T ∂2T
=α 2. (4.1)
∂t ∂z
Following Duahamel’s theorem (Özışık, 1980), the solution which satisfies
the boundary condition far from the target interface and the initial condition
T = Tw0 at (t = 0 ∧ z ∈ [0, ∞]) ∨ (t > 0 ∧ z → ∞), (4.2)
is
" #
Z t
z
T (t) = Tw0 + A(τ )erfc dτ, (4.3)
2 α(t − τ )
p
0
where A(τ ) is a function determined by the conditions at the target interface

z = 0, erfc is the complementary error function and Tw0 is the initial wall
temperature. As usual, the limit z → ∞ denotes a position at a finite distance
much larger than the thickness of the thermal boundary layer in the wall.
Equation (4.3) allows the general solution for the interface temperature Ti (t)
and for the heat flux q̇(t) to be determined
T (τ )
Z t Z t 0
w
Ti (t) = Tw0 + A(τ )dτ, q̇(t) = − √ √i dτ, (4.4)
0 π 0 t−τ
where
w = (4.5)
p
λρcp
is the thermal effusivity of the wall.
4.2 Modelling of temperature and heat flux

evolution
The theoretical model published by Breitenbach et al. (2017b) for the heat
transfer from a single drop and during spray cooling in the film boiling regime
is based on the analysis of the heat conduction in the substrate, heat convection
62
4.2 Modelling of temperature and heat flux evolution
in the liquid region and in an expanding thin vapor layer emerging between
the impacting drop and the very hot substrate. The mass flux of the vapor
generated at the lower liquid interface is determined from the energy balance
at this interface. This model has already been validated by comparison with
experimental results from literature (Wendelstorf et al., 2008). Moreover, the
predicted vapor layer thickness agrees well with the direct measurements of
Tran et al. (2012) and Chaze et al. (2019). Predictions for the evolution of
the heat flux q̇(t) also agree with the accurate measurements based on the
infrared technique (Chaze et al., 2019).
The total heat transferred during the impact of a single drop Qsingle is
determined by integration of the heat flux q̇(t) over the “apparent contact
area” during the contact time. It should be noted that the contact area cannot
be based on the drop spreading diameter, since the free lamella in the remote
regions can levitate (Roisman et al., 2018). Therefore the values of D2 and
D/U are used as the scales for the contact area and for the contact duration,
where U and D are the impact velocity and drop diameter, respectively.
This analysis (Breitenbach et al., 2017b) allows the heat flux during spray
impact in the film boiling regime to be predicted
q̇ = Sw (Ti − Tsat ), (4.6)

ṁ
S = 8.85χ h i, (4.7)
1/2
1 − b + (1 − b)2 + w
p
ρf D10 U 1/2
√
8(Ti − Tsat )2w 2 5w f (Tsat − Tf0 )
w= , b= , (4.8)
πλv ρf L πρf λv L
where L is the latent heat of evaporation, χ is a dimensionless fitting parameter

which depends on the substrate wetting properties and roughness (but is of
order unity), Tsat is the saturation temperature of the liquid and Tf0 is the
initial spray fluid temperature. All terms with the subscript “f” correspond
to the liquid (fluid) component, “w” to the wall and “v” to the vapor.
Since the sprays used in the experiments and in practical applications are
always polydisperse, the average drop diameter and velocity, D10 and U , are
used in the model. Thus the parameter χ inherently accounts also for the
influence of the drop size and velocity distributions.
63
Table 4.1: Typical conditions of spray cooling experiments of this study.

Quantity Value
Ti 340 ◦ C
Tsat 99 ◦ C
w 8432 J/Km2 s1/2
λv 0.0248 W/mK
ρf 998 kg/m3
L 2453 kJ/kg
f 1581 J/Km2 s1/2
Tf0 20 ◦ C
In this study the model (Eqs. (4.6) to (4.8)) is used for prediction of the
evolution of the wall temperature in time. The predictions are then compared
with our experimental data.
For typical experimental conditions of this study, shown in Table 4.1, the
estimated values for w and b defined in Eq. (4.8) are w ≈ 700 and b ≈ 25.
Since b 1 and b2 and w are of the same order of magnitude, the effect of
the dependence of S on the changing temperature Ti in the expression for w
can be neglected. The value of S can be estimated by using Tw0 instead of
Ti (t) in the Eq. (4.8).
Equations (4.4) and (4.6) lead to the following integral equation for the
surface temperature, presented in the dimensionless form
ξ
Θ0 (ζ)
Z
Θ(ξ) + √ dζ = 0, (4.9)
0 ξ−ζ
where the surface temperature is made dimensionless using
Ti (t) − Tsat
Θ= , ξ = tπS 2 , ζ = τ πS 2 . (4.10)
Tw0 − Tsat
64
4.2 Modelling of temperature and heat flux evolution
(a) Dimensionless surface temperature Θ(ξ) (b) Dimensionless heat flux Φ(ξ).
and its derivative Θ0 (ξ).
Figure 4.1: Theoretically predicted evolution of the dimensionless quantities as

functions of the dimensionless time ξ in the film boiling regime.
This equation can be solved subjected to the initial condition Θ(0) = 1. The
analytical solution for Θ(ξ) can be represented as a series
∞
X
Θ(ξ) = 1 + ai ξ i/2 , (4.11)
i=1
2 1 4
a1 = − , a2 =, a3 = − 2 , ... ,
π π 3π
2−i Γ(i + 1)
ai+1 = −ai i+1 i+3 (4.12)
Γ 2 Γ 2
where Γ is the gamma function. The series Eq. (4.12) converges on the interval
0 < i < 22. Therefore it is sufficient to calculate Eq. (4.11) in this interval.
The corresponding heat flux q̇ can be estimated using Eq. (4.4)
q̇ = Sw (Tw0 − Tsat )Φ(ξ), (4.13)

∞
Θ (ζ) 1X
Z ξ 0
Φ(ξ) = − √ dζ, Θ0 (ζ) = ai iζ i/2−1 . (4.14)
0 ξ−ζ 2 i=1
The solution for the dimensionless surface temperature Θ(ξ), its derivative
Θ0 (ξ) and the dimensionless heat flux Φ(ξ) are shown in Fig. 4.1.
65
(a) Dimensionless surface temperature Θ(ξ) (b) Dimensionless heat flux Φ(ξ)
(Eq. (4.10)). (Eq. (4.13)).
Figure 4.2: Evolution of the dimensionless quantities for the stainless steel target
as a function of the dimensionless time ξ for experimental data in the film boiling
regime and for the theoretical solution (Eqs. (4.9) and (4.14)). (Adapted from
Tenzer et al. (2019b), with permission of Cambridge University Press. © 2019
Cambridge University Press.)
4.3 Model validation

In the following section the model is validated against experimental data,
organized according to the material which is being considered.
Stainless steel target In Fig. 4.2a the measured evolution of the dimen-
sionless surface temperature Θ(ξ) as a function of dimensionless time ξ is
plotted for the theoretical prediction (Eq. (4.11)) and for experimental results.
The computation of the theoretical solution is based on the first 50 terms of
the series. It converges on the interval 0 < ξ < 22. The line shown for the
experiments is the mean and the error bars indicate the standard deviation
computed over 49 experiments. For the reduction of the experimental data Ti
and t, χ in (Eq. (4.7)) is fitted to the experimental data using a least square
fit, resulting in χ = 2.2. Only the experiments exhibiting clear film boiling
behaviour are chosen and the experimental data comprising the film boiling
regime are plotted. We skip the experiments showing no film boiling behaviour
because of the previously mentioned limited temporal response of the measure-
66
ment system. The experimental parameters corresponding to the experimental

data span the following ranges: ṁ = 0.5 − 9.1 kg/m2 s, D10 = 43 − 78 µm,
U = 6.7 − 15.9 m/s, Tw0 = 350 − 450 ◦ C and Tf0 = 18 − 80 ◦ C.
In Fig. 4.2b the corresponding evolution of the dimensionless heat flux
for the same experiments is shown and compared to theory. Although there
is obviously a large scatter in the experimental data, very good agreement
between the experiments and theory can be observed. Especially χ, being
close to unity, indicates the performance of the theory, which clearly captures
all main physical players. The shown data consists of different sources
like phase Doppler, patternator and heat flux data, each having numerous
sources of uncertainty. Keeping that in mind, the scatter is acceptable and
therefore the good agreement between experiment and theory indicates a good
understanding of the physics of spray cooling in the film boiling regime.
Hot-work tool steel target This work is motivated amongst others by

the hot forging industry and should contribute to a better design of cooling
processes. Therefore it is mandatory that the developed models are valid for
the materials used in hot forging.
In Fig. 4.3a the dimensionless surface temperature Θ(ξ) is plotted against
the dimensionless time ξ for the experimental data performed with the target
built of hot-work tool steel (1.2365) and the theoretical prediction (Eq. (4.11)).
Figure 4.3b shows the corresponding evolution of the dimensionless heat flux
Φ(ξ).
The thermal properties of the target, which can be found in Table 2.3 are
used for the data reduction and calculation of the theoretical prediction. The
parameter χ is again fitted to the experimental data. The resulting value
χ = 1.9 is different, but still comparable and quite close to the value of the
stainless steel target (as discussed in the previous paragraph). Keeping in
mind the limited amount of data for the hot-work tool steel targets, the
agreement is very good. This indicates the universality of the model.
Again we can identify a very good agreement between experimental data
and theoretical prediction. Compared to Fig. 4.2 the scatter is smaller,
which is again a result of the larger time scales due to the larger thermal
conductivity of the material, as discussed in Section 3.3.5. Due to the low
resistance against corrosion and oxidation, only a few experiments with the
target built of hot-work tool steel could be performed. Therefore the amount
67
(Eq. (4.10)). (Eq. (4.13)).
Figure 4.3: Evolution of the dimensionless quantities for the hot-work tool steel
target as a function of the dimensionless time ξ for experimental data in the film
boiling regime and for the theoretical solution (Eqs. (4.9) and (4.14)).
of data having different spray properties is also much smaller compared

to the experiments with other targets. The spray properties of these 6
experiments only spread across the following range: ṁ = 1.1 − 4.4 kg/m2 s,
D10 = 43−77 µm, U = 7.9−14.9 m/s, Tw0 = 410−449 ◦ C and Tf0 = 19−21 ◦ C.
Nickel target Fig. 4.4 shows again the validation of the model for the film
boiling regime against experimental data. This time the data is obtained with
the target built of nickel. The parameter χ is again fitted to experiments
yielding χ = 2.2, which is the same as for the stainless steel target. As
previously stated in Section 3.3.5, the nickel target leads to a much larger span
of different sprays which can be used for the cooling experiments. This becomes
particularly obvious from the much larger mass flux. The spray properties
underlying the 61 experiments are: ṁ = 0.6−29.5 kg/m2 s, D10 = 41−117 µm,
U = 5.2 − 23.7 m/s, Tw0 = 300 − 462 ◦ C and Tf0 = 18 − 23 ◦ C.
Again there is a very good agreement between experimental data and
theoretical prediction. Especially the strong decrease of the theoretical pre-
diction for small dimensionless times ξ in Fig. 4.4b is well represented by the
experiments. In Fig. 4.4a some small discrepancy between experiment and
prediction of the dimensionless surface temperature exist. These are caused
68
(Eq. (4.10)). (Eq. (4.13)).
Figure 4.4: Evolution of the dimensionless quantities for the nickel target as a
function of the dimensionless time ξ for experimental data in the film boiling
regime and for the theoretical solution (Eqs. (4.9) and (4.14)).
by the long duration of the experiments: The longer the experiments last,
the larger the thermal boundary layer grows. The higher thermal effusivity
supports this increase. Since the developed model is based on the assumption
of a semi-infinite substrate; hence, the long duration experiments are no longer
covered by the model. Nevertheless, the agreement between experiment and
model is still good. In contrast, the dimensionless heat flux is not affected
that much by the violation of the semi-infinity of the target. This is because
the dependence of the heat flux on the surface temperature (the slope in the
film boiling regime in e.g. Fig. 3.3) is rather small for low mass fluxes. Since
only the experiments performed with small mass fluxes are the ones that last
the longest, the influence of the violated assumption of semi-infinity on the
predicted heat flux is rather small.
For pneumatic atomizer For the sake of completeness, Fig. 4.5 shows
the comparison between experimental data and theoretical prediction for
experiments performed with the nickel target and the pneumatic atom-
izer. The spray properties of these 5 experiments cover the following nar-
row range: ṁ = 0.1 − 2.1 kg/m2 s, D10 = 12 − 24 µm, U = 12.7 − 20.8 m/s,
Tw0 = 448 − 450 ◦ C and Tf0 = 20 − 23 ◦ C.
69
(Eq. (4.10)). (Eq. (4.13)).
Figure 4.5: Evolution of the dimensionless quantities for the nickel target spayed
with a pneumatic atomizer as a function of the dimensionless time ξ for experi-
mental data in the film boiling regime and for the theoretical solution (Eqs. (4.9)
and (4.14)).
Figure 4.5a shows poor agreement between experiments and theoretical

prediction of the dimensionless surface temperature. Only a few experiments
correspond to the predicted trend. Again, we can explain this by the long
duration of the experiments, violating the assumption of semi-infinity. Since
most of the experiments in this case have a very low mass flux, the experiments
last very long times. In contrast, the agreement of the dimensionless heat flux
and its theoretical prediction is again very good.
As previously discussed, the validation of the theoretical prediction against
experimental data obtained with various spray parameters and substrate
materials clearly indicates the performance of the model. All influences are
well captured by the theoretical prediction. The main drawback, that the
surface temperature is not well predicted for experiments of long duration,
is not too serious: A simple estimation of the thermal conduction inside
the substrate can identify in which cases the assumption is valid or not.
Nevertheless, the underlying preposition of superimposing single drop impacts
in the film boiling regime is still valid. Especially the fact that the parameter
χ remains in a very narrow range around ≈ 2 and close to unity, indicates
the universality of the model.
70
4.4 Application of the model
To better compare theoretical predictions with experimental results, one

exemplary case is presented in Fig. 4.6 in dimensional form. The spray
parameters for this case are ṁ = 0.9 kg/m2 s, D10 = 43 µm and U = 10 m/s,
impacting onto the stainless steel target. These parameters are used in the
model (Eqs. (4.9) to (4.14)).
These figures indicate an obviously excellent agreement between experiment

and model. Especially the heat flux (Fig. 4.6a) is very well predicted by the
model. Slightly larger deviations arise for the surface temperature (Fig. 4.6b).
The dotted parts of the experimental curve indicate where the measurements
are not valid due to the limited temporal response of the measurement system.
The model is not formulated to account for the Leidenfrost effect; hence, the
theoretical curve is not valid for t > 40.
Examining the initial instants of spray impact (small t in Fig. 4.6a and
high Ti in Fig. 4.6c), the theory leads to a better understanding of the process.
At the very first instants of the spray impact, the heat flux is very high.
Subsequently, it decreases and converges to the well-known linear slope. The
very high heat flux in the first instants can be explained by the fact that
the first droplets impact onto a substrate having no thermal boundary layer,
resulting in an extremely high temperature gradient at the solid-liquid contact
interface. With time the thermal boundary layer in the substrate grows and
the temperature gradient decrease, leading to a near steady state heat flux.
This knowledge is of special interest for pulsed spray cooling, where the
spray is continuously switched on and off. During the period when no spray
impact occurs, heat flows from the deeper part of the substrate towards the
surface and decreases the thermal boundary layer. At the next instant, when
the spray is switched on again, spray impact takes place again at a more
homogeneously heated surface (compared to the case of continuous spraying).
This leads to a heat flux that again starts at a high level. Continuous pulsing
of the spray can therefore increase the amount of heat extracted from the
substrate while using an overall smaller amount of spray liquid.
71
(a) Heat flux evolution with time. (b) Surface temperature evolution with
time.
(c) Heat flux as a function of surface tem-

perature.
Figure 4.6: Exemplary comparison between one experiment and the theoretical
prediction in the film boiling regime. In a) and b) times are only shown up to
shortly after the critical heat flux point is reached.
72
Figure 4.7: Heat flux as a function of the initial spray water temperature. Com-
parison between experimental data and theoretical prediction. Experimental data
is equivalent to those in Fig. 3.17.
Initial spray water temperature In Fig. 4.7 the application of the theo-
retical prediction for experiments performed with various spray water temper-
atures Tf0 is shown. Here the heat flux, evaluated at the surface temperature
Ti = 370 ◦ C, is plotted as a function of Tf0 . The experimental data is equivalent
to those in Fig. 3.17.
Comparing experiment and theory leads to reasonably good agreement. At

low spray temperatures the slope of both lines is comparable. For higher
spray temperatures the model predicates a nearly linear decrease of the heat
flux, whereas the slope of the experimental data lessens. This deviation can
be explained by the following: Since the drops travel some distance between
the orifice of the nozzle and the surface, they are affected by the ambient
air and they cool down. Warmer drops lose more heat than colder ones,
resulting in a higher temperature drop. In the experiments the spray water
temperature is measured upstream of the nozzle and therefore differs from the
true drop temperature upon impact - the impacting drops are colder than Tf0
in the experiments. The effect becomes stronger for higher drop temperatures
and velocities and for smaller droplets. If this effect was accounted for, the
theoretical prediction would correspond better to the experimental results. In
conclusion, the theoretical prediction describes well the influence of the spray
water temperature on the heat flux.
73
Figure 4.8: Evolution of the heat flux as a function of time for an exemplary
experiment. The parameters of the spray are: ṁ = 1.4 kg/m2 s, D10 = 43 µm
and U = 11 m/s. The dashed vertical lines bound the transition boiling regime.
4.5 Transitional spray cooling regime below

the Leidenfrost point
At some instant of time tL the conditions at the surface correspond to the

Leidenfrost point. We identify the Leidenfrost point from the time evolution
of the measurement data as the point when the heat flux reaches its minimum
in the film boiling regime. The value of the surface temperature corresponding
to the Leidenfrost point is not a fixed value and will be discussed in further
detail in Chapter 6.
Denote the surface temperature at the Leidenfrost point as TiL . The
corresponding heat flux q̇L can be estimated using Eq. (4.6).
Further cooling leads to a rapid increase of the heat flux, caused by partial
wetting of the surface. The area of the wetted patches grows rapidly, promoted
by the decreasing surface temperature. The duration of this transitional spray
cooling regime is rather short. During this period, two modes of drop impact
can be observed: drop impact onto dry regions in the film boiling regime and
drop impact onto wetted patches.
74
4.5 Transitional spray cooling regime below the Leidenfrost point
In Fig. 4.8 an exemplary dependence of the heat flux on time is shown.

Here we limit the temporal span to the time between the end of the film
boiling regime and the beginning of the nucleate boiling regime in order to
illustrate the behaviour in the transition boiling regime. The corresponding
spray parameters are: ṁ = 1.4 kg/m2 s, D10 = 43 µm and U = 11 m/s. We
can clearly identify the instant when the Leidenfrost point occurs (tL ), at
around 7.6 s, and the instant when the critical heat flux is reached (tCHF ), at
8.7 s. This short time of slightly more than 1 s between both events is similar
to the rise time of the thermocouples. Therefore, we can obtain no reasonable
temperature readings in the range between both events and can make no
prediction of the surface temperature and the heat flux in the transition
boiling regime or at the critical heat flux.
This short time in the transition boiling regime is very small compared to
the overall cooling process and therefore the heat which is transferred during
this short time can be neglected. In the present case the transition boiling
regime plays no role with respect to the cooling process. Having this short
time in mind, we can imagine the transition boiling regime to be non-existent
and replaced by a jump of the heat flux towards a very high value at the
Leidenfrost point. The typical trend we see in our measurements of heat flux
is only due to the limited temporal resolution of the measurement equipment
and is probably only valid for very low mass flux or cooling rates in general.
The above mentioned time of 1 s is the result of an experiment having a small
mass flux. Typical experiments having a higher mass flux lead to even smaller
times, which support the previously described understanding of the transition
boiling regime in our experiments.
75
76
5 Heat transfer in nucleate boiling
regime
This chapter describes the development and application of a model for the heat
flux in the nucleate boiling regime. First, the underlying assumptions of the
model are presented, followed by the formulation, validation and exemplary
application of the model. Initially the model was developed for the stainless
steel target; however, other target materials are also taken into account.
Parts of this chapter have already been published in Tenzer et al. (2019b).
5.1 Modelling heat flux as a remote

asymptotic solution
For longer times t > tL , during the nucleate boiling regime of spray cooling,
the heat flux can be estimated from Eq. (4.4) accounting for the very short
duration of the transient boiling regime and the very high time derivation
of temperature during this regime. Denote ∆TL as the temperature jump
during the transition boiling. In this case Eq. (4.4) yields
tL 0
Ti,film (τ ) w ∆TL t 0
Ti,nucleate (τ )
Z Z
w w
q̇(t) = − √ √ dτ + √ √ −√ √ dτ. (5.1)
π 0 t−τ π t − tL π tL t−τ
On the right-hand side of Eq. (5.1) the first term is associated with the
thermal history during the film boiling regime, the second term is associated
with the temperature jump ∆TL during the transition regime at τ = tL , and
the last term is based on the temperature evolution during the nucleate boiling
at times tL < τ < t.
In order to model the heat flux in the nucleate boiling regime the values of
∆TL and the evolution of the surface temperature Ti (τ ) (which is required for
the computation of the time derivative Ti0 (τ )) are necessary.
77
5 Heat transfer in nucleate boiling regime
In the estimation of an upper bound for heat flux during the nucleate boiling
regime of single drop impact the temperature at the wetted part of the wall
interface is approximated by the saturation temperature Tsat (Breitenbach
et al., 2017a).
The nucleate boiling regime is characterized by intensive nucleation and
expansion of vapor bubbles. The temperature in the vicinity of the contact line
of the each expanding bubble is close to the saturation temperature. However,
some overheating of the surrounding liquid is required for the bubble growth.
The heat transfer in the liquid phase during the nucleate boiling regime is
governed by the convection in the liquid flow between the bubbles. The
heat mainly goes into vaporization at the bubble interfaces where T = Tsat .
Therefore, the upper bound for the heat flux during nucleate boiling regime
can be estimated invoking the assumption that the temperature at the wetted
wall interface is Tsat . Using this rough estimation of q̇ the predicted duration
of a single drop evaporation in the nucleate boiling regime agrees very well
with numerous experimental data (Abu-Zaid, 2003; Itaru & Kunihide, 1978;
Tartarini et al., 1999).
In this study the upper bound for the heat flux q̇ during spray cooling is
also estimated, as in the case of a single drop impact, using the assumption
that the substrate temperature is equal to the saturation temperature at
t > tL . The third term on the right-hand side of Eq. (5.1), associated with
the time gradient of the surface temperature at t > tL , can be neglected in
comparison to the effect of the temperature jump at the Leidenfrost point.
The temperature jump during the transition boiling regime can be estimated
as ∆TL = TiL − Tsat .
The heat flux can be obtained from Eq. (5.1), neglecting the value of the
last term, in the form
∞
ξL i 1 i−1

w TiL − Tsat Tw0 − Tsat X
q̇ = √ √ − Sw ai iB ; , ξ 2 , (5.2)
π t − tL 2 i=1
ξ 2 2
where B[·; ·, ·] is the incomplete beta function. Equation (5.2) is valid only for
very fast substrate cooling, when the time interval between the Leidenfrost
point and the point corresponding to the critical heat flux is very short. It
78
can be further modified using Eq. (4.10)

" √ √ ∞
#
ξL i 1 i−1

w Tw0 − Tsat S π t − tL X
q̇ = √ √ Θ(ξL ) − ai iB ; , ξ 2 . (5.3)
π t − tL 2 i=1
ξ 2 2
Moreover, at large times, t tL , Eq. (5.3) approaches the following remote

asymptotic solution
√
2S tL

w Tw0 − Tsat
q̇ ≈ √ √ Θ(ξL ) + √ . (5.4)
π t − tL π
Rearranging Eq. (5.4) leads to
2w ∆T 2
T ≡k 2 ≈ t − tL , (5.5)
π q̇(t)
where k is a constant which can be determined from the experiments and

∆T = Tw0 − Tsat .

Stainless steel target In Fig. 5.1 the values of the term T = k2w ∆T 2 /π q̇ 2 ,
measured for various spray parameters and different initial substrate tempera-
tures, are shown as a function of t − tL . The constant k = 1.05 is determined
for all the sets of the operational parameters to best fit the theory. The
parameters of the 86 experiments used in the plotting of the data in Fig. 5.1
are varied over wide ranges: mass flux ṁ from 0.5 to 29.5 kg/m2 s, average
drop diameter D10 from 43 to 78 µm, average impact velocity U from 6.7 to
17.7 m/s, initial wall temperature Tw0 from 350 to 450 ◦ C and spray fluid
temperature Tf0 from 18 to 80 ◦ C. Additionally, a line corresponding to the
average value of T for all the experiments and the error bars indicating one
standard deviation are shown in the graph.
We can identify an excellent agreement between theoretical prediction and
experimental data over a broad range of spray parameters. As predicted by
the remote asymptotic solution (Eq. (5.5)), the term T is very close to the
time increment t − tL in all the experiments.
79
Figure 5.1: Scaling of the heat flux, expressed through T defined in Eq. (5.5),
plotted as a function of time for the stainless steel target. The straight line
has a slope of unity and indicates good agreement between the theory and the
experimental data. (Adapted from Tenzer et al. (2019b), with permission of
Hot-work tool steel target Similarly, Fig. 5.2 shows the comparison
between theoretical prediction and experiential data for the target built of hot-
work tool steel. The spray properties of these 6 experiments cover the following
narrow range: ṁ = 1.1 − 4.6 kg/m2 s, D10 = 43 − 79 µm, U = 7.8 − 14.3 m/s,
Tw0 = 410 − 449 ◦ C and Tf0 = 19 − 21 ◦ C.
The constant k = 1.05 is the same as in the previous case of the stainless
steel target. Again we can identify very good agreement between theoretical
prediction and experimental data for larger times, when the critical heat flux
has been passed.
Nickel target In Fig. 5.3a the same comparison is shown for the target built
of nickel. The spray properties of these 59 experiments span the following
ranges: ṁ = 0.6 − 30.3 kg/m2 s, D10 = 43 − 114 µm, U = 5.1 − 22.5 m/s,
Tw0 = 300 − 462 ◦ C and Tf0 = 18 − 23 ◦ C. In contrast to the other materials,
we identify good agreement between theory and experimental data only for
small times t < 20 s. At larger times the experiments differ from the theoretical
prediction. This can be explained by the thermal boundary layer inside the
substrate. Since the thermal diffusivity of nickel is significantly larger than
those of stainless steel and hot-work tool steel, the thermal boundary layer
growths much faster for experiments performed with nickel. At some instant
80
Figure 5.2: Scaling of the heat flux, expressed through T , defined in Eq. (5.5),
plotted as a function of time for the target built of hot-work tool steel.
the boundary layer reaches the bottom of the target and the assumption of
semi-infinity is violated. Therefore the theory (Eq. (5.5)) is no longer valid
and experimental data and theoretical predictions differ.
In Fig. 5.3b we show only those experiments where the assumption of semi-
infinity is justifiable. If the temperature difference at the bottom between the
beginning and the end of the experiment is larger than 30 ◦ C, the experiments
are treated as significantly not semi-infinite and are not included in Fig. 5.3b.
It turns out that this constraint leads to basically the same amount of
experimental data that spans across the same temporal range (t < 20 s) as
observed in Fig. 5.3a.
Finally, we observe good agreement between experimental data and the-
oretical predictions as long as the assumption of semi-infinity is valid. Our
experiments show that the heat flux in the fully developed nucleate boiling
regime depends significantly on time. It is estimated as
w Tw0 − Tsat
q̇ ≈ √ √ . (5.6)
π t − tL
Note that the measured values for T deviate significantly from the theoret-
ical predictions at small times, associated with the film boiling and transition
regimes, for which the scaling (Eq. (5.5)) is not applicable.
81
(a) All experiments, including those where (b) Only experiments, where the target is
the target is not semi-infinite. semi-infinite.
Figure 5.3: Scaling of the heat flux, expressed through T defined in Eq. (5.5),
plotted as a function of time for the nickel target.
The remote asymptotic solution (Eq. (5.6)) is valid only for a semi-infinite
hot substrate and uniform spray. In practice this means that the thickness
and the width of the substrate are larger than the thickness of the thermal
√
boundary layer in the substrate αt.
5.3 Heat flux and its upper boundary

The measured heat flux during nucleate boiling regime is described very well
by the upper boundary estimation (Eq. (5.6)) obtained from the assumption
that the interface temperature is equal to Tsat . The reason for this agreement
is not immediately clear since some liquid overheat is expected during the
nucleate boiling.
This interesting result can be explained by the fact that the interface tem-
perature oscillates due to the nucleation of multiple bubbles, their expansion
and subsequent collapse (Carey, 2018). The bubble contact lines of the wall-
bounded bubbles, whose temperatures are close to Tsat , quickly propagate
along the interface. The characteristic time of bubble formation tbubble in
the experiments is approximately 1 ms (Breitenbach et al., 2017a). Therefore,
the thickness of the thermal boundary layer associated with a single bubble
82
prediction in the nucleate boiling regime (Eq. (5.6)). Heat flux evolution with
time.
√
event is hbubble ∼ αw tbubble . This value is approximately 70 µm in our
case. In any case, our measurement system is not able to detect such rapid
temperature fluctuations.
At any given location on the wall surface the temperature jumps to the
saturation temperature each time the contact line propagates through this
position. Our measurements allow estimation of only the averaged value of
the temperature fluctuations above the saturation temperature. Nevertheless,
the contribution of the temperature fluctuations (consisting of positive and
negative jumps above Tsat ) to the time averaged heat flux q̇ is negligibly small.

To better compare the theoretical predictions with experimental results, one
exemplary case is shown in Fig. 5.4 in dimensional form. The experiment cor-
responds to that in Section 4.4, but this time we focus on the nucleate boiling
regime. Accordingly, the spray parameters for this case are ṁ = 0.9 kg/m2 s,
D10 = 43 µm and U = 10 m/s.
The dashed line representing the theoretical prediction is calculated using
Eq. (5.6). For clarification it is mentioned that the Leidenfrost point and
time when it occurs tL is taken from the experimental data. In the present
case tL = 40 s.
83
Although the curve shows the entire experiment, the theoretical prediction
is only valid after the Leidenfrost point is passed. This range corresponds to
t > tL = 40 s. At these first instants the predicted heat flux starts at infinity
and strongly decreases. The agreement between experiment and model is
obviously not good in this interval. Keeping in mind the limited temporal
resolution of the measurement system, the disagreement is not unexpected.
A few instants later, roughly at the critical heat flux (maximum heat flux),
the prediction matches almost perfectly the experimental data. The excellent
agreement persists until the end of the experiment.
In summary, we find excellent agreement between experiment and theoretical
prediction (Eq. (5.6)) after the first instants. Furthermore, the very good
agreement leads to a new understanding of the heat transfer in the nucleate
boiling regime during transient spray cooling: The heat flux is limited by the
thermal effusivity w of the substrate and the initial wall temperature Tw0 .
It depends on time. There is no or only a very weak dependence of the heat
flux on the spray properties. In other words, the heat flux is limited by the
thermal inertia of the material. This means that there is no possibility to
influence the heat transfer in the nucleate boiling regime with different spray
properties. Even the smallest mass flux of all our experiments is sufficiently
large enough to lead to this state of saturation where the material limits the
heat flux even if the spray could lead to a higher one.
The complete heat flux dependence on time for the entire experiment is
calculated by combining the heat fluxes predicted by the models for both
regimes (Eqs. (4.9) to (4.14) and (5.6)) as
(
Sw (Tw0 − Tsat )Φ(ξ) for t < tL
q̇S (t) = (5.7)
√
π
√ −Tsat
w Tw0
t−tL
for t > tL .
As described and justified in Section 4.5 the transition boiling regime is skipped
and replaced by a jump at the Leidenfrost point. Again, the only necessary
addition information is tL = 40 s, which is taken from the experiments. For
the sake of completeness, the temporal evolution of the resulting overall heat
flux expressed by Eq. (5.7) is shown in Fig. 5.5. Obviously, the heat flux in
the transition boiling regime is strongly over-estimated due to replacing this
regime with a jump. From the experimental point of view we are not able
to confirm or disprove whether this jump is reasonable or in which way the
84
prediction of the entire cooling process (Eq. (5.7)). Heat flux evolution with time.
jump is attenuated. Since the Leidenfrost point corresponds to the onset of

remaining wetting, we can assume a change of the entire physics governing
heat transfer. Furthermore, the visual observations in Section 3.2 show that
the transition boiling regime exhibits the beginning of the formation of a
closed liquid film at the surface. Both are consistent with a jump of the heat
flux. Nevertheless, it is obvious that the formation of the liquid film is not
an instantaneous event and instead takes some time. The duration is longer
for lower mass fluxes, since the lower mass flux leads to longer times until
the liquid film is formed. Therefore, the agreement between prediction of the
overall heat flux (Eq. (5.7)) and experiment can be expected to improve with
increasing mass flux.
For higher cooling efficiency, the independence of the heat flux on the spray
parameters suggests a cooling process that consists of two stages: Spraying
with high mass fluxes in the film boiling regime to reach a maximum heat flux.
Subsequently, switching to a lower mass flux in the nucleate boiling regime,
reduces the necessary spray fluid while maintaining the heat flux, since it is
no longer influenced by the spray properties.
85
86
6 Leidenfrost point
The Leidenfrost point corresponds to the instant when the minimum heat flux
is reached and afterwards strongly increases. For this reason it is of major
interest to reliably predict this point in any spray cooling process. Moreover,
it is the last remaining part of the boiling curve which has yet to be described
to finally predict the heat flux during the entire spray cooling process.
The following chapter is dedicated to a better understanding of the Lei-
denfrost point. In the first part the experiential results of this study are
analyzed to identify any possible dependence on the various spray param-
eters. Subsequently, our results are compared to published predictions in
literature. Finally we identify the dependence on the substrate and derive a
new understanding and universal prediction for the Leidenfrost point.
Parts of this chapter have already appeared in Tenzer et al. (2020).
6.1 Influences on the Leidenfrost point

The visual observations in Section 3.2 indicate that the Leidenfrost point
corresponds closely to the first instants in which fluid remains on the surface.
At temperatures higher than the Leidenfrost temperature, no sessile drops or
wetted patches are observed on the surface. When the surface temperature
is slightly lower, the amount of liquid remaining on the surface immediately
after the drop impact begins to increase. This trend continues for even lower
temperatures: the amount of remaining liquid rises and larger patches of
liquid form, as shown in Fig. 3.16.
The experimental data of this study span a large parameter range. This
allows quantitative investigation of the Leidenfrost point and its influencing
parameters. The Leidenfrost point and its temperature is determined for each
experiment, where film boiling can be clearly identified by the minimum heat
flux. Figure 6.1 illustrates this procedure, where qmin is the minimum heat
flux and TL the Leidenfrost temperature. These experiments are characterized
87
6 Leidenfrost point
Figure 6.1: Determination of the Leidenfrost point TL as the temperature at the

minimum heat flux qmin .
by a typical linear increasing heat flux in the film boiling regime. Other
experiments not showing this behavior are discarded, since in these cases
the cooling is too intensive and the timescales are too short to be properly
captured by the thermocouples. Although the linear slope is also present
in these discarded experiments, it cannot be detected by the measurement
technique, as previously discussed in Section 3.3.1.
Even though the literature survey in Section 1.2 reveals no clear picture of
the dependencies of the Leidenfrost point, there is general agreement that the
mass flux plays an important role. To check whether this influence is present
in our experimental data Fig. 6.2 shows the Leidenfrost temperature TL as a
function of the local mass flux ṁ. Here the different materials of the target,
stainless steel and nickel, are indicated by different markers and colors. Since
only a few experiments were performed with hot-work tool steel, these results
are not shown here and in the following discussion. Although the data scatters
a lot, there is no dependence of the Leidenfrost temperature on the mass
flux for both materials. This clearly becomes apparent from the experiments
performed with nickel. For the very large range of mass fluxes the Leidenfrost
temperature remains constant. Due to high cooling rate in the experiments
performed with stainless steel, the data only spans a narrow range of mass flux.
But even in this case no influence is visible. Compared to literature we detect
no influence of the mass flux on the Leidenfrost temperature. Additionally,
88
6.1 Influences on the Leidenfrost point
Figure 6.2: Leidenfrost temperature TL as a function of the mass flux ṁ. (as can
be read in Tenzer et al. (2020).)
the Leidenfrost temperature is higher for stainless steel than for nickel. This
leads to the assumption that the Leidenfrost point is dependent on the target
material.
The observations in Section 3.2 and the modeling in the film boiling regime
in Chapter 4 have shown that spray impact in the film boiling regime consists of
a superposition of single drop impacts. Since the Leidenfrost point temporally
follows the film boiling regime and starts to take place during this regime, we
can assume that the Leidenfrost point is connected to the impact of a single
drop.
To check this assumption Fig. 6.3 shows the dependence of the Leidenfrost
temperature on the spray properties: mean drop diameter D10 in Fig. 6.3a
and mean drop velocity U in Fig. 6.3b. In both cases no dependence can be
observed. Despite the scatter, the Leidenfrost temperature is constant and
independent of the mean drop diameter or mean drop velocity of the spray.
The shown data also includes the experiments performed with the pneumatic
atomizer. Due to the smallest drop diameter of ≈ 20 µm, these points are
clearly visible in Fig. 6.3a. In this case the Leidenfrost temperature is also
constant and comparable to the other experiments. It is remarkable that even
for sprays produced by different principles of atomization, the Leidenfrost
temperature is constant.
89
6 Leidenfrost point
(a) Mean droplet diameter D10 . (b) Mean droplet velocity U .
Figure 6.3: Dependence of the Leidenfrost temperature TL on the spray properties.

(as can be read in Tenzer et al. (2020).)
(a) Weber number We. (as can be read in (b) Reynolds number Re.
Tenzer et al. (2020).)
Figure 6.4: Dependence of the Leidenfrost temperature TL on dimensionless num-

bers.
90
6.2 Comparison with literature
Furthermore, the influence of two combinations of these quantities are

presented as dimensionless numbers: the single drop Weber number We
in Fig. 6.4a and the single drop Reynolds number Re in Fig. 6.4b, both
defined in Section 2.2.3. In both cases there is again no dependence of the
Leidenfrost point on the Weber number or the Reynolds number. The lowest
Weber number of the experiments is Wemin = 25. The literature survey
in Section 1.2 showed an influence of the Weber number on the Leidenfrost
temperature during single drop impact. This dependence shows an asymptotic
behavior towards large Weber numbers: The influence is large for small Weber
numbers and becomes very small for large ones. The independent and constant
Leidenfrost temperature in Fig. 6.4a indicates, that Wemin = 25 of this
experiments is already sufficiently large enough and in the asymptotic range.
Thus, the Weber number has no identifiable influence on the Leidenfrost
temperature anymore.
Following the previous discussion, we conclude that there is no influence
of any spray parameter on the Leidenfrost temperature, which is in strong
contrast to the findings in literature. Except for some small scatter the
temperature at the Leidenfrost point is constant in our study. This scatter can
be explained by the inhomogeneity of the spray: The atomization process of
a conventional atomizer is not perfectly uniform due to mechanical tolerances
and imperfections. Therefore larger drops or even large non-spherical ligaments
occur randomly. These non-uniformities of the spray result in a strong local
and non-homogeneous heat flux which influences the local wetting behavior
at the surface. The randomness of this process is not captured by the spray
characterization or by the two-dimensional heat flux calculation. Thus, their
effect is not visible in the previously discussed dependencies of the Leidenfrost
point in Figs. 6.2 to 6.4. Instead, this is likely the explanation for the observed
scatter of the observed Leidenfrost temperature.

As already discussed in Section 1.2 there exist only a few models to describe or
predict the Leidenfrost point. In the following section our results are compared
to those models or correlations found in literature that are explicitly meant
for sprays.
91
6 Leidenfrost point
Figure 6.5: Comparison between own data and the correlation proposed by Al-
Ahmadi & Yao (2008) (Eq. (1.2)).
Figure 6.5 shows the experimental data of this study in comparison to the
correlation proposed by Al-Ahmadi & Yao (2008) which presumes an influence
of the mass flux (Eq. (1.2)). As discussed before, the data of this study shows
no influence of the mass flux on the Leidenfrost temperature. Therefore
there is obviously no agreement between our experiments and the proposed
correlation. Instead the correlation strongly over-predicts the Leidenfrost
temperature of up to 800 ◦ C.
In Fig. 6.6 the experimental data of this thesis is compared to the model
proposed by Bernardin & Mudawar (2004) (Eq. (1.6)). This model predicts an
influence of the droplet velocity. As can be seen, the Leidenfrost temperature
predicted by the model and the experimental data span approximately the
same range. But again, the experimental data show no influence of the droplet
velocity. In contrast the model forecasts that the Leidenfrost temperature
increases with an increasing droplet velocity. We can identify no agreement
between prediction and our experimental data.
As already shown in Section 1.2, Yao & Cox (2002) plotted the experimental
data of various researchers (Yao & Choi, 1987; Choi & Yao, 1987; Hoogendoorn
& den Hond, 1974; Mizikar, 1970; Ito et al., 1991; Shoji et al., 1984; Cox &
Yao, 1999) as a function of the spray Weber number Wes . They assumed that
the Leidenfrost temperature depends on Wes and fitted a correlation resulting
in Eq. (1.5). Figure 6.7 shows an adaption of their original illustration. To
provide a good comparison to the original illustration, the logarithmic axes
92
Figure 6.6: Comparison between own data and the correlation proposed by
Bernardin & Mudawar (2004) (Eq. (1.6)).
Figure 6.7: Comparison between own data, those of other researchers and the
correlation proposed by Yao & Cox (2002) (Eq. (1.5)). (Adapted from Yao &
Cox (2002), with permission of Taylor & Francis. © 2002 Taylor & Francis.)
93
6 Leidenfrost point
and their scale are kept the same. In addition the experimental data of the
present study is added. The correlation seems to fit quite well to the ensemble
of all data. Taking a closer look at the individual data sets of each group, the
agreement is however much worse. Especially the data of this study (stainless
steel and nickel), is not well described by their correlation. Instead, the
data presented from each study could be considered as being approximately
constant and independent of the spray Weber number.
In summary, there is no agreement between models and empirical correla-
tions published in the literature and the data of this study. All publications
maintain that the Leidenfrost temperature is significantly influenced by spray
properties. This assertion cannot be corroborated by the data from the present
study, in fact the present results indicates that the Leidenfrost temperature
is independent of any spray property.
6.3 Theoretical prediction of the Leidenfrost

point
The previous discussion of the experimental results has shown that the
Leidenfrost point is independent of the spray properties. In contrast it is
influenced by the substrate material.
This leads to the central question: What happens at the Leidenfrost point?
This question is first pursued by examining possible mechanisms for stabilizing
the vapor-liquid interface occurring in the film boiling condition. These
mechanisms are sketched in Fig. 6.8. One mechanism is the enhancement of
vaporization in the thinner regions of the vapor layer due to the higher heat
flux there, as shown schematically in Fig. 6.8a. Such a mechanism counteracts
the Rayleigh-Taylor instability. Thus, the transition to the nucleate boiling
regime is not possible, since the heat flux, and thus the evaporation rate,
become infinite as soon as the vapor layer thickness tends to zero.
Another potential reason for instability of the film boiling regime could be
the homogeneous nucleation which is initiated in the liquid at a certain temper-
ature (Avedisian, 1985), as illustrated in Fig. 6.8b. This assumption has been
used in models for the Leidenfrost temperature (Wang et al., 2019b; Bjornard
& Griffith, 1977; Wang et al., 2019a), in particular, for the explanation of
the dependence of the Leidenfrost temperature on the thermal effusivity of
94
6.3 Theoretical prediction of the Leidenfrost point
Figure 6.8: Possible mechanisms of film boiling: (a) stabilization of a liquid-vapor

interface by the vapor generation and (b) homogeneous nucleation at a certain
temperature. A single bubble (c) can also expand along the substrate due to
strong vaporization at the contact line (d). (as can be read in Tenzer et al.
(2020).)
the target material. The estimated homogeneous nucleation temperature is

Th = 306 ◦ C, computed from the data for the Leidenfrost temperature (Wang
et al., 2019a). Another estimation (Avedisian, 1985) for the homogeneous
nucleation temperature is Th = 202 ◦ C. However this assumption contradicts
the observations of film boiling below Th on substrates with very high thermal
effusivity, where the Leidenfrost point approaches the saturation temperature.
To confirm the appearance of low Leidenfrost temperatures, we performed a
simple experiment by placing a sessile water drop on a polished copper surface
at various substrate temperatures Ti . The exemplary outcomes are shown
in Fig. 6.9. In Fig. 6.9b at Ti = 138 ◦ C we can clearly identify Leidenfrost
phenomena: The drop is separated from the surface by a vapor layer. The
lifetime of the drop is very high. At a lower surface temperature of Ti = 124 ◦ C
in Fig. 6.9b the drop immediately starts to evaporate after first contact
with the surface. Extensive nucleation and formation of secondary droplets
is visible. We were able to identify Leidenfrost behavior at temperatures
Ti > 125 ◦ C. Therefore the measured Leidenfrost temperature for water drops
on copper targets is 124 ◦ C for a sessile drop or 134 ◦ C for impacting drops in
95
6 Leidenfrost point
(a) Ti = 124 ◦ C (b) Ti = 138 ◦ C
Figure 6.9: A sessile drop on a heated copper surface at different surface tempera-
tures Ti .
a spray (Ito et al., 1991), which are both lower than the limit for superheat of
distilled water. Moreover, neither of these hypotheses can explain the observed
significant influence of nano-structures, surface morphology or wettability of
the substrate on the Leidenfrost temperature, (Auliano et al., 2016; Kruse
et al., 2013; Romashevskiy et al., 2016; Kim et al., 2012; Takata et al., 2005).
Consider however the first stage of heat transfer, just after the contact of
the liquid with the solid substrate. In the nucleate boiling regime the contact
is followed by the emergence of numerous bubbles on the solid substrate as a
result of heterogeneous nucleation. The heat transfer can be described by the
heat conduction in two expanding thermal boundary layers in the liquid and
the substrate respectively. The thickness of the thermal boundary layer is
√
hw ∼ αw t where αw is the thermal diffusivity of the corresponding material,
liquid or solid. This phenomenon is depicted schematically in Fig. 6.8c.
The typical heat flux at the solid-liquid interface in the nucleate boiling
regime is determined by the thickness of the boundary layer in the solid
substrate and by the fact that the temperature of the vapor bubble interface
is close to the saturation temperature Tsat
w (Tw0 − Tsat )
q̇ ∼ √ , (6.1)
t
96
where Tw0 is the initial substrate temperature prior to contact with the liquid
and w is the thermal effusivity of the wall material. Equation (6.1) has been
recently validated by comparison with the experimental data for single drops
(Breitenbach et al., 2017a) and sprays (Tenzer et al., 2019b).
At some elevated wall temperature, characterized by intensive heating,
vaporization fronts appear instead of local bubbles and expand along the
substrate with a certain velocity. For a high enough velocity, the front leads
to the formation of a thin, near-wall vapor layer (as shown in the sketch in
Fig. 6.8d). The propagation velocity of the vaporization front, depending on
the heating rate and liquid properties, can be constant (Okuyama et al., 2006;
Avksentyuk & Ovchinnikov, 2000; Stutz & Simões-Moreira, 2013; Aktershev
& Ovchinnikov, 2011) or can grow exponentially (Staszel & Yarin, 2018) in
time.
We can roughly estimate the characteristic velocity Uv ∼ q̇/ρf L of the
vaporization front using Eq. (6.1)
w (Tw0 − Tsat )
Uv ∼ √ , (6.2)
ρf L t
where L is the latent heat of evaporation and ρf is the density of the liquid.
A similar approach has been successfully used to predict the velocities of
secondary drops in the thermal atomization regime (Breitenbach et al., 2018a;
Roisman et al., 2018).
The only available characteristic thickness in the liquid region associated
with the vaporization front is the thickness of the thermal boundary layer in
√
the fluid hf ∼ αf t, where αf is the thermal diffusivity of the fluid. Therefore,
the Weber and the capillary numbers based on the typical vaporization front
velocity (Eq. (6.2)) are singular at t → 0, which means the influence of surface
tension on the formation of the vaporization front is initially negligibly small.
The Reynolds number Rev = ρf hf Uv /νf , in contrast, is finite
√
w αf (Tw0 − Tsat )
Rev = . (6.3)
νf L
The viscous stresses therefore play an important role during the entire process
of the vaporization front formation. The characteristic superheat ∆T ∗ at which
the inertia and viscous terms are comparable can be determined assuming
97
6 Leidenfrost point
Figure 6.10: Substrate superheat at the Leidenfrost point, TL − Tsat , as a function

of ∆T ∗ for sessile drops. The data for different liquids, from this study and from
Baumeister & Simon (1973); Bernardin & Mudawar (1999); Emmerson (1975);
Wachters et al. (1966); Wang et al. (2019a), are listed in Table 6.1. (as can be
read in Tenzer et al. (2020).)
Rev = 1 in Eq. (6.3)
νf L
∆T ∗ = √ (6.4)
w αf
In Fig. 6.10 the experimental data for TL − Tsat of sessile drops of different
liquids on various substrates is shown as a function of ∆T ∗ . The liquid
properties for calculation of ∆T ∗ are taken at the saturation temperature.
For data taken from the literature, the substrate materials and their thermal
properties, type of experiment and the resulting Leidenfrost temperature are
summarized in Table 6.1. The thermal properties of the designated substrate
materials are taken from general literature and handbooks. Since these
properties are temperature dependent, they have been calculated at the
corresponding Leidenfrost temperature. In all cited studies the surface is
polished. Since the spray or drop impact occurs at the surface, the thermal
properties of the plating material is used in the case of a plated substrate,
(Yao & Choi, 1987; Choi & Yao, 1987; Yao & Cox, 2002; Bernardin et al.,
1997). Data from Wang et al. (2019a) were measured for different Weber
numbers: those with a large We were treated as spray data, whereas those
with a small We is treated as sessile drops.
98
Figure 6.11: Mean substrate superheat at the Leidenfrost point T L − Tsat as a

function of ∆T ∗ for spray cooling. The data for distilled water from this study
and from Hoogendoorn & den Hond (1974); Shoji et al. (1984); Ito et al. (1991);
Yao & Choi (1987); Choi & Yao (1987); Yao & Cox (2002); Bernardin et al.
(1997); Tran et al. (2012); Wang et al. (2019a) are listed in Table 6.1. (as can be
read in Tenzer et al. (2020).)
It seems that for a wide range of targets the value of TL − Tsat can be
predicted rather well as TL − Tsat = ∆T ∗ . For wall materials with very
low thermal effusivity the values for the Leidenfrost temperature deviate
significantly from the prediction. In both cases, for water and for ethanol, the
measured wall temperatures are above the critical temperature. In this case
additional physics has to be accounted for.
The corresponding data for spray impact or high Weber number drop
impact are shown in Fig. 6.11. The superheat associated with the Leidenfrost
temperature also follows a linear dependence on ∆T ∗ . The best fit of the
data is T L − Tsat = 1.51∆T ∗ . The Leidenfrost temperature in the impacting
drops increases, since the thickness of the thermal boundary in the liquid and
the propagation of the vaporization front are influenced by the flow in the
spreading drop (Roisman, 2010).
The fact that the value of TL − Tsat is comparable with ∆T ∗ for a range
of different liquids and substrates indicates that the initiation of the film
boiling can indeed be explained by the propagation of the vaporization front
along the substrate, assuming that the wall temperature does not exceed the
critical value for the studied liquid. Note that propagation of the vaporization
front is also influenced by the conditions at the moving contact line. This
99
6 Leidenfrost point
explains the sensitivity of the Leidenfrost temperature to the substrate micro-

and nano-morphology and wettability, identified in the existing literature. In
summary the theoretical prediction describes very well the dependence of the
Leidenfrost temperature on thermal fluid and substrate properties.
Table 6.1: Overview the thermal properties of the materials and the value of the
Leidenfrost temperature. (as can be read in Tenzer et al. (2020).)
Study Type of Fluid Substrate w , TL ,
experiment material Ws1/2 /m2 K ◦
C
This study Spray Water Stainless steel 8.8501e+03 342
1.4841
This study Spray Water Nickel Alloy 1.7892e+04 286
201
Hoogendoorn Spray Water Stainless steel 9.0193e+03 371
& den Hond 1.4301
(1974)
Shoji et al. Spray Water Nickel 1.8569e+04 321
(1984)
Ito et al. Spray Water Copper 3.6476e+04 134
(1991)
Yao & Choi Spray Water Copper with 1.7831e+04 222
(1987) chrome plating
Choi & Yao Spray Water Copper with 1.7831e+04 253
Yao & Cox Spray Water Copper with 1.7831e+04 266
Bernardin Drop Water Copper with 2.7701e+04 225
et al. (1997) chain gold plating
Tran et al. Drop Water Silicon wafer 9.7544e+03 480
(2012)
Wang et al. Drop Water FeCrAl 6.5664e+03 445
(2019a)
Wang et al. Drop Water Sintered SiC 1.5733e+04 350
(2019a)
Wang et al. Drop Water Zr-4 5.1511e+03 531
(2019a)
100
Table 6.1 continued

C
This study Sessile Water Copper 3.6476e+04 124
drop
Baumeister Sessile Water Pyrex 2.1030e+03 700
& Simon drop
(1973)
Baumeister Sessile Water Brass 1.9804e+04 233
& Simon drop
(1973)
Baumeister Sessile Water Gold 2.8174e+04 184
& Simon drop
(1973)
Baumeister Sessile Water Aluminium 2.0594e+04 210
& Simon drop
(1973)
Bernardin Sessile Water Aluminium 2.0594e+04 175
& Mudawar drop
(1999)
Emmerson Sessile Water Monel 9.5390e+03 316
(1975) drop
Bernardin Sessile Water Silver 3.2925e+04 157
& Mudawar drop
(1999)
Bernardin Sessile Water Graphite 1.1690e+04 310
& Mudawar drop
(1999)
Wachters Sessile Water Gold 2.8174e+04 170
et al. (1966) drop
Wang et al. Sessile Water FeCrAl 6.5664e+03 379
(2019a) drop
Wang et al. Sessile Water Sintered SiC 1.5733e+04 310
(2019a) drop
101
6 Leidenfrost point
Table 6.1 continued

C
Wang et al. Sessile Water Zr-4 5.1511e+03 415
(2019a) drop
Baumeister Sessile Ethanol Pyrex 2.1030e+03 360
& Simon drop
(1973)
Baumeister Sessile Ethanol Aluminium 2.0594e+04 155
& Simon drop
(1973)
Baumeister Sessile Ethanol Stainless steel 9.0193e+03 190
& Simon drop
(1973)
Bernardin Sessile Acetone Aluminium 2.0594e+04 134
& Mudawar drop
(1999)
Bernardin Sessile Benzene Aluminium 2.0594e+04 175
& Mudawar drop
(1999)
Bernardin Sessile FC-72 Aluminium 2.0594e+04 90
& Mudawar drop
(1999)
102
7 Conclusion and outlook
The present study deals with the investigation of transient spray cooling of a
heated thick target. Spray cooling is used in a great variety of applications like
cooling of high power electronics or quenching during steel making. Although
widely used, the overall physical picture is not complete and predictions for the
heat transfer are mainly of empirical nature. The drawback of these predictions
is the lack of universality and therefore limited range in which they produce
reliable results. To overcome this lack of knowledge several experiments of
transient spray cooling were conducted. The initially homogeneously heated
target was continuously cooled by various kinds of sprays. The local heat flux
and surface temperature were acquired and compared to visual observations of
the hydrodynamics during spray impact captured by a high speed observation
system.
In the film boiling regime the spray impact can be treated as single drop
impacts onto a dry surface and thus, can be regarded as a superposition of
single drop impacts. With decreasing temperature, and after the Leidenfrost
point has been passed, first small patches of water remain immediately after
the drop impact. Subsequently, in the transition boiling regime, the amount
of liquid remaining on the surface increases. Since the time spent in transition
boiling is very short, it is reasonable to simplify and replace this condition by
a jump of the heat flux. After passing the maximum heat flux (critical heat
flux) the entire surface is covered by a liquid film.
The mass flux of the impinging spray turns out to be the key factor for
achieving high cooling performance, whereas the mean droplet velocity and
diameter have only a smaller influence. Oblique spray cooling of a surface,
when the main spray direction is inclined towards the surface normal, results
in a lower heat transfer compared to the perpendicular impact. The decrease
turns out to be linearly dependent on the normal component of the spray
mean velocity. Furthermore, the heat transfer is influenced by the initial spray
water temperature. Cooling with hot water results in an decreased heat flux
103
and therefore lower cooling performance. Although this influence is present, it

is rather small compared to that of the mass flux. Another strong influencing
parameter is the material of the target substrate. Experiments on a nickel
surface exhibited much higher heat fluxes compared to those on a stainless
steel target. Thus, the heat flux is influenced by the thermal effusivity of the
substrate: A higher thermal effusivity results in a higher heat flux.
The analysis of the heat transfer during spray cooling leads to a predictive
model for the heat flux and surface temperature in the film boiling regime. The
model is based on the superposition of single drop impacts and the assumption
of a semi-infinite target. It accounts for different surface materials, spray
properties, spray temperature and initial wall temperatures. There is excellent
accordance between experimental results and theoretical predictions for all
experimental data, covering a large range of parameters.
In the nucleate boiling regime the heat flux dependence on time indicates
some kind of a universal character, which is nearly independent of the spray
parameters. Assuming an instantaneous jump of the surface temperature
to saturation temperature and a semi-infinite wall, results in a heat flux
dependence on time, which is only influenced by the substrate material and
the initial temperature. This upper bound of the heat flux in the nucleate
boiling regime agrees well with experimental data as long as the semi-infinite
assumption is valid. The upper bound is reached, even when spraying with
low mass fluxes. This leads to the conclusion that in many cases the material
is the limiting factor for heat flux and that further increasing the spray mass
flux is of little benefit in the nucleate boiling regime.
A consequence of this insight is that it may be beneficial to construe a

cooling strategy consisting of two stages: In the film boiling regime a dense
spray having a high mass flux is used to achieve a high cooling efficiency. When
the Leidenfrost point is passed and the nucleate boiling regime is reached,
the mass flux is then reduced, since the heat flux is limited by the substrate
and not by the spray. This would reduce the amount of spray fluid necessary
while maintaining a high cooling performance. This is of paramount interest
when special fluids are used instead of ordinary water. To what extent such a
strategy would be advantageous will depend on the specific properties of the
spray liquid, substrate material and temperature range of operation.
104
Measurements of the Leidenfrost point, identified as the minimum heat
flux, indicate an independence on the spray characteristics, which is in strong
contrast to findings in literature. Instead, the Leidenfrost point is solely
influenced by the material of the target surface. A theoretical prediction
based on the formation of a vaporization front, assuming that the inertial
and viscous effects are comparable, agrees well with the experimental data
for sessile drops and sprays.
The general outcomes and derived models from this study provide a com-
prehensive set of predictions and new knowledge for developing and improving
reliable cooling processes which can be adapted to the special needs of each
individual cooling application.
The fact that the heat flux at the first instants of spray impact above the
Leidenfrost point is extremely high suggests another novel cooling strategy
based on the results of this study. Since the thermal boundary layer inside
the substrate has a significant influence on the cooling performance, it is
promising to use pulsating sprays for increasing the cooling efficiency, i.e. the
spray is continuously switched on and off. During the off-phase heat from the
lower portions of the substrate can flow to the top and decrease the thermal
boundary layer thickness. When the spray is again switched on, the spray
impact takes place on a substrate which has a thin boundary layer, leading
to an initial higher heat flux. Depending on the exact material properties
and temperature range, this may lead to an increased overall heat flux while
decreasing the required amount of spray fluid. The model developed in this
study for the film boiling regime accounts for the very high heat flux at the
initial phase of spray impact and can therefore be used to design and assess
such spray strategies.
Nevertheless, a number of topics should be addressed in future investiga-
tions: Due to the short timescales in the transition boiling regime it seems
reasonable to neglect this regime in some cases. But in others this simplifi-
cation leads to large errors. Since the physics of the formation and growth
of the liquid patches is not yet completely understood, no predictive models
exist. Furthermore, there must be some lower bound under which no complete
wetting of the surface can accrue. This boundary has to be identified and
described. Therefore future studies should deal with the transition boiling
regime.
105
Although the upper bound for the heat flux in the nucleate boiling regime
agrees well with the experimental data in this study, it is in fact an upper bound
which can only be valid if certain conditions are met. These circumstances
must be investigated so that models can be developed which describe when
to used this upper boundary. Additionally, a prediction for the heat flux, if
the upper boundary is not reached, has to be developed.
The theory for the Leidenfrost temperature is only valid for temperatures
below the critical temperature of the fluid, which is sufficient for many
industrial applications. Future investigations should find explanations for the
physical processes to achieve a prediction covering also the cases when the
temperature exceeds the critical value. Furthermore, the experimental data
found in literature for different fluids is often quite old and rather sparse.
Acquisition of an accurate and complete set of Leidenfrost temperatures with
dependence on various fluids and substrates could further enhance research
in this field.
106
Nomenclature
Capital roman letters

A m2 surface
A(τ ) ◦
C/s function
B - incomplete beta function
C1 , C2 - empirical constants
D m drop diameter
D10 m mean drop diameter
L J/kg heat of evaporation
Q J heat
Re - Reynolds number
ReS - spray Reynolds number
Rev - Reynolds number associated with the
vaporization front
S s−1/2 parameter
T ◦
C temperature
Tf0 ◦
C initial fluid temperature
Th ◦
C homogeneous nucleation temperature
Ti ◦
C surface temperature
TiL ◦
C surface temperature at Leidenfrost point
TL ◦
C Leidenfrost temperature
TL ◦
C mean Leidenfrost temperature
TLd ◦
C dynamic Leidenfrost temperature
TLs ◦
C static Leidenfrost temperature
Tsat ◦
C saturation temperature
Tw0 ◦
C initial wall temperature
∆T ◦
C temperature difference
∆TL ◦
C temperature difference
∆T ∗ ◦
C characteristic superheat
T s predicted parameter
107
U m/s mean droplet velocity

Uβ m/s normal component of the droplet velocity
during oblique spray impact at β
Uv m/s velocity of the vaporization front
We - Weber number
WeS - spray Weber number
Lowercase roman letters
ai - series, i ∈ [1, 2, ..., ∞]
b - dimensionless parameter
cp J/kgK specific heat at constant pressure
dT m diameter of target
hT m height of target
hbubble m thermal boundary layer in substrate due to
vaporization bubble
hi m boundary layer thickness in the phase i, i ∈
[f, w]
i - control variable
k - fitting paramter
m kg mass
ṁ kg/m2 s local mass flux
q̇ W/m2 local heat flux
q̇L W/m2 local heat flux at Leidenfrost point
q̇S W/m2 predicted local heat flux of entire cooling
experiment
r m radial coordinate of the target
t s time
tCHF s time at critical heat flux
tL s time at Leidenfrost point
tbubble s characteristic time of bubble formation
w - dimensionless parameter
z m wall-normal coordinate of the target
Capital greek letters
Θ - dimensionless surface temperature
108
Φ - dimensionless heat flux
Lowercase greek letters
α m2 /s thermal diffusivity
αi m2 /s thermal diffusivity of the phase i, i ∈ [f, w]
β ◦
spray impact angle
J/s1/2 m2 K thermal effusivity
i J/s1/2 m2 K thermal effusivity of the phase i, i ∈ [f, w]
ζ - dimensionless paramter
η kg/ms dynamic viscosity
λ W/mK thermal conductivity
λv W/mK thermal conductivity of the phase v
ν m2 /s kinematic viscosity
νf m2 /s kinematic viscosity of phase f
ξ - dimensionless time
ξL - dimensionless time at Leidenfrost point
ρ kg/m3 density
ρf kg/m3 density of phase f
σ kg/s2 surface tension
σf kg/s2 surface tension of phase f
τ s parameter
χ - fitting parameter
Abbreviations
CHF critical heat flux
DPSS diode pumped solid state
f fluid
film film boiling regime
HS high speed
IHCP inverse heat conduction problem
L Leidenfrost
min minimum
nucleate nucleate boiling regime
PC personal computer
PD phase Doppler
PDA phase Doppler analyzer
109
v vapor
w wall
110
Bibliography
Abu-Zaid, M. (2003). An experimental study of the evaporation characteristics
of emulsified liquid droplets. Heat and Mass Transfer, 40(9):737–741.
Aktershev, S. P. & Ovchinnikov, V. V. (2011). Modeling of the vaporization

front on a heater surface. Journal of Engineering Thermophysics, 20(1):77–
88.
Al-Ahmadi, H. M. & Yao, S. C. (2008). Spray cooling of high temperature

metals using high mass flux industrial nozzles. Experimental Heat Transfer,
21(1):38–54.
Albrecht, H.-E., Borys, M., Damaschke, N., & Tropea, C. (2013). Laser
Doppler and Phase Doppler Measurement Techniques. Springer, Berlin
Heidelberg.
Araneo, L. & Tropea, C. (2000). Improving phase Doppler measurements in

a diesel spray. In SAE Technical Paper 2000-01-2047.
Auliano, M., Fernandino, M., Zhang, P., & Dorao, C. A. (2016). The Leiden-
frost phenomenon on silicon nanowires. In ASME 2016 14th International
Conference on Nanochannels, Microchannels, and Minichannels, ICNMM
2016. American Society of Mechanical Engineers.
Aursand, E., Davis, S. H., & Ytrehus, T. (2018). Thermocapillary instability

as a mechanism for film boiling collapse. Journal of Fluid Mechanics,
852:283–312.
Avedisian, C. T. (1985). The homogeneous nucleation limits of liquids. Journal

of Physical and Chemical Reference Data, 14(3):695–729.
Avksentyuk, B. P. & Ovchinnikov, V. V. (2000). Shape of the vapor bubble

upon explosive boiling. Journal of Applied Mechanics and Technical Physics,
41(2):317–318.
111
Bibliography
Bar-Cohen, A., Arik, M., & Ohadi, M. (2006). Direct liquid cooling of
high flux micro and nano electronic components. Proceedings of the IEEE,
94(8):1549–1570.
Baumeister, K. J. & Simon, F. F. (1973). Leidenfrost temperature—Its
correlation for lipid metals, cryogens, hydrocarbons, and water. Journal of
Heat Transfer, 95(2):166–173.
Bernardin, J. D. & Mudawar, I. (1996). An experimental investigation into
the relationship between temperature-time history and surface roughness in
the spray quenching of aluminum parts. Journal of Engineering Materials
and Technology, Transactions of the ASME, 118(1):127–134.
Bernardin, J. D. & Mudawar, I. (1999). The Leidenfrost point: experimen-
tal study and assessment of existing models. Journal of Heat Transfer,
121(4):894.
Bernardin, J. D. & Mudawar, I. (2004). A Leidenfrost point model for
impinging droplets and sprays. Journal of Heat Transfer, 126(2):272.
Bernardin, J. D., Stebbins, C. J., & Mudawar, I. (1997). Mapping of impact
and heat transfer regimes of water drops impinging on a polished surface.
International Journal of Heat and Mass Transfer, 40(2):247–267.
Bertola, V. (2015). An impact regime map for water drops impacting on heated
surfaces. International Journal of Heat and Mass Transfer, 85:430–437.
Biance, A.-L., Clanet, C., & Quéré, D. (2003). Leidenfrost drops. Physics of
Fluids, 15(6):1632.
Bjornard, T. A. & Griffith, P. (1977). PWR blowdown heat transfer. In
Winter annual meeting of the American Society of Mechanical Engineers,
pages 17–41, Atlanta, GA, USA.
Breitenbach, J., Kissing, J., Roisman, I. V., & Tropea, C. (2018a). Character-
ization of secondary droplets during thermal atomization regime. Experi-
mental Thermal and Fluid Science, 98:516–522.
Breitenbach, J., Roisman, I. V., & Tropea, C. (2017a). Drop collision with a
hot, dry solid substrate: Heat transfer during nucleate boiling. Physical
Review Fluids, 2(7):074301.
112
Bibliography
Breitenbach, J., Roisman, I. V., & Tropea, C. (2017b). Heat transfer in the
film boiling regime: Single drop impact and spray cooling. International
Journal of Heat and Mass Transfer, 110:34–42.
Breitenbach, J., Roisman, I. V., & Tropea, C. (2018b). From drop impact
physics to spray cooling models: a critical review. Experiments in Fluids,
59(3):55.
Cai, C., Mudawar, I., Liu, H., & Si, C. (2020). Theoretical Leidenfrost point
(LFP) model for sessile droplet. International Journal of Heat and Mass
Transfer, 146:118802.
Carey, V. P. (2018). Liquid Vapor Phase Change Phenomena: An Introduction

to the Thermophysics of Vaporization and Condensation Processes in Heat
Transfer Equipment. Taylor and Francis, New York, 2. ed edition.
Castanet, G., Caballina, O., & Lemoine, F. (2015). Drop spreading at the
impact in the Leidenfrost boiling. Physics of Fluids, 27(6).
Cebo-Rudnicka, A., Malinowski, Z., & Buczek, A. (2016). The influence

of selected parameters of spray cooling and thermal conductivity on heat
transfer coefficient. International Journal of Thermal Sciences, 110:52–64.
Celata, G. P., Cumo, M., Mariani, A., & Zummo, G. (2006). Visualization
of the impact of water drops on a hot surface: effect of drop velocity and
surface inclination. Heat and Mass Transfer, 42(10):885–890.
Chandra, S. & Avedisian, C. T. (1991). On the collision of a droplet with a

solid surface. Proceedings of the Royal Society A: Mathematical, Physical
and Engineering Sciences, 432(1884):13–41.
Chaze, W., Caballina, O., Castanet, G., Pierson, J. F., Lemoine, F., & Maillet,
D. (2019). Heat flux reconstruction by inversion of experimental infrared
temperature measurements – Application to the impact of a droplet in
the film boiling regime. International Journal of Heat and Mass Transfer,
128:469–478.
Chen, R.-H. H., Chow, L. C., & Navedo, J. E. (2002). Effects of spray charac-
teristics on critical heat flux in subcooled water spray cooling. International
Journal of Heat and Mass Transfer, 45(19):4033–4043.
113
Bibliography
Chen, S. J. & Tseng, A. A. (1992). Spray and jet cooling in steel rolling.
International Journal of Heat and Fluid Flow, 13(4):358–369.
Cheng, W. L., Zhang, W. W., Chen, H., & Hu, L. (2016). Spray cooling
and flash evaporation cooling: The current development and application.
Renewable and Sustainable Energy Reviews, 55:614–628.
Choi, K. J. & Yao, S. C. (1987). Mechanisms of film boiling heat transfer of
normally impacting spray. International Journal of Heat and Mass Transfer,
30(2):311–318.
Cox, T. L. & Yao, S. C. (1999). Heat transfer of sprays of large water
drops impacting on high temperature surfaces. Journal of Heat Transfer,
121(2):446–450.
Ebadian, M. A. & Lin, C. X. (2011). A review of high-heat-flux heat removal
technologies. Journal of Heat Transfer, 133(11):110801.
Emmerson, G. S. (1975). The effect of pressure and surface material on the
Leidenfrost point of discrete drops of water. International Journal of Heat
and Mass Transfer, 18(3):381–386.
Estes, K. A. & Mudawar, I. (1995). Correlation of sauter mean diameter and
critical heat flux for spray cooling of small surfaces. International Journal
of Heat and Mass Transfer, 38(16):2985–2996.
Gottfried, B. S., Lee, C. J., & Bell, K. J. (1966). The Leidenfrost phenomenon:
film boiling of liquid droplets on a flat plate. International Journal of Heat
and Mass Transfer, 9(11):1167–1188.
Hiroyasu, H., Kadota, T., & Senda, T. (1974). Droplet evaporation on a
hot surface in pressurized and heated ambient gas. Bulletin of JSME,
17(110):1081–1087.
Hofmann, J. (2019). Experimental investigation of the Leidenfrost point during
spray- and single drop impact. Master’s thesis, Technische Universität
Darmstadt, Darmstadt, Germany.
Hoogendoorn, C. J. & den Hond, R. (1974). Leidenfrost temperature and
heat transfer coefficients for water sprays impinging on a hot surface. In
IFTH International Heat Transfer Conference, volume 4, pages 135–138.
114
Bibliography
Itaru, M. & Kunihide, M. (1978). Heat transfer characteristics of evaporation

of a liquid droplet on heated surfaces. International Journal of Heat and
Mass Transfer, 21(5):605–613.
Ito, T., Takata, Y., Mousa, M. M. M., & Yoshikai, H. (1991). Studies on the
water cooling of hot surfaces (experiment of spray cooling). Memoirs of the
Faculty of Engineering, Kyushu University, 51(2):119–144.
Jerome, B. P. (1960). Transition boiling heat transfer from a horizontal

surface. Technical report, Massachusetts Institute of Technology, Heat
Transfer Laboratory no. 17, Cambridge, Mass.
Kakac, S. & Bon, B. (2008). A review of two-phase flow dynamic instabilities

in tube boiling systems. International Journal of Heat and Mass Transfer,
51:399–433.
Kaupe, T. (2017). Design and setup of a spray patternator for measuring

the mass flux density. Master’s thesis, Technische Universität Darmstadt,
Darmstadt, Germany.
Kim, H., Truong, B., Buongiorno, J., & Hu, L. W. (2012). Effects of
micro/nano-scale surface characteristics on the Leidenfrost point tempera-
ture of water. Journal of Thermal Science and Technology, 7(3):453–462.
Kim, J. (2007). Spray cooling heat transfer: The state of the art. International
Journal of Heat and Fluid Flow, 28(4):753–767.
Klinzing, W. P., Rozzi, J. C., & Mudawar, I. (1992). Film and transition
boiling correlations for quenching of hot surfaces with water sprays. Journal
of Heat Treating, 9(2):91–103.
Kruse, C., Anderson, T., Wilson, C., Zuhlke, C., Alexander, D., Gogos, G., &
Ndao, S. (2013). Extraordinary shifts of the Leidenfrost temperature from
multiscale micro/nanostructured surfaces. Langmuir, 29(31):9798–9806.
Labergue, A., Gradeck, M., & Lemoine, F. (2015). Comparative study

of the cooling of a hot temperature surface using sprays and liquid jets.
International Journal of Heat and Mass Transfer, 81:889–900.
Lefebvre, A. (1988). Atomization and Sprays. CRC Press, Boca Raton.
115
Bibliography
Leidenfrost, J. G. (1966). On the fixation of water in diverse fire. International

Liang, G. & Mudawar, I. (2017a). Review of drop impact on heated walls.

International Journal of Heat and Mass Transfer, 106:103–126.
Liang, G. & Mudawar, I. (2017b). Review of spray cooling – Part 1: Single-

phase and nucleate boiling regimes, and critical heat flux. International
Journal of Heat and Mass Transfer, 115(September):1174–1205.
Liang, G. & Mudawar, I. (2017c). Review of spray cooling – Part 2: High

temperature boiling regimes and quenching applications. International
Journal of Heat and Mass Transfer, 115:1206–1222.
Mizikar, E. A. (1970). Spray cooling investigation for continuous casting of

billets and blooms. Iron and Steel Engineer, 47(6):53–60.
Monde, M., Arima, H., Liu, W., Mitutake, Y., & Hammad, J. A. (2003). An
analytical solution for two-dimensional inverse heat conduction problems
using Laplace transform. International Journal of Heat and Mass Transfer,
46(12):2135–2148.
Mudawar, I. (2001). Assessment of high-heat-flux thermal management

schemes. IEEE Transactions on Components and Packaging Technologies,
24(2):122–141.
Mudawar, I. & Deiters, T. A. (1994). A universal approach to predicting

temperature response of metallic parts to spray quenching. International
Nimi, M., Pri, O., Ulivanju, K., Raudensky, M., Hnizdil, M., Hwang, J. Y.,
Lee, S. H., & Kim, S. Y. (2012). Influence of the water temperature on
the cooling intensity of mist nozzles in continuous casting. Materiali in
tehnologije, 46(3):311–315.
Nižetić, S., Čoko, D., Yadav, A., & Grubišić-Čabo, F. (2016). Water spray cool-
ing technique applied on a photovoltaic panel: The performance response.
Energy Conversion and Management, 108:287–296.
116
Bibliography
Nukiyama, S. (1966). The maximum and minimum values of the heat Q

transmitted from metal to boiling water under atmospheric pressure. Inter-
national Journal of Heat and Mass Transfer, 9(12):1419–1433.
Okuyama, K., Kim, J. H., Mori, S., & Iida, Y. (2006). Boiling propagation of
water on a smooth film heater surface. International Journal of Heat and
Mass Transfer, 49(13-14):2207–2214.
Özışık, M. N. (1980). Heat conduction. John Wiley & Sons, New York,
Chichester, Brisbane, Toronto.
Pham, M. K. (2018). Study of the influence of the fluid temperature on the
heat transfer and hydrodynamics during spray cooling. Bachelor’s thesis,
Technische Universität Darmstadt, Darmstadt, Germany.
Pola, A., Gelfi, M., & La Vecchia, G. M. (2013). Simulation and validation of
spray quenching applied to heavy forgings. Journal of Materials Processing
Technology, 213(12):2247–2253.
Puschmann, F. & Specht, E. (2004). Transient measurement of heat transfer
in metal quenching with atomized sprays. Experimental Thermal and Fluid
Science, 28(6):607–615.
Quéré, D. (2013). Leidenfrost dynamics. Annual Review of Fluid Mechanics,
45(1):197–215.
Roisman, I. V. (2010). Fast forced liquid film spreading on a substrate:
Flow, heat transfer and phase transition. Journal of Fluid Mechanics,
656:189–204.
Roisman, I. V., Breitenbach, J., & Tropea, C. (2018). Thermal atomisation of
a liquid drop after impact onto a hot substrate. Journal of Fluid Mechanics,
842:87–101.
Romashevskiy, S. A., Agranat, M. B., & Dmitriev, A. S. (2016). Thermal
training of functional surfaces fabricated with femtosecond laser pulses.
High Temperature, 54(3):461–465.
Sargunanathan, S., Elango, A., & Mohideen, S. T. (2016). Performance
enhancement of solar photovoltaic cells using effective cooling methods: A
review. Renewable and Sustainable Energy Reviews, 64:382–393.
117
Bibliography
Shoji, M., Wakunaga, T., & Kodama, K. (1984). Heat transfer from a heated
surface to impinging subcooled droplets : Heat transfer characteristics
in non-wetting region. Transactions of the Japan Society of Mechanical
Engineers Series B, 50(451):716–723.
Sienski, K., Eden, R., & Schaefer, D. (1996). 3-D electronic interconnect pack-
aging. In IEEE Aerospace Applications Conference Proceedings, volume 1,
pages 363–373. IEEE.
Sozbir, N., Chang, Y. W., & Yao, S. C. (2003). Heat transfer of impacting
water mist on high temperature metal surfaces. Journal of Heat Transfer,
125(1):70.
Sozbir, N., Yigit, C., Issa, R. J., Yao, S.-C., Guven, H. R., & Ozcelebi,
S. (2010). Multiphase spray cooling of steel plates near the Leidenfrost
temperature-experimental studies and numerical modeling. Atomization
and Sprays, 20(5):387–405.
Spiegler, P., Hopenfeld, J., Silberberg, M., Bumpus, C. F., & Norman, A.
(1963). Onset of stable film boiling and the foam limit. International
Staat, H. J. J., Tran, T., Geerdink, B., Riboux, G., Sun, C., Gordillo, J. M.,
& Lohse, D. (2015). Phase diagram for droplet impact on superheated
surfaces. Journal of Fluid Mechanics, 779:R3.
Staszel, C. & Yarin, A. L. (2018). Exponential vaporization fronts and critical

heat flux in pool boiling. International Communications in Heat and Mass
Transfer, 98:171–176.
Stutz, B. & Simões-Moreira, J. R. (2013). Onset of boiling and propagating

mechanisms in a highly superheated liquid - The role of evaporation waves.
International Journal of Heat and Mass Transfer, 56(1-2):683–693.
Takata, Y., Hidaka, S., Cao, J. M., Nakamura, T., Yamamoto, H., Masuda,
M., & Ito, T. (2005). Effect of surface wettability on boiling and evaporation.
Energy, 30:209–220.
118
Bibliography
Tartarini, P., Lorenzini, G., & Randi, M. R. (1999). Experimental study of

water droplet boiling on hot, non-porous surfaces. Heat and Mass Transfer,
34(6):437–447.
Tenzer, F. M., Hofmann, J., Roisman, I. V., & Tropea, C. (2020). Leiden-
frost temperature in sprays: role of the substrate and liquid properties.
Manuscript submitted for publication, (arXiv:2001.05426).
Tenzer, F. M., Pham, M. K., Roisman, I. V., & Tropea, C. (2019a). Fast spray
cooling: Effect of the liquid temperature. In HEFAT 2019, 14th Interna-
tional Conference on Heat Transfer, Fluid Mechanics and Thermodynamics,
pages 706–710, Wicklow, Ireland.
Tenzer, F. M., Roisman, I. V., & Tropea, C. (2018). Spray cooling in the
nucleate boiling regime: Hydrodynamics and heat transfer. In ICLASS
2018, 14th Triennial International Conference on Liquid Atomization and
Spray Systems, page 123, Chicago, IL.
Tenzer, F. M., Roisman, I. V., & Tropea, C. (2019b). Fast transient spray
cooling of a hot thick target. Journal of Fluid Mechanics, 881:84–103.
Testa, P. & Nicotra, L. (2009). Influence of pressure on the Leidenfrost

temperature and on extracted heat fluxes in the transient mode and low
pressure. Journal of Heat Transfer, 108(4):916.
Tran, T., Staat, H. J. J., Prosperetti, A., Sun, C., & Lohse, D. (2012). Drop
impact on superheated surfaces. Physical Review Letters, 108(3):1–5.
Tropea, C. (2011). Optical particle characterization in flows. Annual Review

of Fluid Mechanics, 43(1):399–426.
Tropea, C., Xu, T.-H., Onofri, F., Géhan, G., Haugen, P., & Stieglmeier,
M. (1996). Dual-mode phase-Doppler anemometer. Particle & Particle
Systems Characterization, 13(2):165–170.
Wachters, L. H., Bonne, H., & van Nouhuis, H. J. (1966). The heat transfer
from a hot horizontal plate to sessile water drops in the spherodial state.
Chemical Engineering Science, 21(10):923–936.
119
Bibliography
Wang, Z., Qu, W., Xiong, J., Zhong, M., & Yang, Y. (2019a). Investigation on
effect of surface properties on droplet impact cooling of cladding surfaces.
Nuclear Engineering and Technology.
Wang, Z. F., Xiong, J., Yao, W., Qu, W., & Yang, Y. (2019b). Experimental
investigation on the Leidenfrost phenomenon of droplet impact on heated
silicon carbide surfaces. International Journal of Heat and Mass Transfer,
128:1206–1217.
Wendelstorf, J., Spitzer, K. H., & Wendelstorf, R. (2008). Spray water cooling
heat transfer at high temperatures and liquid mass fluxes. International
Journal of Heat and Mass Transfer, 51(19-20):4902–4910.
Wolter, S. (2018). Spray characterization of one- and two-component nozzles

for the investigation of spray cooling of hot surfaces. Master’s thesis,
Technische Universität Darmstadt, Darmstadt, Germany.
Woodfield, P. L., Monde, M., & Mitsutake, Y. (2006). Improved analytical

solution for inverse heat conduction problems on thermally thick and
semi-infinite solids. International Journal of Heat and Mass Transfer,
49(17-18):2864–2876.
Xu, F. & Gadala, M. S. (2006). Heat transfer behavior in the impingement

zone under circular water jet. International Journal of Heat and Mass
Transfer, 49(21-22):3785–3799.
Yang, J., Chow, L. C., & Pais, M. R. (1996). Nucleate boiling heat transfer
in spray cooling. Journal of Heat Transfer, 118(3):668.
Yao, S. C. & Cai, K. Y. (1988). The dynamics and Leidenfrost temperature

of drops impacting on a hot surface at small angles. Experimental Thermal
and Fluid Science, 1(4):363–371.
Yao, S. C. & Choi, K. J. (1987). Heat transfer experiments of mono-dispersed

vertically impacting sprays. International Journal of Multiphase Flow,
13(5):639–648.
Yao, S. C. & Cox, T. L. (2002). A general heat transfer correlation for

impacting water sprays on high-temperature surfaces. Experimental Heat
Transfer, 15(4):207–219.
120
Bibliography
Yarin, A. L., Roisman, I. V., & Tropea, C. (2017). Collision phenomena in

liquids and solids. Cambridge University Press, Cambridge.
Zuber, N. (1958). On the stability of boiling heat transfer. Transactions of

the American Society of Mechanical Engineers, 80.
121
122
List of Figures
1.1 Heat transfer coefficient of different cooling technologies. . . . 2

1.2 Exemplary temporal evolution of the temperature 1 mm below
the surface of a workpiece during a hot forging process. . . . 3
1.3 Exemplary evolution of the heat flux as a function of the surface
temperature during spray cooling. . . . . . . . . . . . . . . . 5
1.4 Dependence of the Leidenfrost temperature on the spray Weber
number. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1 Schematic of experimental setup. . . . . . . . . . . . . . . . . 14

2.2 Detailed sketch of the atomizer and water supply system used
for spray generation. . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 Mean droplet diameter and velocity as a function of radial
position. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 Mass flux as a function of the radial position. . . . . . . . . . 20
2.5 Droplet velocity versus diameter for various atomizer in the
center region below the atomizer. . . . . . . . . . . . . . . . . 21
2.6 Heated target equipped with watertight housing and ventilation
slit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.7 Heated target and additional components inside the watertight
housing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.8 Sketch of the sectional view of the heated target showing the
thermocouple positions. Dimensions are in mm. . . . . . . . . 27
2.9 Sketch of the thermocouples having an open measuring tip
which is aligned with the shield. . . . . . . . . . . . . . . . . 28
2.10 Field of view for configuration with Tamron 180 mm macro
optical lens. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.11 Field of view for configuration with Questar QM-100 long
distance microscope objective lens. . . . . . . . . . . . . . . . 29
123
List of Figures
3.1 Calculated temperature T inside the substrate as a function of

the depth z for different times t. . . . . . . . . . . . . . . . . 32
3.2 Typical temporal evolution of the surface temperature Ti (t)
and instantaneous local heat flux q̇(t). . . . . . . . . . . . . . 33
3.3 Typical evolution of the instantaneous local heat flux q̇(t) as a
function of the surface temperature Ti (t). . . . . . . . . . . . 33
3.4 Spray cooling regimes at different surface temperatures Ti (t)
around the Leidenfrost point. . . . . . . . . . . . . . . . . . . 34
3.5 Phenomena of spray impact regimes at different surface tem-
peratures Ti (t). . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.6 Position of the wetting front that follows the Leidenfrost point
at different instants of time t. . . . . . . . . . . . . . . . . . . 39
3.7 Sketch of the hydrodynamics occurring during spray impact. 41
3.8 Influence of mass flux ṁ on the heat transfer. . . . . . . . . . 43
3.9 Influence of the mean droplet velocity on the heat transfer. . 44
3.10 Influence of different spray produced by a pneumatic atomizer. 45
3.11 Sketch of oblique spray impact. . . . . . . . . . . . . . . . . . 46
3.12 Influence of the spray impact angle β. . . . . . . . . . . . . . 47
3.13 Scaled heat flux in the film boiling regime q̇β /q̇ as a function
of the scaled impact velocity Uβ /U for different spray impact
angles β. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.14 Heat flux dependence on the surface temperature for different
spray water temperatures. . . . . . . . . . . . . . . . . . . . . 48
spray water temperatures. Only film boiling regime, Leidenfrost
point and beginning of transition boiling regime are shown. . 50
spray water temperatures. The inserts show the conditions at
the surface at the corresponding surface temperature. . . . . 51
3.17 Heat flux as a function of the initial spray water temperature.
The data and spray parameters are equal to Fig. 3.15. . . . . 53
3.18 Influence of the initial wall temperature Tw0 for a small mass
flux. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.19 Influence of the initial wall temperature Tw0 for a large mass
flux of ṁ = 9.3 kg/m2 s. . . . . . . . . . . . . . . . . . . . . . 55
124
List of Figures
3.20 Influence of the substrate material on the heat transfer for a

constant rather sparse spray. . . . . . . . . . . . . . . . . . . 56
3.21 Influence of the substrate material on the heat transfer for a
constant rather dense spray. . . . . . . . . . . . . . . . . . . . 57
3.22 Influence of the surface roughness on the heat transfer, for a
small mass flux of 1 kg/m2 s. . . . . . . . . . . . . . . . . . . . 58
3.23 Influence of the surface roughness on the heat transfer, for a
small mass flux of 3.1 kg/m2 s. . . . . . . . . . . . . . . . . . . 60
4.1 Theoretically predicted evolution of the dimensionless quanti-

ties as functions of the dimensionless time ξ in the film boiling
regime. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2 Evolution of the dimensionless quantities for the stainless steel
target as a function of the dimensionless time ξ. . . . . . . . . 66
4.3 Evolution of the dimensionless quantities for the hot-work tool
steel target as a function of the dimensionless time ξ. . . . . . 68
4.4 Evolution of the dimensionless quantities for the nickel target
as a function of the dimensionless time ξ. . . . . . . . . . . . 69
4.5 Evolution of the dimensionless quantities for the nickel tar-
get spayed with a pneumatic atomizer as a function of the
dimensionless time ξ. . . . . . . . . . . . . . . . . . . . . . . . 70
4.6 Exemplary comparison between one experiment and the theo-
retical prediction in the film boiling regime. . . . . . . . . . . 72
4.7 Heat flux as a function of the initial spray water temperature. 73
4.8 Evolution of the heat flux as a function of time for an exemplary
experiment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.1 Scaling of the heat flux plotted as a function of time for the
stainless steel target. . . . . . . . . . . . . . . . . . . . . . . . 80
target built of hot-work tool steel. . . . . . . . . . . . . . . . 81
nickel target. . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
retical prediction in the nucleate boiling regime. . . . . . . . . 83
125
List of Figures

retical prediction of the entire cooling process. . . . . . . . . . 85
6.1 Determination of the Leidenfrost point TL as the temperature

at the minimum heat flux qmin . . . . . . . . . . . . . . . . . . 88
6.2 Leidenfrost temperature TL as a function of the mass flux ṁ. 89
6.3 Dependence of the Leidenfrost temperature TL on the spray
properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6.4 Dependence of the Leidenfrost temperature TL on dimensionless
numbers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6.5 Comparison between own data and the correlation proposed
by Al-Ahmadi & Yao (2008) (Eq. (1.2)). . . . . . . . . . . . . 92
6.6 Comparison between own data and the correlation proposed
by Bernardin & Mudawar (2004) (Eq. (1.6)). . . . . . . . . . 93
6.7 Comparison between own data, those of other researchers and
the correlation proposed by Yao & Cox (2002). . . . . . . . . 93
6.8 Possible mechanisms of film boiling. . . . . . . . . . . . . . . 95
6.9 A sessile drop on a heated copper surface at different surface
temperatures Ti . . . . . . . . . . . . . . . . . . . . . . . . . . 96
6.10 Substrate superheat at the Leidenfrost point, TL − Tsat , as a
function of ∆T ∗ for sessile drops. . . . . . . . . . . . . . . . . 98
6.11 Mean substrate superheat at the Leidenfrost point T L − Tsat
as a function of ∆T ∗ for spray cooling. . . . . . . . . . . . . . 99
126
List of Tables
2.1 Overview of atomizers used for the experiments and their
operational parameters. . . . . . . . . . . . . . . . . . . . . . 16
2.2 Parameters of the phase Doppler measurement system. . . . . 17
2.3 Overview of targets and material parameters used of cooling
experiments. . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.1 Typical conditions of spray cooling experiments of this study. 64
6.1 Overview the thermal properties of the materials and the value
of the Leidenfrost temperature. . . . . . . . . . . . . . . . . . 100
127

Dissertation Tenzer

Uploaded by

Copyright:

Available Formats

Dissertation Tenzer

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dissertation Tenzer

Uploaded by

Copyright:

Available Formats

Heat transfer during transient spray cooling:

An experimental and analytical study

Vom Fachbereich Maschinenbau

Fabian Michael Tenzer, M.Sc.

Berichterstatter: Prof. Dr.-Ing. Cameron Tropea

Bitte zitieren Sie dieses Dokument als:

Dieses Dokument wird bereitgestellt von TU Prints, E-Publishing-Service

Veröffentlicht unter CC BY-SA 4.0 International:

Spray cooling features a very high, homogeneously distributed cooling per-

In dieser Arbeit wird die transiente Sprühkühlung eines heißen, dicken

2 Experimental setup and methods 13

3 Typical results of transient spray cooling 31

3.3.5 Wall thermal properties . . . . . . . . . . . . . . . . . 55

4 Heat transfer in film boiling regime 61

5 Heat transfer in nucleate boiling regime 77

7 Conclusion and outlook 103

List of Figures 123

List of Tables 127

Figure 1.1: Heat transfer coefficient of different cooling technologies. (Adapted

Figure 1.2: Exemplary temporal evolution of the temperature 1 mm below the

To gain a first impression of the temperatures and timescales during hot

In Fig. 1.2 the temperature reading from this experiment is plotted as a

1.2 Fundamentals of spray cooling

The Leidenfrost point The following section gives an overview of the

TLd = TLs + C1 WeC2 (1.1)

TL = 536.8 ṁ0.116 . (1.2)

Here TL is the Leidenfrost temperature and ṁ the mass flux. Additionally,

Reynolds and spray Weber numbers are defined as

where D, η, ρ and σ are a mean diameter, dynamic viscosity, density and

The correlation as well as the underlying data is shown in Fig. 1.4.

TL = 162 + 24.3U 0.64 . (1.6)

The model is applied to experimental data of Klinzing et al. (1992) and

Influence of the initial spray water temperature Spray cooling in-

1.3 Objective and outline of this thesis

Spray cooling is used in many different industrial applications because of

In Chapter 5 a predictive model for the heat transfer in the nucleate

2.1 Overview of the setup

2.2 Spraying test rig

2.2.1 Phase Doppler technique

Table 2.2: Parameters of the phase Doppler measurement system.

2.2.3 Results of experimental spray characterization

Exemplary results of droplet diameter and velocity Phase Doppler

(a) Mean droplet diameter D10 . (b) Mean droplet velocity U .

Influence of the target as a displacement body. Since the phase

Figure 2.4: Mass flux as a function of the radial position. Position r = 0 mm

Overview of the spray parameter range By changing the type of at-

2.3 Cooling test rig

2.3.1 Mechanical and thermodynamic construction

• The dimensions of the heated target enlarge approximately 1 mm be-

The above mentioned difficulties lead to a rather complicated construction,

2.3.2 Measurement technique for heat flux and surface

possible due to the presence of the non-transparent target. Infrared imaging

2.3.3 Observation system

3.1 Measurements of the heat flux

Figure 3.1: Calculated temperature T inside the substrate as a function of the

3.2 Visualization of the spray impact

Figure 3.5: Phenomena of spray impact regimes at different surface temperatures

At larger times the substrate is completely covered by a thin boiling liquid

Far field observation By decreasing the magnification of the optical sys-

3.3 Influence of various parameters

q̇ = Sw (Ti − Tsat ), (4.6)

q̇ = Sw (Tw0 − Tsat )Φ(ξ), (4.13)

Moreover, at large times, t tL , Eq. (5.3) approaches the following remote