Group Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Title: 2.6.1 - Measurement of Water Thickness by using Neutron Radiography and Simulation of Gas-Velocity Distributions in a PEFC
ALL VOLUMES CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102023/00062
 Material Information
Title: 2.6.1 - Measurement of Water Thickness by using Neutron Radiography and Simulation of Gas-Velocity Distributions in a PEFC Multiphase Flows with Heat and Mass Transfer
Series Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Physical Description: Conference Papers
Creator: Murakawa, H.
Sugimoto, K.
Asano, H.
Takenada, N.
Publisher: International Conference on Multiphase Flow (ICMF)
Publication Date: June 4, 2010
 Subjects
Subject: polymer electrolyte fuel cell
water thickness
neutron radiography
network analysis
gas velocity distribution
 Notes
Abstract: Fuel gas (hydrogen gas) and oxidant gas (air) are supplied to a Polymer Electrolyte Fuel Cell (PEFC). Condensation may occur in the cathode side, since air is super-saturated by the fuel cell reactions. If condensed water exists in a gas diffusion layer (GDL) or the gas channels, it may affect the fuel cell performances because of blocking the oxygen from reaching the cathode reaction site. In order to clarify water effects on performances of a PEFC, visualization and quantitative measurements of water distributions in a PEFC were carried out by means of neutron radiography. The cell voltage and the pressure drop between the inlet and the outlet of air were simultaneously measured in single- and three-serpentine channels. From the results, it was found that the water easily accumulate in the GDL under rib rather than that under channel at beginning of the operation. The maximum water accumulate in the GDL is about 150 μm. Furthermore, a network analysis of gas-velocity distribution is applied for the experimental results. It analyzes the gas-velocity distribution depending on the flow resistance which is the pressure drop. Applying the measured data of water thickness, pressure drop in the gas channel and the GDL can be obtained. From the calculation, pressure drop between inlet and outlet of air was compared with the experimental results.
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: VID00062
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: 261-Murakawa-ICMF2010.pdf

Full Text

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


Measurement of Water Thickness by using Neutron Radiography
and Simulation of Gas-Velocity Distributions in a PEFC


Hideki Murakawa, Katsumi Sugimoto, Hitoshi Asano and Nobuyuki Takenaka

Department of Mechanical Engineering, Kobe University
1-1 Rokkodai, Nada, Kobe, 657-8501 Japan
murakawa@mech.kobe-u.ac.jp


Keywords: polymer electrolyte fuel cell, water thickness, neutron radiography, network analysis, gas velocity distribution


Abstract

Fuel gas (hydrogen gas) and oxidant gas (air) are supplied to a Polymer Electrolyte Fuel Cell (PEFC). Condensation may occur
in the cathode side, since air is super-saturated by the fuel cell reactions. If condensed water exists in a gas diffusion layer
(GDL) or the gas channels, it may affect the fuel cell performances because of blocking the oxygen from reaching the cathode
reaction site. In order to clarify water effects on performances of a PEFC, visualization and quantitative measurements of
water distributions in a PEFC were carried out by means of neutron radiography. The cell voltage and the pressure drop
between the inlet and the outlet of air were simultaneously measured in single- and three-serpentine channels. From the results,
it was found that the water easily accumulate in the GDL under rib rather than that under channel at beginning of the operation.
The maximum water accumulate in the GDL is about 150 tpm. Furthermore, a network analysis of gas-velocity distribution is
applied for the experimental results. It analyzes the gas-velocity distribution depending on the flow resistance which is the
pressure drop. Applying the measured data of water thickness, pressure drop in the gas channel and the GDL can be obtained.
From the calculation, pressure drop between inlet and outlet of air was compared with the experimental results.


Introduction

A Polymer Electrolyte Fuel Cell (PEFC) consists of a
Membrane Electrode Assembly (MEA) sandwiched with
Gas Diffusion Layers (GDLs) and separator plates as shown
in Figure 1. The GDL is porous media made of carbon fibers.
Fuel gas (hydrogen gas) and oxidant gas (air) are supplied to
the PEFC. At the anode, protons and electrons are generated,
while at the cathode the protons and electrons recombine to
form water. Condensation may occur in the cathode side,
since air is super-saturated by the fuel cell reactions. If
condensed water exists in the GDL or the gas channels, it
may affect the fuel cell performances because of blocking
the oxygen from reaching the cathode reaction site.
However, relation between water distributions in the PEFC
and the cell performances are not completely understood,
and further investigations are required. Many researchers
have tried to investigate water transport mechanism in the
PEFC by using a transparent fuel cell. Liu et al. (2006)
showed water movement with gas-flow rate and cell voltage.
Spernjak et al. (2007) also investigated the relation between
the flow-field flooding at cathode with the cell voltage. A
new parameter called wetted area ratio was introduced to
characterize channel flooding (Hussai & Wang 2009).
However, there may be differences of electrical
characteristics between transparent and actual PEFCs.
Neutron radiography is one of the effective tools for
observing the water distributions in an in-situ PEFC. Many
researchers have tried to measure the water transport
phenomena by using neutron radiography (Ueda et al. 2008,
Turhan et al. 2006, Hickner et al. 2006, Kim et al. 2006,


Satija et al. 2004, Ludlow et al. 2006, Sakata et al. 2009),
and showed the water distributions inside the PEFC.
However, effects of neutron scattering at the PEFC must be
removed from the obtained data for evaluating the
quantitative water thickness. The authors have tried to
measure the quantitative water thickness in the PEFC
(Murakawa et al. 2009) by using umbra method (Takenaka
et al. 2001). These measurements are important for
evaluating the water effect on gas-velocity distributions.
In order to clarify the water effects on performances of the
PEFC, visualization and the quantitative measurements of
water distributions in a PEFC were carried out by means of
a neutron radiography facility at JRR-3 in JAEA (Japan
Atomic Energy Agency). The cell voltage and the pressure
drop between the inlet and the outlet of air were
simultaneously measured. From the results, water
accumulation mechanism in the GDL under rib and channel
is discussed. Maximum water thickness in the GDL is one
of the important parameter, and it is estimated from the
experimental results. Furthermore, a network analysis of
gas-velocity distribution is carried out. It analyzes the
gas-velocity distribution depending on the flow resistance
which is the pressure drop. Applying the measured data of
water thickness, pressure drop in the gas channel and the
GDL can be obtained. From the calculation, pressure drop
between inlet and outlet of air is compared with the
experimental results.









Nomenclature


channel height (m)
channel width (m)
hydraulic equivalent diameter (m)
characteristic pore diameter (m)
superficial gas velocity (m s-')
channel length (m)
pressure (Pa)
velocity (m s 1)


Greek letters
a void fraction
e porosity
K permeability (m2)
Kk Kozney constant
A friction drag coefficient
ip viscosity (Pa s)
p density (kg m3)

Experiment and Data Analysis

A visualization fuel cell for the neutron radiography used in
this research is shown in Figures 2(a) and (b). The PEFC is
compliant with JARI (Japan Automobile Research Institute)
standard PEFC. Materials or thickness of the holding and
the separator plates, and the current collectors were changed
for increasing neutron transmission without influence on the
fuel cell performances. Concretely speaking, the holding
plates and the current collectors were made of aluminum.
The GDL was TGP-H-060 (Toray Ind.) with thickness of
190 pm. The MEA was Nafion NR-212 with thickness of
51 pm and area of 50x50 mm2. Temperature of the PEFC
was kept at 80 OC using rubber heaters. In order to
investigate effects of channel geometry, two kinds of
channel geometry as shown in Figure 3 were used for the
measurement. The one is single-serpentine, and the other is
three-serpentine. Size of the channel depth and width was 1
mm, and the channel area was 52x53 mm2.
The neutron radiography was carried out at JRR-3 in JAEA.
The pictures were taken by a cooled CCD camera (PIXES
1024, Princeton Instruments) with a resolution of
1024x1024 pixels and gray scale of 16 bit. The exposure
time was set at 12 sec, and pictures were taken at 15 sec
intervals. The spatial resolution which corresponds to 1
pixel was 108 tpm. Pressure drop between inlet and outlet of
the air, and the cell voltage were simultaneously measured.
Umbra method (Takenaka et al. 2001) was practiced for the
analysis of water distributions by using neutron absorber
grids made of boron as shown in Figure 4. Width and space
of the grids are equally 3 mm. Using the neutron absorber
grids to remove an influence of scattered neutron in the
PEFC, quantitative measurements can be conducted. Images
were computed between after and before the PEFC
operations for quantifying the generated water, and average
water-thickness of 1 x3 mm2 area in both the channel and the
rib, time series of water distributions were obtained.


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

Anode Cathode
S> Gas Channel e


H2


mm 0 Rib
S H2 H H20



GDL-190gm MEA-50gtm
Figure 1: Mechanism of a Polymer Electrolyte Fuel Cell.


Air inlet


2
H2 inlet


H2 outle


N
Air outlet
Air outlet


Neutron
beam





MEA+GDL Separator


(a) Images of the PEFC


MEA GDL Gasket
Holding plate
(b) Structure of the PEFC
Figure 2: Visualization cell for neutron radiography.


Air inlet


Air inlet


52mm 52mm


Ill
EM




Air outlet Air outlet
(a) Single-serpentine (b) Three-serpentine
Figure 3: Schematic diagram of gas channels.






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


Figure 4: Umbra method by using a neutron absorber
grid.

Water Distributions in the PEFC

Figure 5 shows results of time series of cell voltage,
pressure drop and area-average water thickness all over the
measurement area in single-serpentine channel. The
area-average water thickness was divided into channel and
rib positions. The water thickness at channel represents total
water in the MEA, the GDL and the channel. In contrast, the
water thickness at rib represents water in the MEA and the
GDL. Accordingly, difference of water thickness between
channel and rib indicates existence of condensed water
mainly in the channel. Furthermore, production of water in
the MEA and the GDL is obtained from the water thickness
at rib. The experimental conditions were current density of
200 mA/cm2, oxygen flow-rate of 227 cc/min (utilization:
36.5%) and hydrogen flow-rate of 400 cc/min (utilization:
9.5%). The relative humidity (RH) of both air and hydrogen
was set at 81%. After pre-operation for setting condition of
the cell voltage, the time was referred to as starting time. A
glance at the result reveals that the pressure drop increases
with water thickness until 10 min from the starting time.
Although the pressure drop gradually decreases with
increasing the area-average water thickness, the cell voltage
is almost the constant. The average water-thickness at
channel and rib is almost the constant around 90 min. Two
dimensional water distributions are shown in Figures 6. At
30 min, accumulations of water around the channel corer
are confirmed. During the operation, it increased until 110


Z -U


Current density 200 nA/cm
8 Oxygen 227 cc/rmn Hydrogen 400 cc/nm 2

6 Cell voltage 15
4 .- .1
2 10
2: Pressure drop

0 50 100 150
Time [min]

(a) Cell voltage and pressure drop

)0. .. .... ..... I30
Current density 200mnA/cm
Oxygen 227 cc/irm Channel
)0 Hydrogen 400 cc/mn -20


)0 Rib -10
0Rib


Time[min]
(b) Average water thickness


Figure 5:
channel.


Air inlet



1350
1200
1050
750
460

0

Air outlet
(a) 2min (b) 30min


[Wn]
I1500
iron
1350
1200
wo
600
450

0


Iwo 1650
2 1200


fso
70



(c) 70min (d) 110min
Figure 6: Water distributions in single-serpentine channel.

min. The water plugs sometimes moved, and water ejection
was confirmed.
Results in three-serpentine channel are shown in Figures 7.
It should be noted that range of the pressure drop in
three-serpentine channel is lower than that in
single-serpentine. It is clear that the cell voltage and the
pressure drop have large fluctuation. And the pressure drop
increases as the cell voltage decreases. It can be confirmed
that the cell voltage has a sudden recovery when the
pressure drop suddenly decreases. For more detail
examination, graph between the times (1) and (2) around 22
min in Figure 7(a) is closed-up as shown in Figure 8. The
cell voltage and the pressure drop are linked with each other,
and the pressure drop suddenly decreases with increasing of
the cell voltage. Two dimensional water distributions are
shown in Figures 9. At 22 min, there are some water liquid
plugs around middle of the channels. After the 15 sec, it can
be confirmed that a water plug moved to exit of the channel.
After the 30 sec, movement of water plugs and water
ejection were confirmed with increasing of the cell voltage.
This result indicates the fact that the water plugs may reduce
the cell performance, and efficient ejection of the water is
required for supply of the gas.


0


0
0U
o
o,
o|


150


Measurement results in single-serpentine






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


(1)' (2) Current density 200 mA/c2m 2

-2
0.8 / ...
0.6 C,
0.6 Pressure drop
0.4-1

Cell voltage
0 20 40 60 80
Time [min]
(a) Cell voltage and pressure drop
300C ,,,. 300
Current density 200nmA/cmi
Oxygen 227 cc/mm
200- Hydrogen 400 cc/mm -200
Channel .2

o100- P,- -100


0 20 n 60n s8


Time[min]
(b) Average water thickness
Figure 7: Measurement results in three-serpentine channel.


Figure 8: Close-up of
three-serpentine channel.


Time [min]
cell voltage


and pressure drop in


1350
1200


600
SOO
450

0


(a) 22min


(b) 22minl5sec


I[ml
1350
1200
1050
900
750
No
4SO
150
0.


(c) 22min30sec (d) 22min45sec
Figure 9: Water distributions in three-serpentine channel.



Mechanism of Water Accumulation and Ejection in
the GDL

In order to understand the mechanism of water
accumulation in the GDL, time-series of local water
thickness are shown in Figures 10. The experimental


(a) Case of water accumulation


EI

5'


)


Time [s]
(b) Case of drainage from upstream
Figure 10: Time-series of local water-thickness at a channel.

conditions were current density of 300 mA/cm2, oxygen
utilization of 36.5% and hydrogen utilization of 13.1% in
single-serpentine channel. Typical two-cases of the
water-thickness changes in different position are picked up.
In the case of the Figures 10(a), water gradually generated
from the starting time. The water thickness takes maximum
value around 3550 s, and it suddenly decreases. It represents
that the water gradually accumulated in the GDL and the
channel, and the water ejected to downstream. It must be
noted that the water thickness doesn't became 0 after the
time. It takes around 150 tpm after the water ejections.
Maximum water thickness in the GDL can be estimated as
about 150 pm taking into account thickness and porosity of
the GDL. It can be considered that the water in the only
channel was ejected at the time, and the water in the GDL
under the channel still existed. Furthermore, if water
accumulates over 600 pm at a channel, the water is easily
ejected to downstream under the flow condition. On the
other hand, it was observed that water thickness around
6000 s suddenly changes as shown in Figure 10(b). This
indicates that a water plug passed at the channel, and the
water in the GDL under the channel is almost the constant.
Local water thicknesses at channel are compared with that
at rib. Figures 11(a)-(c) represent the comparison of water
thickness at a channel and at a rib next to the channel
position. If water thickness suddenly changes at a channel, it
indicates that a water plug passed through the position. In
this case, the water at a rib next to the channel is slightly
affected by the water ejection. In particular case of Position
II, water thickness in the GDL became thicker because of a
water ejection in the channel. Movement of a water plug in
a channel increases the water thickness in the GDL. It is
apparent if the water in the GDL is less than 50 pm. On the
other hand, with increases of water thickness in a GDL, the
effects of the water movement in a channel to the water in
the GDL becomes little. Although movement of a water
plug affects the water thickness in the GDL around 2100 s
in the case of Position III, water accumulation after 2800 s
doesn't affect the water thickness in the GDL next to the
channel.


a


sO






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


100 200 300 400 500 600
Water Thickness under Channels[ rim]


Time [s]
(b) Position II


Time [s]
(c) Position III
Figure 11: Comparison of water thickness at the channel
and the rib.

Estimation of water accumulation in the GDL is important
to calculate the network analysis based on the pressure drop
in the GDL. Figure 12 represents the relation of water
thickness in the GDL between under rib and under channel.
The experimental conditions were current density of 200
mA/cm2, oxygen flow-rate of 227 cc/min (utilization:
36.5%) and hydrogen flow-rate of 400 cc/min (utilization:
9.5%). The horizontal axis indicates water thickness under
channel, and the vertical axis indicates ratio of water
thickness between under rib and under channel next to the
rib. Early time of the starting time, the ratio widely
distributes. The ratio of 1 means that water in the GDL
under rib and channel equally accumulated. However, there
are many places that take over 1 at 100 s. It represents that
water easily accumulated in the GDL under rib than that in
the channel at early stage. With increasing of the PEFC
operation, the distributions tend to around 0.5-1.5 at 500 s.
From the results, it is also confirmed that water in the GDL
under rib is easily accumulated than that under the channel.
As the PEFC operation, condensed water ejected from the
GDL to the channel. As a result, water accumulated in the
channel, and the higher water thicknesses were confirmed.
A line expresses a full capacity of the GDL water. As
porosity and thickness of the GDL is 0.78 and 190 pm,
maximum water thickness in the GDL is about 150 pm.
Therefore, the line indicates the theoretical maximum values
of water in the GDL. At 700 s, many of the plots distribute
around the full water line. In particular, if water exists in a
channel, water in the GDL next to the channel has also
maximum water thickness. This tendency is more apparent
in higher water thickness in the channel.


Figure 12: Relation of water thickness under rib and
under channel.

Basic Equations of Network Analysis of
Gas-Velocity Distributions

As shown in the experimental results, the water thickness,
the pressure drop between inlet and outlet of air and the cell
voltage are strongly linked with each other. Furthermore,
accumulation of water may affect the gas supply and the cell
performances. If gas flows non-uniformly in the GDL
because of the generated water, the current density has also
non-uniformity.
For analyzing gas-velocity distribution in the PEFC,
network modeling was developed as shown in Figure 13.
Air is supplied from the inlet, and the flow is distributed in
the channel and the GDL. The flow distributes in each
calculation volume depending on the flow resistance which
is the pressure drop.
The basic equations are conservation of mass and pressure
drops in the channel and the GDL. The pressure drop in the
channel can be obtained by the following equation.
L jw2
AP =2 (1)
D 2
where L is the channel length, D is the hydraulic equivalent
diameter, p is the density, u is the velocity. A is the friction
drag coefficient, and can be expressed as
64
2~ k, (2)
Re
k is the geometric coefficient and is 0.889 in square duct. If


Figure 13: Network modeling of gas-velocity distribution
in the PEFC.


(a) Position I






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


water exists in the channel with void fraction of a, u/a
and D' are used instead of u and D. And the D'is expressed
as
D'= (a+b)D,
Ca+ (3)

a and b are the channel height and width.
On the other hand, transport of gas in the GDL is obtained
by the following Darcy's equation.
J, =- VP, (4)

where K is the permeability, Jg is the superficial velocity that
takes into account the facial porosity s and a, and is
expressed as
J= usa, (5)
where u is the average gas velocity in the porous media. The
permeability can be obtained from Carman-Kozeny theory
and expressed as (Lister & Djilali 2005);
d2 3
K = p (6)
*, (1-i )2
where dpor is the characteristic pore diameter and Kk is
Kozney constant, which is evaluated from a shape factor
and tortuosity factor. If the water exist uniformly in the
GDL,

(1 a):(7)
',., (1- ea)2

For calculating the proposed network analysis, the void
fraction in the PEFC is required. Therefore, the values are
obtained by the experimental data of neutron radiography.
The calculating conditions are shown in Table 1. General
values of a GDL are used for Kk and dpor,. The gas channel is
three-serpentine with area of 52 x 53 mm2, and the size
geometries are shown in Figure 14. The channel and rib
-width and -height are equally 1 mm, and thickness of the
GDL is 190 gm. For adopting the umbra method which
employs the boron grid with width and space of 3 mm, the
size of calculation mesh is 3 mm in horizontal direction
except the channel comers. For adjusting the channel comer,
mesh size was set at 1 mm around the comer. On the other
hand, vertical mesh is 1 mm that is the same width with the
channel and the rib.

Table 1 Calculating conditions
S: 0.78
Thickness of the GDL : 190 pm
Kk :5
dpo, : 12.9 pm
Channel geometry : Three-serpentine


Rib -


/GDL


4190um


1mm


Figure 14: Geometries of the gas channel.

Two-dimensional water thickness in the PEFC can be
obtained by the neutron radiography. However, the
information is integrated values along the neutron beam.
Therefore, the thickness at the channel includes the water
both in the channel and the GDL. If the water thickness is
larger than maximum water-thickness in the GDL, the
excess water is considered to be in the channel. From the
data, void fraction distribution in the channel and the GDL
was calculated. For the gas transportation between the
channel and the GDL, half thickness of the GDL is used in
Eq.(4). Furthermore, the change of the GDL thickness
owing to the holding pressure, and anisotropy of the GDL
are neglected.

Results and discussions

In order to validate the model, pressure drop in single-phase
flow in the PEFC was compared with experimental results
as shown in Figure 15. The channel geometry is
three-serpentine. Air was supplied into the PEFC, and the
pressure drop between the inlet and the outlet was measured
at 80 C without generation of the electricity. The gas-flow
rate is expressed under 0.1 MPa, 0 C and air-RH of 81%.
The result reveals that the model has a good agreement with
the experimental results under air-flow rate of 400 cc/min.
However, with increase of the gas-flow rate, the data has
large difference between the experiments and analysis. The
pressure drop of the analytical data does not increase
linearly with the gas-flow rate. This indicates that the
pressure loss include not only the friction loss in the channel.
It may indicate that the pressure drop at the channel comers
cannot be neglected under higher air-flow rate. Typical air
flow-rate of single cell of the JARI standard PEFC is less
than 400 cc/min, and the experimental condition of neutron
radiography is 227 cc/min. Therefore, a simulation and
measurements in two-phase flow can be compared.

10
S Calculation results
A Channel+GDL Experimental results
2 6-
4A
2- A


0 200 400 600
Flow rate[cc/min]


800 1000


Figure 15: Comparison of pressure drop in single-phase
flow.






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


Jg: .02[rrn s]

a[-]

0.9
0.8
0.7


(a) Superficial gas-velocity and void fraction distributions
in the Channel


z z zI I I I I I I I I I I IIz z


(b) Superficial gas-velocity and void fraction distributions
in the GDL






Jgz[m/s]
S0.005
0 004
0.003
0.002
0 001
0
-0.001
-0 002
-0 003
-0.004
-0.005


(c) Superficial gas-velocity distribution between
the channel and the GDL

Figure 15: Gas-velocity distributions in single-phase flow.


Calculation results in single-phase flow at 227 cc/min are
shown in Figures 15. Figures 15(a), (b) represent the
void-fraction distributions, and the vectors indicate the
superficial gas-velocity (Jg) distribution. For understanding
the velocity information between the channel and the GDL,
Figure 15(c) represents the superficial gas velocity between
the channel and the GDL. The negative value is direction
from the channel to the GDL. It must be noted that the flow
velocity in the channel is much larger than that in the GDL.
Therefore, the gas mainly flows in the channel, and a little
gas shortcut the GDL under the rib. This can be confirmed
from the flow vectors in the GDL either. The shortcut-flow
occurs because of the pressure drop in the channel.
Figures 16 and 17 show the void-fraction and gas-velocity
distributions at 10 and 30 min from the starting time.
Experimental conditions are current density of 200 mA/cm2,
oxygen flow-rate of 227 cc/min, air-RH of 81% and
hydrogen flow-rate of 400 cc/min. The tendency of the flow


in the channel is almost the same with that in single-phase
flow. As shown in the experimental results, the water
gradually accumulates around the central area. It is obvious
from Eq.(7) that the permeability at the GDL decreases by
the existing water. Therefore, the flow rate in the GDL
decreases with void fraction. The gas moves to the channel
for avoiding water in the GDL. Superficial gas-velocity
distributions between the channel and the GDL also
decrease with the void fraction. At 30 min, much water
accumulated around center region in the GDL. As a result,
gas supply into the GDL decreases with increasing of the
water. As water accumulates in a channel position, gas flows
in the other channels.


Jg : 3.5 [rrs]

c[-]

0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0


U
U.....


U
U.....


U
U.....


U
U


. . . . ..IIIIIIn l





. . . . . I I l] I











-I-
Jg 3 5 [m/s]

a[-]


0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0


(a) Superficial gas-velocity and void fraction distributions
in the channel


Jg: O.01[m/s]

a[-]

0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0


(b) Superficial gas-velocity and void fraction distributions
in the GDL






Jgz[m/s]
0.001
0.0008
0.0006
0.0004
0.0002
0
-0.0002
-00004



-0.001


(c) Superficial gas-velocity distribution between
the channel and the GDL
Figure 16: Calculation results at 10 min.


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



Jg 3 5 [m/s]




09
Clo


(a) Superficial gas-velocity and void fraction distributions
in the channel


Jg: 0.01[m/s]

a[-]

09
08
07
0.6
05
0.4
03
02
0 1
0


(b) Superficial gas-velocity and void fraction distributions
in the GDL






Jg [m/s]
0,001
0.0008
0.0006
0.0004
0.0002
0
-0.0002
-0 0004
-0 0006
-0 0008
-0 001






(c) Superficial gas-velocity distribution between
the channel and the GDL
Figure 17: Calculation results at 30 min.


'Eli
MEN.
MMM..L7

MINE MMMIMMMIMM
.... ..... ....
- MENE - MENE MEN -

MINE
.............MIS

1 177777 r7lIg



Minn 14
ME,


MEN
NONE




""onni

MIS
MINE !is
MINE MENEM
MMMI MINE = = 111111mominEMEME
............. MIS - -- --- -


7:1::::
7








Experimental result Current density 200 mA/cm
---Anahticalresult Oxygen 227 cc/mm
Hydrogen 400 cc/mn

o---------------- -- -
P)
0 20 40 60 80
Time [min]
Figure 18: Result of pressure drop between inlet and
outlet of air.

Pressure drop between inlet and outlet of air is compared
with calculated results as shown in Figure 18. The
calculations were conducted every 15 sec. The experimental
results include both fluctuations of higher and lower
frequencies. The measurement of water thickness is average
over 12 sec which is exposure time of a camera. Therefore,
fluctuation of short period less than 12 sec cannot be
analyzed in the calculation. The analytical results showed
that the large fluctuation of the pressure drop appears
because of the water ejection, and large fluctuation can be
simulated. However, the value of the pressure drop has large
difference between the experimental and analytical results
though the initial pressure drop is good agreement with the
each other. Fluctuation of the pressure drop is mainly caused
by the water ejection, and the large fluctuation can be
considered as water ejection was occurred at the time. The
reason of the pressure difference can be considered that the
model doesn't include the pressure drop in two-phase flow.
Furthermore, treatment of the pressure drop in the channel
corer is one of the problems. For more investigation,
models are required to deal the water shape in the channel,
and to predict the pressure drop around the channel corer at
higher flow rate.

Conclusions

Water behavior in an operating Polymer Electrolyte Fuel
Cell (PEFC) was visualized by using neutron radiography.
The cell voltage and the pressure drop between the inlet and
the outlet of air were simultaneously measured in single-
and three-serpentine channels. The results in
three-serpentine were applied for a network analysis to
obtain the gas-velocity distributions. From the results, the
followings are obtained;
* Water easily accumulates in the GDL under rib than
that under the channel at the beginning of the PEFC
operation.
Water ejection was confirmed from a local water
thickness. Two types of water ejection were confirmed.
If water gradually condensed and ejected at a channel,
water in the GDL still existed after water ejection in the
channel.
Flow rate in the GDL decreases with increasing of void
fraction. The gas moves to the channel for avoiding the
water in the GDL.
Analytical results showed that the large fluctuation of
the pressure drop appears because of water ejection,
and large fluctuation can be simulated. However, value
of the pressure drop has large difference between the
experimental and analytical results. For more
investigation, two-phase flow and channel corer
modeling of pressure drop is required.


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


Acknowledgements

The authors acknowledge the New Energy and Industrial
Technology Development Organization (NEDO) for their
financial support.

References

Hickner, M.A., Siegel, N.P & Chen, K.S., Exploring Liquid
Water Distribution and Local-Heating Effects in an
Operating PEM Fuel Cell Using Neutron Radiography, Proc.
8th World Conf. on Neutron Radiography,
#T323(CD-ROM)(2006)

Hussaini, I.S. & Wang, C.-Y., Visualization and
quantification of cathode channel flooding in PEM fuel cells,
J. Power Sources, Vol. 187, 444-451(2009)

Kim, T, Kim, J., Kim, M., Sim, C., Lee S.W. & Jeon, J.,
Research for Water Removal at PEMFC by using Neutron
Imaging Technique at NRF, HANARO, Proc. 8th World
Conf. on Neutron Radiography, #T324(CD-ROM)(2006)

Litster, S. & Djilali, N., Kinetic Theory of Plasma, in B.
Sunden and M. Faghiri (ed.), Transport phenomena in fuel
cells, chap. 5, WIT Press, Boston(2005)

Liu, X., Guo, H. & Ma, C., Water flooding and two-phase
flow in cathode channels of proton exchange membrane fuel
cells, J. Power Sources, Vol. 156, 267-280(2006)

Ludlow, D.J., Calebrese, C.M., Yu, S.H., Dannehy, C.S.,
Jacobson, D.L., Hussey, D.S., Arif, M., Jensen, M.K. &
Eisman, GA., PEM fuel cell membrane hydration
measurement by neutron imaging, J. Power Sources, Vol.
162, 271-278(2006)

Murakawa, H., Ueda, Y., Yoshida, T., Sugimoto, K., Asano,
H., Takenaka, N., Mochiki, K., likura, H., Yasuda, R. &
Matsubayashi, M., Effect of water distributions on
performances of JARI standard PEFC by using neutron
radiography, Nucl. Inst. and Methods in Physics Research -
A, Vol. 605(1-2), 127-130( 2009)

Sakata, I., Ueda, T, Murakawa, H., Sugimoto, K., Asano, H.,
Takenaka, N., Yasuda, R., Tomura, T. & Shiozawa, M.,
Three-dimensional observation of water distribution in
PEFC by neutron CT, Nucl. Inst. and Methods in Physics
Research- A, Vol. 605(1-2), 131-1 1 i",

Satija, R., Jacobson, D.L., Arif, M. & Werner, S.A., In situ
neutron imaging technique for evaluation of water
management systems in operating PEM fuel cells, J. Power
Sources, Vol. 129, 238-245ki "l4)

Spernjak, D., Prasad, A.K. & Advani, S.G, Experimental
investigation of liquid water formation and transport in a
transparent single-serpentine PEM fuel cell, J. Power
Sources, Vol. 170, 334-344(2007)

Takenaka, N., Asano, H., Fujii, T. & Matsubayashi, M., A
Method for Quantitative Measurement by Thermal Neutron





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

Radiography, Nondestructive Testing and Evaluation, Vol.
16(2-6), 345-354(2001)

Turhan, A., Kowal, J.J., Heller, K., Shi, L., Brenizer, J. &
Mench, M.M., Interaction of Design, Materials, and
Interfacial Forces on Liquid Water Storage and Distribution
in Polymer Electrolyte Fuel Cells, Proc. 8th World Conf. on
Neutron Radiography, #T142(CD-ROM)(2006)

Ueda, T., Takenaka, N., Asano, H., Tanimoto, K., Mochiki,
K., Kawabata, Y. & Matsubayashi, M., Visualization and
Measurement of Dynamic Water Behavior in Polymer
Electrolyte Fuel Cell by Neutron Radiography J. Power
and Energy Systems, Vol. 2(3), 997-1008(2008)




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 - - mvs