Group Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Title: 6.6.4 - Water Quenching of Particle-Laden hot Supersonic Gas Flows
Full Citation
Permanent Link:
 Material Information
Title: 6.6.4 - Water Quenching of Particle-Laden hot Supersonic Gas Flows Multiphase Flows with Heat and Mass Transfer
Series Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Physical Description: Conference Papers
Creator: Rakel, T.
Schaber, K.
Publisher: International Conference on Multiphase Flow (ICMF)
Publication Date: June 4, 2010
Subject: quenching
particle-laden gas
crossflow injection
Abstract: In the course of an innovative concept for high throughput gas-phase synthesis of oxide nanoparticles a new supersonic quenching system is developed to instantaneously stop nanoparticle growth. Supersonic quenching combines high gas dynamical cooling rates of more than dT/dt=-106 K/s with a total enthalpy reduction through evaporation of an injected liquid. The hot gas stream is accelerated to Mach 2 or higher and subsequently water is injected. Due to partial evaporation within the supersonic regime a deceleration without temperatures rising to former levels is possible. Until present there are no publications covering this particular application. In this paper the experimental facility is described and the technical proof of concept is presented by means of experimental results. Measured pressure profiles verify supersonic flow conditions of the generated disperse two-phase mixture, which even expands further downstream of the injection. Temperatures in the subsonic domain prove sufficient evaporation and indicate a homogeneous phase as well as temperature distribution. The presented numerical results confirm to experimental data regarding the pressure profile. As static temperatures are not measureable the numerical results can be used to evaluate the achieved integral cooling rates, which are in the order of 105 K/s. The nanoparticles leaving the supersonic quench are verifiable spherical and non-aggregated.
General Note: The International Conference on Multiphase Flow (ICMF) first was held in Tsukuba, Japan in 1991 and the second ICMF took place in Kyoto, Japan in 1995. During this conference, it was decided to establish an International Governing Board which oversees the major aspects of the conference and makes decisions about future conference locations. Due to the great importance of the field, it was furthermore decided to hold the conference every three years successively in Asia including Australia, Europe including Africa, Russia and the Near East and America. Hence, ICMF 1998 was held in Lyon, France, ICMF 2001 in New Orleans, USA, ICMF 2004 in Yokohama, Japan, and ICMF 2007 in Leipzig, Germany. ICMF-2010 is devoted to all aspects of Multiphase Flow. Researchers from all over the world gathered in order to introduce their recent advances in the field and thereby promote the exchange of new ideas, results and techniques. The conference is a key event in Multiphase Flow and supports the advancement of science in this very important field. The major research topics relevant for the conference are as follows: Bio-Fluid Dynamics; Boiling; Bubbly Flows; Cavitation; Colloidal and Suspension Dynamics; Collision, Agglomeration and Breakup; Computational Techniques for Multiphase Flows; Droplet Flows; Environmental and Geophysical Flows; Experimental Methods for Multiphase Flows; Fluidized and Circulating Fluidized Beds; Fluid Structure Interactions; Granular Media; Industrial Applications; Instabilities; Interfacial Flows; Micro and Nano-Scale Multiphase Flows; Microgravity in Two-Phase Flow; Multiphase Flows with Heat and Mass Transfer; Non-Newtonian Multiphase Flows; Particle-Laden Flows; Particle, Bubble and Drop Dynamics; Reactive Multiphase Flows
 Record Information
Bibliographic ID: UF00102023
Volume ID: VID00166
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: 664-Rakel-ICMF2010.pdf

Full Text

Paper No 7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

Water Quenching of Particle-Laden hot Supersonic Gas Flows

T. Rakel* and K. Schaber*

Institute fuir Technische Thermodynamik und Kaltetechnik, Karlsruhe Institute of Technology, Karlsruhe, Germany
Engler-Bunte-Ring 21, D-76131 Karlsruhe, Germany
thomas.rakel & karlheinz.schaber

Keywords: quenching, supersonic, particle-laden gas, crossflow injection, CFD


In the course of an innovative concept for high throughput gas-phase synthesis of oxide nanoparticles a new supersonic
quenching system is developed to instantaneously stop nanoparticle growth. Supersonic quenching combines high gas
dynamical cooling rates of more than dT/dt=-106 K/s with a total enthalpy reduction through evaporation of an injected liquid.
The hot gas stream is accelerated to Mach 2 or higher and subsequently water is injected. Due to partial evaporation within the
supersonic regime a deceleration without temperatures rising to former levels is possible. Until present there are no
publications covering this particular application. In this paper the experimental facility is described and the technical proof of
concept is presented by means of experimental results. Measured pressure profiles verify supersonic flow conditions of the
generated disperse two-phase mixture, which even expands further downstream of the injection. Temperatures in the subsonic
domain prove sufficient evaporation and indicate a homogeneous phase as well as temperature distribution. The presented
numerical results confirm to experimental data regarding the pressure profile. As static temperatures are not measureable the
numerical results can be used to evaluate the achieved integral cooling rates, which are in the order of 10' K/s. The
nanoparticles leaving the supersonic quench are verifiable spherical and non-aggregated.


The general goal of a quenching system is the reduction of
the gas temperature by the evaporation of an injected liquid
phase. A supersonic water quenching system combines a
laval-nozzle and a quenching system with water injection in
one setup. Within the divergent convergent laval-nozzle a
hot gas is accelerated into a supersonic regime, which leads
to very high cooling rates due to the conversion of enthalpy
into kinetic energy. Cooling rates of more than
dT/dt=-106 K/s can be reached. This is especially of interest
for kinetic controlled processes taking place at critical
midrange temperatures. As a deceleration would lead back
to the initial high total temperature, the specific total
enthalpy of the gas flow must be partially reduced within
the supersonic flow regime. Evaporating an injected liquid,
as done in general quenching systems, is due the high latent
heat an effective way to reduce the total enthalpy of the gas
flow. Thus the supersonic quenching system consists out of
a gas dynamical cooling and an evaporative cooling section.
The supersonic quenching system is developed in the frame
of the project "Gasdynamically induced nanoparticle
synthesis", funded by the Deutsche Forschungs-
gemeinschaft (DFG). The goal of the novel reactor concept
is the production of non-agglomerated oxide nanoparticles
with a narrow particle size distribution. The reactor concept
differs clearly from known industrial production processes
like flame synthesis (Wegner & Pratsinis, 2003) or hot-wall
synthesis. But nevertheless it is a high throughput reactor

with gas flow rates up to 100 g/s. In Fig. 1 the reactor setup,
first presented by Grzona et al. (2009), is shown.
Characteristic is the consecutive connection of two
laval-nozzles. The hot gas generated by a methan burner
enters the first laval-nozzle together with the injected
precursor and accelerates to supersonic speed, suppressing
the ignition of the precursor due to the acceleration of the
gas flow. The rise in temperature at the shock front then
initiates the particle growth. Within the reaction volume the
particles grow due to nucleation and coagulation processes,
whereby the gas speed and reactor length defines the
adjustable reaction time. The supersonic quenching system
terminates the particle growth instantly through the high
cooling rates. The specific total enthalpy of the
particle-laden hot gas is reduced by the total evaporation of
the injected water. By the means of a subsequent filter the
nanoparticles are separated from the humid exhaust gas.


IBurner Laval-Nozzle

po= 6 12 bar Mag 100 g/s
Shock To= 1400- 1500 K M r6 gs Water

Reaction volume SupersonicQuench

Reaction volume Supersonic Quench

Figure 1: Reactor setup (Grzona et al. 2009) reproduced
with authors permission

Paper No


area (mm2)
nozzle diameter (mm)
channel height (mm)
impulse correction factor
mass flow (g s-')
Mach number
pressure (bar)
momentum flux ratio
temperature (K)
time (ms)
Weber number
coordinate (mm)
coordinate (mm)

Greek letters
p density (kg m'3)

0 total
jet water nozzle
max maximum
min minimum
QL quench lance
out outlet
stat static
* nozzle throat

Regarding the supersonic quench there are in the literature
until present, to the knowledge of the authors, no directly
comparable test cases published. Of course there are various
publications on supersonic compressible gas flows or
droplet (Renksitzbulut 1991, Aggarwal & Peng 1995) and
spray (Abramzon & Sirignano 1989) evaporation. There are
also publications addressing liquid crossflow injection into
subsonic (Wu et al. 1997) and supersonic (Yates 1972,
Billig 1993) gas streams, whereby most supersonic cases
relate to scramjet combustion. In all cases the mass flux
ratio of gas to liquid is higher and the liquid is not cooling
the hot gas by evaporation. Publications with gas cooling by
the evaporation of liquids on the other hand deal only with
low gas velocities and do not consider or take advantage of
gas dynamical cooling effects (Siepmann & Gusewell 2000).
Therefore the supersonic quenching system is a completely
new application.

Supersonic quench design

A detailed schematic sketch of the supersonic quenching
system is illustrated in Fig. 2. As indicated by the colour
gradient, the gas temperature decreases due to the
acceleration into the supersonic regime. To prevent an anew
temperature rise the quench water is injected directly into
the supersonic domain of the laval-nozzle, where the static
gas temperature is low. The single-phase water nozzles are
located on the wall and on the tip of a so called quench
lance extending from the outlet into the nozzle, in each case
circumferential distributed. Despite the water injection the
main gas flow retains in a supersonic condition and only the
directly from the dispersed water phase covered areas are in
a subsonic regime. Shock waves are initiated by the water

7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

jets as illustrated by unsteady temperature rises in Fig. 2.
The shock front is normal to the flow direction at the water
jet outlet, but as the water jet is distracted with further wall
distance the angle of the shock decreases. Because an
oblique shock only leads to a reduction of the Mach number
without a change to a subsonic regime, the main gas flow
stays in the supersonic regime. Nevertheless the shock
induces an increase in temperature and pressure, whereby
the temperature rise at the jet front is directly damped by the
water. The continuous liquid jet breaks up close to the
nozzle as well as the generated interim ligaments and
droplets which disintegrate quickly to smaller droplets due
to the high relative velocities. Thus the specific gas-water
interphase increases, which is a prerequisite for high heat
and mass transfer fluxes. With the expansion of the
dispersed water phase the water coverage of the cross
section rises downstream from the injection. Hereby the
continuous gas phase stays in a supersonic regime. A
transition to a subsonic regime is not possible without
hitting former temperature levels, until a certain amount is
evaporated and the dispersed phase is homogenously
distributed. The more water is evaporated in the supersonic
domain, the lower the maximum temperatures are in the
subsonic domain. Thus a certain distance and residence time
within the supersonic regime is a prerequisite.

Gas Water

S Walter
1300 T [K] 300
Figure 2: Schematic sketch Supersonic quenching system

The inner contour of the test facility has a round cross
section and with the round quench lance an annular gap is
formed. Starting from a 45 mm diameter at the inlet the
contour is converging to 13.25 mm at the nozzle throat.
From there the contour is diverging until a diameter of 156
mm at the exhaust pipe to the filter.

Experimental Facility

In the course of the novel reactor concept a complete new
experimental facility was built. Especially the requirements
for low fabrication tolerances in combination with high
pressures and temperatures within the reactor are
challenging. For details regarding the general experimental
setup the reader is referred to Grzona et al. (2009). This
paper focuses on the supersonic quenching system. The
flow chart in Fig. 3 displays the relevant components. A
compressor supplies continuously dry pressurized air to a
high pressure buffer vessel and methane is obtained from a
gas bottle set. Both streams are measured by coriolis mass
flow meters. Burning a methane-air mixture creates the
required hot air stream. With increasing gas flow rates the
pressure within burner and reactor builds up automatically.
The ratio between mass flow and pressures depends, except
for the temperature, mainly on the nozzle throats. The throat
of second laval-nozzle from the quenching system therefore
defines the pressure within the reaction volume and equally
the first laval-nozzle throat for the burner. General operating

7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

- 10mm

Water (Deionized) -----F P
Figure 3: Flow chart with relevant components

parameters of the reactor are briefly specified in Fig. 1. The
quench water is deionized, as it is totally evaporated, and
pressurized by a rotary vane pump, which is controlled by a
frequency inverter to adjust the water mass flow rate. The
total water flow rate is measured by a coriolis mass flow
meter next to the pump. Additionally pressure and
temperature are recorded. Afterwards the flow splits up to
feed the quench lance and the wall nozzles, of which the
later ones all have individual feed pipe. The partitioned
mass flow rates are measured by manual rotameters. To
control the facility there is plenty of measurement technique
installed. For this purpose the pressure as well as the
temperature is recorded at the inlet to the reactor, at the inlet
to the supersonic quenching system and at the outlet.
Within the quenching system multiple temperature and
pressure gauges are installed. The pressure is measured
through 1 mm holes at the wall by using two Scanivalve
Sensor Arrays, each with 16 ports. Temperatures are
detected in situ with thermocouples of either 0.5 or 1 mm.
As unmodified thermocouples in the supersonic two-phase
domain of the quench did not withstand the conditions, an
extra protective pipe with an outer diameter of 1 mm and
0.2 mm wall-thickness encloses the 0.5 mm thermocouples.
The 1 mm thermocouples without a protective pipe are used
in the subsonic domain. At a time six thermocouples are
circumferentially distributed at 8 locations downstream of
the water injection. For each position 4 thermocouples were
equipped with an adjustment unit to variegate the radial
position. Between these 8 positions the other thermocouples
are equally partitioned. Beside temperature and pressure
ports there are three optical access points, which are
normally closed by contour fitting metal dummies. Thus one
has to consider the disturbance initiated by planar windows.
Generally it is complicated to apply complex optical
measurement techniques to this particular problem, because
an exact nozzle contour in combination with wearing flow
conditions leads to a massive steel construction.

Another test facility was used for pilot tests, regarding the
massive water injection into supersonic gas flows within a
laval-nozzle. These experiments were carried out in
cooperation with a project partner, the German Aerospace
Agency (DLR), Institute of Aerodynamics and Flow
Technology, Division Wind Tunnels. The facility has a
rectangular cross section and features a wide optical
Plexiglas access. Thus the total gas temperatures were
limited to 343 K at the inlet and water evaporation was
inevitably not taken into account. Fig. 4 displays a basic
sketch of the experimental setup. In this setup the water is
only injected trough nozzles on the moveable quench lance.

Figure 4: Setup for optical cold gas experiments (at DLR)

Numerical Scheme

Beside the experimental research the supersonic quenching
system is analyzed using computational fluid dynamics
(CFD) methods. The widespread commercial available
software Ansys CFX 12.1 is used. The main motivation and
advantage is the availability of all flow variables within a
solution. Especially the static temperature distribution of the
compressible two-phase flow, which is experimentally not
detectable in the quench system, is of interest.
A challenging detail to capture within the model is the
crossflow water injection into supersonic hot gas flows.
Modeling approaches for the injection into subsonic
crossflows were topic of various publications. Detailed
models focusing especially on the injection even take the
primary breakup into account (Rachner et al. 2007). Another
also widely used approach is the direct injection of droplets
with an initial diameter equivalent to the nozzle diameter
(Madabhushi R.K. 2003). Thus only the secondary breakup
is taken into account. The latter approach is supported by
the results of Wu et al. (1997), who investigated the breakup
processes of liquid jets and described the column fracture
using the time scale know for aerodynamic secondary
breakup of spherical droplets. Concerning the supersonic
quench concept the modeling of liquid jet penetration is
important to capture the macroscopic phase distribution. Jet
trajectories have been widely studied (Wu et al. 2007) and
there is a common agreement that the liquid to air
momentum ratio q is the main actuating variable. At the
same time the momentum ratio can be interpreted as a
quotient of the particular Weber numbers We.

S Welrqud Piqudv iquid
Wegas Ogas Vgas

Due to a wide range of conditions there are still diverse
functions for the jet trajectories. Liquid injection into
supersonic streams was studied by Yates C.L. (1972) and
lead to the following jet trajectory correlation.

y/d = 1,15 qln(1+6(x/d))

The correlation calculates the injection depth y at a position
in flow direction x normalized by the nozzle diameter d and
depends only on the momentum ratio q. The calculated
trajectories are in good agreement with our experimental
results from the DLR test facility. A further important fact
for modeling the supersonic quenching system is the
repercussion of the liquid jet onto the compressible gas flow
and the thereby initiated shock waves.

As the injection region is comparatively small to the entire

Paper No


Paper No

quench setup the simplification of injecting droplets with a
diameter equal to the orifice is arguable. The droplets are
modeled as dispersed Lagrange particles and the continuous
Euler gas-phase is simulated as an ideal gas using the
Reynolds-Averaged-Navier-Stokes (RANS) equations. The
impulse exchange between the phases is calculated in both
directions (two-way coupled), changes in the drag
coefficients due to droplet distortion are considered (Liu et
al. 1993) and the secondary droplet breakup is modeled
using the Cascade Atomization and Breakup Model (CAB).
Due to the two-way coupling shock waves initiated at the
liquid interface are captured. For turbulence modeling the
two-equation Shear Stress Transport (SST) is used.
Turbulent dispersion is considered according to the
approach of Gosman & Ioannides (1983). As the gas-phase
is assimilating the evaporating water it consists of an ideal
mixture of air and water vapor. Heat and mass transfer
between gas and dispersed water phase is modeled using the
widespread approach of Ranz & Marshall (1952), in which
the mass transfer model considers the Stefan convection.
Modeling the continuous jet as a row of droplets is essential
for the general Euler-Lagrange approach. On the downside
this simplifications has an impact on the jet deflection
through the impulse exchange model. Thus an unmodified
simulation is not capable to meet experimental jet
trajectories and is over predicting the penetration. A
modification of the impulse transport model would affect
the entire domain and is therefore not appropriate. Instead
an impulse correction factor j=v2dropleV2et is introduced to
reduce the initial impulse of the droplets and to adjust the
modeled jet penetration to the Yates correlation. In this
manner only a boundary condition has to be corrected.
Direct comparisons between experiments and CFD
simulations prove the concept.

Results and Discussion

The presented supersonic quench facility is the result of
pilot tests and the experiences made with an intermediate
supersonic quench design. The results from the intermediate
design are not discussed.
The results from the test facility (Fig. 4) allow an insight
into the proceeding actions due to the wide optical access.
As mentioned the gas is slightly preheated to approximately
343 K and therefore only the liquid injection process
without evaporation is studied. Fig. 5 illustrates a Schlieren
image of the water injection into a supersonic gas flow at a
momentum flux ratio of 1.44. The Mach number at the
water jet position is 2.2, which relates to a gas velocity of
585 m/s.

Figure 5: Schlieren Image Water injection into a cold
supersonic crossflow

7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

On the quench lance the water is injected via 8
circumferentially distributed 0.5 mm diameter nozzles. As
Schlieren images visualize gas density gradients the shock
wave at the front of the water jets is clearly observable.
Downstream of the injection the water covered areas are
expanding. But despite the apparently decreasing height
above the water contour the gas stays in a supersonic regime
and the massive water injection is not blocking the
laval-nozzle. The meanwhile contraction of the water
covered areas seems to be related to the unintentional small
wall edge, which initiates an oblique shock and therefore
proves the supersonic flow conditions. The flow condition
illustrated in Fig. 5 is time-invariant and thus constant
quench conditions can be achieved on a macroscopic scale.

In the following all experimental results correspond to the
supersonic quenching system from the hot gas reactor
presented in Fig. 1 and 3. One crucial reactor operating
parameter is the total pressure or respectively the gas mass
flux. In Fig. 6 the static pressure divided by the initial total
pressure is plotted against the length x in flow direction.
Point of origin is the throat of the laval-nozzle where sonic
speed is reached. The water nozzles on the moveable
quench lance are in all cases positioned at x=95 mm. The
position is equivalent to an isentropic calculated Mach
number of 2.2.



E 0.4-
E 0.2-

- po=4.05; T0=1245 K; Mg, =62.5 g/s; Mwater=20.6 g/s
p ,6=5.02; T0=1245 K; M.,=77.3 g/s; Mwater=24.4 g/s
po=6.32; T0=1273 K; Mgas=97.3 g/s; M w^e=32.0 g/s

0 100 200 300 4600 500 600 700
Distance to A* / (mm)
Figure 6: Static pressure profiles in the supersonic quench

The curves in Fig. 6 prove the fact that the gas phase
remains in a supersonic regime downstream of the water
injection. This conclusion can be drawn from decreasing
static pressures and a concurrently widening cross section.
The waterjets set off oblique shocks, as illustrated in Fig. 2,
which are detected downstream at 116 mm. Upstream at
x=101 mm the injection is not sensed. With an increasing
total pressure the transition from supersonic to subsonic
shifts downstream. The absolute pressure at the shift or
respectively shock position is similar and in the range of
0.25 bar. In each case a shock train shifts the flow into the
subsonic regime instead of a single shock wave, which
results in a curved pressure rise. The two phase flow
supersonic region expands, as the supersonic domain is
elongated with increasing total pressures and the point of
water injection is fixed. As mentioned before this increases
the time respectively distance for the water to be distributed
and to evaporate a higher percentage. This leads to lower
temperatures within the subsonic domain, which is aspired
regarding the avoidance of critical high or mid-range

Paper No

temperatures. The outlet pressure is in each case 1.08 bar,
therefore all dimensionless plots converge to different
values at the outlet.

Instead of increasing gas mass fluxes the supersonic two
phase flow domain can also be elongated by moving the
water injection upstream to lower Mach numbers. Therefore
the quench lance is continuously variable. According to this
there are three axial wall nozzle positions. On the other side
one has to consider, that the gas dynamic cooling domain is
shortened. The water is injected at lower Mach numbers and
therefore higher static temperatures. In Fig. 7 the influence
of the injection position is analyzed by plotting the static
pressure over the length of the quenching system. For all
four data curves the operating conditions are equal. A gas
mass flux of 62.5 g/s and a total temperature of 1263 K
result in a total pressure of 4 bar. The water injection
position is varied between 35 and 135 mm, corresponding to
an isentropic calculated Mach number range from 1.6 to 2.5.


0 2-




- XQLjet=35 mm; MaQLt=1.6 P0=4.0
SXQLet=55 mm; MaQL,et=1.9 To=12
-XQLjet=95 mm; MaQ,,et=2.2 Tout=4
XQLet=135 mm; Ma QL,,=2.5 as=


0 100 200 300 400 500
Distance to A* / (mm)
7: Static pressure profiles for differ

Moving the quench lance upstream intensifies the pressure
rise due to straightening oblique shocks, which relates to the
increasing percentage of water covered area. Experimental
results from Fig. 7 prove that the water injection is not
causing a shift into subsonic flow, which is crucial. But
nevertheless lead sharper pressure rises consequently to
sharper local temperature rises. On the one hand the
temperatures may exceed a critical temperature, but on the
other the flow expands further into the supersonic regime.
Thus the residence time at the higher temperature level is
comparatively low. The second rise in pressure, especially
obvious for x=35 mm, is due to the geometrical reduction of
the widening of the cross section. The influence decreases
the larger the outer diameter is compared the quench lance.
In the case of the rear position at x=135 mm the positively
low impact on the pressure and temperature profile is
overcompensated by the short distance within the supersonic
domain leading to higher temperatures in the subsonic
domain. Downstream the injection all curves converge into
each other. Although the water is injected over a range of
100 mm the shift into the subsonic domain occurs at the
same position.

Another important parameter is the division of the injected
water between the wall and the center nozzles, whereby the
goal is an equally distributed dispersed water phase.

4 bar

7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

Especially at the shift position and within the subsonic
regime the phase distribution must be homogenized. An
uncovered area in the subsonic domain, for example near
the wall, would inevitably result in high temperatures.
Before discussing the recorded temperature profiles from
the supersonic quench, two influencing phenomena have to
be taken into account regarding the temperature
measurement. As already described, thermocouples are used
to detect the temperature. Within the supersonic domain
these thermocouples are not capable of detecting the static
temperature, because they are invasive and thus an
obstruction to the flow. The thermocouple generates a
stagnation point and a shock wave. Due to the
recovery-effect the measured temperature is between the
total temperature and the static temperature. The second
impact factor is the multiphase flow. A thermocouple inside
the dispersed flow is partially wetted by liquid, which
evaporates at wet-bulb temperature. The measured
temperature is inevitable between the gas and the liquid
temperature, whereby the liquid has a dominating influence
due to the higher heat conductivity.

63 K The influence of the wall water mass flux onto the
98 K temperature profile is analyzed by means of Fig. 8. The
32.5 g/s filled symbols connected by solid lines represent the
18.7 g/s averaged temperature at a certain cross section. In each case
the annular gap is divided by thermocouples into 5 equal
concentric rings plus the inner and outer wall temperatures.
Especially near the water injection the measured
temperature can be dominated by either hot gas or cold
liquid. Thus the individual minimal and maximal
temperature value is additionally plotted at each position
600 700 (unfilled symbols connected by dashed lines). Comparing
the different cases in Fig. 8 clearly indicates higher
ent injection averaged temperatures without wall injection. But taking the
pressure profile from Fig. 6 into account shows that these
high temperatures are measured within the supersonic

domain, as the supersonic domain ends at x=200 mm in the
case of 4 bar.



E 600 -




kWall.water=0 g/s-
S -- lwater=2.8 g/s
Wall,water=5.6 gs
\. .- Wl.watr y

- TMn and TMax

.TMin and TMax
--T Mn and Tmax


Po=4.05 bar
To=1244 K

0 100 200 300 400 500 600 700 800
Distance to A* / (mm)

Figure 8: Water wall mass flux
temperature profiles

influence on the

Thus the actual static temperature is much lower and
uncritical regarding the undesired nanoparticle growths.
Injecting 2.8 g/s, which corresponds to 13.5 % of the water
mass flux, clearly decreases the measured temperatures. A
further rise to 5.6 g/s lowers the averaged temperatures to
around 450 K, which is far below the inlet temperature of


Paper No

1244 K. Although the absolute temperature values are not
meaningful within the compressible two phase flow, they
are a good indicator for the phase distribution within the
supersonic quench. Low averaged temperatures are
equivalent to a good coverage of the cross section with
dispersed water. The additionally plotted minimal
temperatures are within the supersonic regime mainly below
373 K. This indicates the presence of a dominating fraction
of liquid water, for example around a disintegrating liquid
jet. The maximal temperatures indicate, whether there are
areas without water coverage. Without wall injection the
maximum temperatures in the subsonic domain downstream
of x=200 mm exceed values of around 750 K, which is
undesired and should to be avoided. In all other cases the
entire subsonic domain is sufficiently covered by the
dispersed water. Towards the outlet both minimum and
maximum temperatures converge to the averaged
temperature values. This indicates a nearly homogeneous
temperature profile at the last two cross sections. In each
case the averaged temperatures only differs by 6 degrees at
the most between these two positions, which is another
indicator for the nearly total evaporation of the water. This
is additionally supported by the low temperature difference
towards the further downstream measured outlet
temperature, here alternatively plotted at the x=800 mm.
The latter mainly applies to the two cases with wall
injection, at which the influence of heat losses through the
meanwhile uninsulated piping should be taken into account.
To sum up, the temperature profile from Fig. 8 proves the
sufficient phase distribution, the avoidance of high
temperatures within the subsonic domain and the total
evaporation of the injected water.

The geometry for the numerical simulations is identical to
the experimental setup, thus a direct comparison of both
experiment and CFD simulation is possible. Concerning the
temperature profile a direct comparison throughout the
domain is not possible, due to the reasons mentioned in the
course of the discussion of Fig. 8. Thus the main benefit of
the simulation is to gain access to the static temperature.
Pressure profiles on the other side can be compared without
reservation. Numerical as well as available experimental
results for one characteristic operating point are shown in
Fig. 9. The water is injected from four wall nozzles at a total
rate of 2.8 g/s (x=57.6 mm) and from the quench lance
nozzles at a total rate of 21.7 g/s (x=115 mm), similar to the
case presented in Fig. 8. The area averaged Mach number,
pressure and temperatures are graphed over the distance to
the nozzle throat. Additionally the Mach number is shown in
terms of color on a two dimensional cut through the domain
featuring two injection nozzles on the quench lance. Not
covered within the plane are the wall nozzles. The Mach
number color plot underlines the combination of gas
dynamic quench and the evaporative cooling. It illustrates
the injection with small local subsonic domain within the
jets and how the two-phase flow is again accelerated to
higher Mach numbers. As the jets totally disintegrate into
dispersed droplets the entire cross section is captured by the
supersonic regime before the shift to the subsonic regime
occurs. The graphed values are area averaged and allow a
qualitative analysis of the numerical results. At the position
of the quench lance water injection the Mach number drops
due to the initiated oblique shock wave. Simultaneously the

7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

static pressure and temperature rises. Because of the lower
mass flow rate the wall water injection impact is
comparatively small. The increasing quench lance diameter
leads afterwards to a short distance of nearly constants
values until the maximum diameter is reached. Downstream
the Mach number rises again due to the further increasing
cross section. The shift to the subsonic regime finally
sharply decreases the Mach number and the pressure rises
up to the outlet pressure level. But important is the
temperature profile and the fact that the evaporation of the
water is suppressing an anew temperature rise as described
in the conceptual design. Comparing experimental and
numerical pressure profile generally validates the simulation.
The model is able to cover the occurring processes. Only the
shift position from super- to subsonic flow is slightly over
predicted. The two useable experimental temperatures at the
outlet are below the calculated temperature profile. Thus the
evaporation is under predicted by the applied model and the
real temperatures should be even lower than the plotted
numerical temperature profile. So the temperature in the
subsonic domain should not exceed the simulated value
750 K. Regarding the nanoparticle growth a temperature
limit of 773 K is aspired. The additionally plotted total
temperature decreases solely because of the water
evaporation. Thus sharp negative gradients point out regions
with high evaporation rates, which occur near the injection
and within the shock train.

0 05 10 1,5 20 2,5
254 c


5, 0 100 200 300 400 500 600 700
Distance to A*/(mm)
41 \4 CFD

1000 --CFD Static Temperature
oo 000 CFD Total Temperature
S9002 Experiment
Mo 400. ---------~ -* --- ----
0 100 200 300 400 500 600 700
Distance to A*/(mm)
Figure 9: CFD results with experimental validation data

Because the velocity within the quenching system is varying
by nearly two orders of magnitude a plot of the
characteristic values over the length is not ideal. Plotting the
numerical calculated temperature profile over the time on a
streamline, as illustrated in Fig. 10, gives a better
understanding of the accomplished high cooling rates. The
conditions in Fig. 10 correspond to Fig. 9.

21300 At=0.6
| 1100-
S100- AT=480K

0.0 0.5 10 15 20 25 3.0 3.5 4.0 45 50
Time on streamline / (ms)
Figure 10: Chronological sequence of the static temperature

M1=77 3 g/s
M _r=21 7 g/s
M_,_=2.8 g/s

Paper No

Due to the high velocities in the supersonic regime the
residence time within it is relatively short. For the analysis a
point of reference at the inlet is chosen, where the
temperature has only dropped by two percent compared to
the total temperature. With it the residence time till the end
of the supersonic domain is less than ten percent. The
integral cooling rate up to the temperature peak in the
subsonic domain, as indicated in Fig. 10, adds up to more
than 8 105 K/s. But even the integral cooling rate from inlet
to outlet of the quenching system is in the order of 105 K/s.

Figure 11: Nanoparticle probe (TEM) at the quench outlet
(Goertz et al. 2009) reproduced with authors permission

Three samples of the generated SiO2 nanoparticle product
are displayed in Fig. 11 to document the proof of concept
regarding the prevention of aggregate structures. The
corresponding probes were taken downstream of the
supersonic quenching system. The nanoparticles in Fig. 11
have a circular shape and are not aggregated. A detailed
analysis of these particle results are presented by Goertz et
al. (2009). The nanoparticle production process is analyzed
in detail by Giglmaier et al. (2010) using a 3-D numerical
simulation, which is also presented on the ICMF 2010.


The concept of the novel supersonic quenching system is
approved by the presented experimental results. Despite the
water injection the generated multiphase flow is further
accelerating to higher Mach numbers. The phase
distribution is qualitatively described using the temperature
measurements, but a quantitative phase distribution
measurement is not possible. On the other hand crucial hot
spots can be detected by the temperature sensors. The
homogeneous temperature distribution at the outlet indicates
a nearly total evaporation, but further validation is needed.
The small discrepancy between numerical and experimental
pressure profile could be related to the applied evaporation
model or unidentified impact factors within the complex
flow problem. Generally the results are in qualitative good
agreement. Thus the numerical results allow the analysis of
experimentally not accessible variables.


This work is support by the German Research Foundation
(DFG) under grant PAK 75/2 "Gasdynamically induced
nanoparticle synthesis" which is greatly acknowledged.


Abramzon, B. & Sirignano, W.A. Droplet vaporization
model for spray combustion calculations. Internat. Journal
of Heat and Mass Transfer, Vol. 32, 1605-1618 (1989)

7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

Aggarwal, S.K. & Peng, F.A. Review of Droplet Dynamics
and Vaporization Modeling for Engineering Calculations.
Journal of Engineering for Gas Turbines and Power, Vol. 17,
453-461 (1995)

Billig, F.S. Research on supersonic combustion. Journal of
Propulsion and Power, Vol. 9, 499-514 (1993)

Giglmaier M., Al-Hasan, N.S. & Schnerr, GH. 3-D
Simulation of the Production of Gas-phase synthesized
Non-aggregated Spherical Nano-particles in Continuous
Gasdynamic Flow. Proceedings International Conference on
Multiphase Flow, Tampa, Fl, USA (2010)

Goertz, V, Korp, D., Al-Hasan, N., Giglmaier, M. & Nirschl,
H. Experimental study of gasdynamically induced
nanoparticle synthesis by use of adapted sampling probes.
submitted to Chemical Engineering Science (2009)

Gosman, A.D. & Ioannides E. Aspects of computer
simulation of liquid-fueled combustors, Journal of Energy,
Vol. 7, 482-490 (1983)

Grzona et al. Gas-phase synthesis of non-agglomerated
nanoparticles by fast gasdynamic heating and cooling.
Shock Waves, Vol. 2, Proceedings, 857-862 (2009)

Liu, B., Mather, D. & Reitz, R.D. Effects of Drop Drag and
Breakup on Fuel Sprays. SAE Technical Paper, No. 930072

Madabhushi, R.K. A model for numerical simulation of
breakup of a liquid jet in crossflow. Atomization and Sprays,
Vol. 13, 413-424 (2003)

Rachner, M., Becker, J., Hassa, C. & Doerr, T Modelling of
the atomization of a plain liquid fuel jet in crossflow at gas
turbine conditions. Aerospace Science and Technology, Vol.
6, 495-506 (2002)

Ranz, W. E. & Marshall, W. R. Evaporation from Drops 1.
Chemical Engineering Progress Vol. 48, 141-146 (1952)

Renksitzbulut, M. A mass transfer correlation for droplet
evaporation in high temperature flows, Chemical
Engineering Science, Vol. 46, 2351-2358 (1991)

Siepmann, S. & Gusewell, M. Droplet size distributions in a
laboratory quench, Chemie Ingenieur Technik, Vol. 72,
1531-1534 (2000)

Wegner, K. & Pratsinis S.E. Scale-up of nanoparticle
synthesis in diffusion flame reactors. Chemical Engineering
Science, Vol. 58, 4581-4589 (2003)

Wu, PK., Kirkendall, K.A. & Fuller, R.P Breakup
Processes of Liquid Jets in Subsonic Crossflow. Journal of
Propulsion and Power, Vol. 13, 64-74 (1997)

Yates, C.L. Liquid Injection into a Supersonic Flow. Aero
Propulsion Lab, Vol. 1, AFAPL-TR-71-97, Wright-Patterson
AFB, OH (1972)

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.7 - mvs