ENGLISH

Examination Information Page

Written examination

General information about the exam:

Subject code: FM1015
Subject name: Modelling of Dynamic Systems
Examination date: 01.12.2023
Examination time: 09:00-12:00
Total hours: 3
Responsible course manager: Bernt Lie
Campus: Porsgrunn
Faculty: Technology, Natural Sciences and Maritime Sciences
No. of assignments: 20 Problems (= questions)
No. of attachments: 1
No. of pages incl. information page and attachments: front page + 16 (1-17)

Permitted aids:
No aids permitted. [Except PC to get access to this document in FLOWhandin + Excel file where
answers must be put, and use of FLOWhandin to upload answers.]

Information regarding attachments:

Multiple Choice exam: insert answers (A-D) in attached Excel sheet, and upload in Excel file format.

Comments:
Your answers to Problem 1-20 (A-D) must be inserted into Excel sheet, and uploaded as .xlxs file.

Exam assignment:
Each of 20 Problems has optional answers A, B, C, D. Make your choice of A-D.
If the exam isn’t in digital form, please select the type of examination paper:
Spreadsheets ☐ Line sheets ☐

THE CANDIDATE MUST VERIFY THAT THE TASK SET IS COMPLETE ON THEIR OWN.

1/1
 Exam: FM1015 Modelling of Dynamic Systems

Date: Friday, December 1, 2023

Form of exam: This is a Digital Exam of type Multiple Choice. Each exam question
(denoted “problems” in the text below) is followed by 4 possible answers (“choice” in
the tables related to questions). Some details about your answers and how they will be
assessed:

• A single correct choice gives full score on a problem; any other choice (an incorrect
choice, no choice, several choices) gives zero score on the particular problem.

• All problems are given the same weight.

• Most problems are posed so that it is necessary to do background calculations to

choose the correct answer (A–D).

• The background calculations shall not be submitted — the only submission in the
Digital Exam are the choices (A–D) in each problem.

• You need to insert the answers in the attached Excel sheet, and upload this Excel
sheet which includes your answers, in the exam system.

• The file you upload must be in Excel file format, i.e., *.xlsx format. (Reason:
the files will be automatically parsed for grading.)

• A total score of 0-39% gives grade F (fail), score 40-49% gives grade E, score 50-59%
gives grade D, score 60-79% gives grade C, score 80-89% gives grade B, and score
90-100% gives grade A.

Aids: No aids allowed, except the necessary use of PC for using WISEflow, and the use
of Excel for filling in answers.

This exam contains questions (denoted problems) in the Problems part. The Problems
part consists of 20 problems, each with 4 alternative choices (A–D).
Contact person for questions during the exam:
Name: Bernt Lie
Phone: 41 80 77 44
Room: B-257b

2
 Figure 1: Fuel Cell with acid electrolyte and hydrogen fuel.

Problems
Hydrogen (H2 ) and oxygen (O2 ), when mixed together, form an explosive mixture1 which
reacts quicly and violently. If mixed with other, inert gases such as nitrogen (e.g., in
air), the mixture is less explosive, and may become harmless when the partial pressure of
hydrogen is below a certain threshold value.
It is of interest to try to control this violent reaction so that we can utilize the energy in
the mixture. This can be done by putting the reacting gases (H2 , O2 ) in different volumes,
separated by a membrane. Such a membrane may be physically damaged if the total
pressure difference across the membrane is too large (for the fuel cell discussed below, the
pressure difference should be less than ca. 0.2 bar). Also, if the partial pressure difference
for one gas is large across the membrane, this particular gas may “leak” (permeate) through
the membrane.
A fuel cell can be used to take out the energy of hydrogen and oxygen: a certain type
of membrane that under normal conditions blocks permeation of any species except proton
(H+ ) can be used (PEM = Proton Exchange Membrane). Then some half-reactions occur
on each side of the membrane, accelerated by a catalyst. The fuel cell includes a system
for feeding of reactants (H2 , O2 ) to each side of the membrane, and for removing the
product (H2 O). Note that normal PEM systems require the membrane to be humid (=
contain some water) in order to work. Because of this, some water is normally added to
both sides of the membrane.

Problem 1. Reactions in Fuel Cell.

Consider a Fuel Cell where hydrogen fuel (H2 ) is added to the so-called anode volume,
while oxygen (O2 ) is added to the cathode volume (typically via air), where an (acid
electrolyte) membrane separates the two volumes and only allows for transport of proton
(H+ ) from the anode side to the cathode side, Fig. 1.
At the anode interface between gas, some catalyst, and the proton conducting mem-
brane (the so-called triple-point junction), hydrogen H2 reacts in a half-cell reaction as
follows:
H2 → 2H+ + 2e− . (1)
1
A H2 + O2 mixture is sometimes known as “knallgas” — taken from German and Scandinavian
languages = bang-gas

3
Proton H+ can move through the membrane towards the cathode interface. The electron
e− , on the other hand, moves through the external electric circuit and eventually ends up
at the cathode.
At the cathode triple-point junction, oxygen O2 reacts with the proton coming through
the electrolyte and the electron coming via the external electric circuit in the following
half-cell reaction:
O2 + 4H+ + 4e− → 2H2 O. (2)
Combining the two half-cell reactions, the overall reaction is sometimes written as

2H2 + O2 → 2H2 O (3)

where the product is water, but other possibilities exist.

Let the species be ordered according to the vector (H2 , O2 , H2 O). It is of interest to
formulate the stoichiometric matrix ν for the overall reaction in Eq. 3, as well as the
positive stoichiometric parameter for the electron, νe− (Note: νe− > 0).
Problem 1: Which of the following pairs of ν and νe− is incorrect? Choice
A: ν = 1, − 2 , 1
and νe− = 2
1

B: ν = −1, − 12 , 1
and νe− = 2

C: ν = − 12 , − 41 , 12 and νe− = 1
D: ν = (−2, −1, 2) and νe− = 4
▲

Problem 2. Equilibrium voltage in Fuel Cell

The so-called chemical potential µj of species j in a mixture is given as molar Gibbs’

energy G̃◦j of species j at standard state plus a term RT ·ln aj where R is the gas constant,
T is absolute temperature (i.e., in K), and aj is the activity. With µj ≡ G̃j , we have

G̃j = G̃◦j + RT · ln aj . (4)

For ideal gas,

pj
aj = (5)
p◦
where pj is the partial pressure of species j, while p◦ is a standard state scaling pressure
(normally either 1 atm or 1 bar). In the liquid case, for an ideal solution

aj = x j (6)

where xj is the mole fraction xj = nj /n of the liquid.

Next, the molar Gibbs energy of reaction ∆r G̃ (at arbitrary T , pj ) is given as
X
∆r G̃ = νj G̃j , (7)
j

while the standard molar Gibbs energy of reaction (i.e., at arbitrary T and standard
pressure p◦ ) is denoted ∆r G̃◦ . The equilibrium voltage from the cathode side (positive)
to the anode side (negative) is then

∆r G̃
un = − (8)
F νe−

4
where F is Faraday’s number.
Problem 2: Equilibrium voltage — which expression is incorrect? Choice
 
ν
A: un = −∆r G̃◦ − RT j ln aj j / (F νe− )
P
 Q 
νj
B: un = −∆r G̃◦ − RT ln a
j j / (F νe− )
 
ν
C: un = −∆r G̃◦ − RT j ln aj j / (F νe− )
Q
 Q 
νj /νe−
D: un = −∆r G̃◦ /νe− − RT ln a
j j /F
▲

Problem 3. Gas vs. liquid form of water product.

In the group project, it was assumed that water was present in the cathode volume at
saturation pressure, which is a pure function of temperature:
pcH2 O = psat
H2 O (T ) .

Saturation pressure for water may be given by the Antoine equation,

 sat 
pH 2 O BH2 O
log10 = A H 2 O − (9)
p◦ T + CH2 O − 273.15 K
where T is absolute temperature (K) and
AH2 O = 5.11564
BH2 O = 1687.537 K
CH2 O = 230.17 K.

This implies that in the expression for G̃H2 O in Eq. 4, the water activity was set to
psat
acH2 O = H2◦O .
p
In much of the relevant published literature, instead it is assumed that water is present
in the cathode volume as condensed water, i.e., water in liquid form. Because the other
involved species (oxygen, nitrogen, etc.) are (almost) completely in gas form, this implies
that the liquid phase is pure water, i.e., xcH2 O ≈ 1, and the activity of water is unity:
acH2 O = xcH2 O = 1.
An interesting question is: can both assumptions on cathode water state lead to the
same equilibrium voltage, or is the assumption of liquid water leading to acH2 O = 1 just
introduced for convenience?
To help us assess this, under the assumption of constant heat capacity c̃p,j , we have:
 
T
◦ ◦ ◦ ◦ ◦
G̃j = ∆f G̃j − ∆f S̃j (T − T ) + c̃p,j T − T − T · ln ◦ . (10)
T
For water in gas phase:
∆f G̃H2 O,g = −228.572 kJ/mol
∆f S̃H2 O,g = 188.7 J/mol K
c̃p,H2 O,g = 33.58 J/mol K

5
 Figure 2: Hydrogen Fuel Cell standard state equilibrium voltage u◦ .

while for water in liquid phase:

∆f G̃H2 O,ℓ = −237.129 kJ/mol

∆f S̃H2 O,ℓ = 69.91 J/mol K
c̃p,H2 O,ℓ = 75.29 J/mol K.

Problem 3: Assuming H2 O product in gas or liquid form — which of Choice

the following statements is correct?
A: The assumption of H2 O in liquid form is only introduced for
convenience, and leads to a slightly different equilibrium voltage un
compared to assuming gas.
B: Whether cathode H2 O is in liquid or gas form, in both cases we get
identical equilibrium voltage un .
C: The assumption of H2 O in gas or liquid form leads to different
equilibrium voltages un . However, if we take into account the energy
balance and condensation of the H2 O in gas form, this will modify the
temperature and will gradually lead to the same equilibrium voltage un .
D: The H2 O content on the anode side will also affect the equilibrium
voltage un , and needs to be taken into account to get a correct
equilibrium voltage un .
▲

Problem 4. Pressure effect on equilibrium voltage.

Assuming that water is produced in gas phase, the standard state equilibrium voltage
u◦ is as in Fig. 2.
If any of the partial pressures pH2 , pO2 , or pH2 O deviate from the standard state p◦ =
1 bar, there will be a correction term to the equilibrium voltage as indicated in Problem 2.

6
We will assume that cathode water is at the saturation pressure (water is produced at the
cathode), which for T = 80 ◦ C is pcH2 O ≈ psat
H2 O ≈ 0.4 bar, while the anode water pressure
is paH2 O ≈ 21 psat
H2 O ≈ 0.2 bar (the membrane must contain some water to work, and some
water is normally added to the anode volume).
For PEM fuel cells, the total anode pressure pa and total cathode pressure pc should
be equal, and not deviate by more than ∆pac = |pa − pc | ≤ 0.2 bar to avoid mechanical
damage to the membrane. Typical electrode pressures pa , pc lie in the range of [1, 5] bar.
Assume that we operate the PEM fuel cell with electrode pressures pa = pc = 5 bar,
that we feed the anode with pure hydrogen (+ some water), while we feed the cathode
with air (+ some water) where the oxygen mole fraction is approximately 0.2. Assume
that we feed sufficient air to the cathode side to maintain mole fraction oxygen of 0.2 in
the cathode volume.
What is the approximate change up in equilibrium voltage away from the standard
state voltage (u◦ ; Fig. 2) due to pressure conditions as indicated above? Use gas constant
R = 8.31 J/(mol · K), Faraday’s constant F = 9.648 · 104 C/mol, and assume operation at
T given by 80 ◦ C.
Problem 4: Approximate change up in equilibrium voltage due to Choice
pressure conditions specified above is
A: up ≈ 0.12 V
B: up ≈ −33 mV
C: up ≈ 0.31 V
D: up ≈ 37 mV
▲

Problem 5. Transfer function for Activation voltage loss.

On the convoluted interface between a charged eletrode and the electrolyte, there will
be a concentration of ions in the electrolyte with opposite charge as to that of the electrode.
This creates a capacitance, and there will be such a capacitance on both electrodes. In
general, capacitance is given by
Adl
Cdl = εdl (11)
δdl
where εdl is the permittivity of the space between the ions in the double layer, Adl is the
area of the double layer, and δdl is the separation distance between the ions in the double
layer. Due to large area Adl and short distance δdl , the capacitance Cdl can be quite large,
and in the order of a few Farads. Theoretical expressions exist for Cdl .
In the group project, activation voltage loss was only considered on one electrode.
Figure 3 shows a linearized electric circuit equivalent of the membrane kinetics, with
external current i, activation resistances Raa , Rac at the anode and cathode electrode,
respectively (given by linearizing the Butler-Volmer equation, or similar), double layer
capacitors Cdla
, Cdl
c
at the anode and cathode electrodes, respectively, and membrane
resistance Rm . Figure 3 shows the relevant currents i (green color), iar , iadl , icr , and icdl (red
color), ans well as relevant voltages uaa , ur , and uca (blue color). Because the depicted
circuit in Fig. 3 is linear, the order of the blocks is irrelevant, and we will neglect the
membrane resistor Rm in the sequel; we will set the total activation voltage ua equal to
the sum of activation voltage losses on the anode (uaa ) and the cathode (uca ). Because
the course syllabus for course FM1015 Modelling of Dynamic Systems does not include
modeling of electric circuits, the model for the activation voltages is given below for

7
 Figure 3: Equivalent linear electric model for kinetics of membrane.

convenience:
duaa
a
Cdl = iadl (12)
dt
uaa = Raa iar (13)
i = iar + iadl (14)
c
c dua
Cdl = icdl (15)
dt
uca = Rac iar (16)
i = icr + icdl (17)
ua = uaa + uca . (18)

It is of interest to find the transfer function h (s) from external current i to total
activation voltage loss ua , i.e.,

ua (s) = h (s) · i (s) .

Problem 5: Which expression for the transfer function from i (s) to Choice
ua (s), denoted h (s), is correct?
Raa Rac
A: h (s) = a s
1+Raa Cdl
· c s
1+Rac Cdl
1+ (C a +Cdlc )s
B: h (s) = (Raa + Rac ) 1+Ra C a sdl 1+R
( a dl )( ac Cdlc s)
Raa Rac +(Raa +Rac )s
C: h (s) = a
(Cdl + Cdlc
) 1+R
( aa Cdla s)(1+Rac Cdlc s)
Raa Rac
D: h (s) = a s + 1+Rc C c s
1+Raa Cdl a dl
▲

Problem 6. Double layer capacitance.

Using a more accurate but more complex model than the Butler-Volmer equation,
step changes in external current i leads to the responses in activation voltage losses uaa
(Wang-HOR) and uca (Wang-ORR) as depicted in Fig. 4. In particular, observe (i) the
differences in activation voltage losses at the electrodes, and (ii) the differences in time
constants.
The transfer function in Problem 5 has two time constants, Ta = Raa Cdl a
and Tc =
Ra Cdl at the anode and cathode, respectively. In the sequel, assume for simplicity that
c c
a
Cdl c
= Cdl = Cdl .
Given the large difference in gain and time constants for the transfer functions, propose
a simplified transfer function hsimp (s) for the activation voltage loss, and suggest the

8
Figure 4: Step responses in activation voltages at the anode (uaa ) and cathode (uca ), re-
spectively.

9
corresponding simplified capacitor Csimp to use.
Problem 6: Simplified transfer function hsimp (s) with overall capacitor Choice
Csimp — which of the following pairs is correct?
A: hsimp (s) = Rac / (1 + Rac Csimp s), Csimp = Cdl /2
B: hsimp (s) = Raa / (1 + Raa Csimp s), Csimp = Cdl /2
C: hsimp (s) = Rac / (1 + Rac Csimp s), Csimp = Cdl
D: hsimp (s) = Raa / (1 + Raa Csimp s), Csimp = Cdl
▲

Problem 7. Electric time constants.

Based on the two plots in Fig. 4, what are approximate values for time constants in
the two plots (see ua in the Wang-HOR plot and ua in the Wang-ORR plot)?
Problem 7: Approximate electric time constants for the anode and the Choice
cathode, which pair is reasonable?
A: Ta ≈ 0.5 ms, Tc ≈ 0.1 s
B: Ta ≈ 2 ms, Tc ≈ 10 s
C: Ta ≈ 0.01 s, Tc ≈ 0.5 s
D: Ta ≈ 100 ms, Tc ≈ 2 s
▲

Problem 8. Anode species balance.

Figure 5 illustrates the most common gas subsystem for a fuel cell: hydrogen is supplied
from a high pressure storage system via an inlet anode pressure controller, alternatively a
flow controller. Effluent hydrogen from the anode volume is returned to the inlet via a fan
(low anode pressure) or compressor (high anode pressure). The purge valve is normally
closed, but can be opened for a short time to purge out accumulated impurities in the
hydrogen gas.
Oxygen is normally supplied via atmospheric air (approximately 20% oxygen, 80%
nitrogen), either by fan (low cathode pressure) or compressor (high cathode pressure).
The fan/compressor flow rate can be controlled (speed control), and the effluent volume is
controlled to pressure pec (so-called back-pressure controller). If pure oxygen were supplied
from a high pressure storage tank (conceptually possible), the oxygen feed system would
have to be similar to the hydrogen supply system in the figure.
Assume that the well mixed anode volume Va contains only H2 O + H2 , and that the
H2 O (T ) where the saturation pressure is given by
water content is fixed to be paH2 O = 21 psat
Antoine’s model as in Problem. 3.
Assuming ideal gas and constant temperature, and known influent ṅiH2 , the hydrogen
species balance leads to the following model for hydrogen pressure:
Problem 8: Dynamic model for hydrogen pressure under given Choice
conditions — which model is correct?
dpH2 pH2 e
A: dt
= V̇ − RT
Va a
ṅi − RT
Va H 2
ṅg
Va H2
dpH2 p
B: dt
= RT i
ṅ − VHa2 V̇ae + RT
Va H 2
ṅg
Va H2
dpH2 p
C: dt
= RT i
ṅ − VHa2 V̇ae
Va H 2
dpH2 pH2 i
D: dt
= Va i
V̇ + RT
Va a
ṅH2 − RT ṅg
Va H 2
▲

10
 Figure 5: Gas subsystem for membrane.

Problem 9. Anode effluent flow rate.

Hydrogen is a precious fuel, and should not be wasted. For the gas subsystem in
Fig. 5, assume for simplicity that ṅiH2 is a known/controlled function of time. Assume
furthermore that the inlet pressure pia to the anode volume is known and constant, and
that a fan or compressor manages to keep a constant pressure increase in the recycle flow
so that ∆prc
a = pa − pa is constant. From Bernoulli’s steady state energy balance, a typical
i e

valve model for the effluent flow rate would be

s
∆pea
V̇ae = Ka · . (19)
p◦

Consider the term V̇ae in Problem 8. Modify the model in Eq. 19 so that (i) the flow
unidirectional (i.e., only can flow out), and (ii) give expressions for ∆pea .
Problem 9: Modified anode effluent flow rate V̇aa — which set of Choice
expressions is correct?
A: V̇ae = Ka · pmax (∆pea , 0) /p◦ ; ∆pea = pa − pia ; pa = pH2 + paH2 O
p

B: V̇ae = Ka · pmin (∆pea , 0) /p◦ ; ∆pea = pa − pia ; pa = pH2 + paH2 O

C: V̇ae = Ka · max (∆pea , 0) /p◦ ; ∆pea = pa − pea ; pa = pH2 + paH2 O ;
pea = pia − ∆p rc
a
D: V̇ae = Ka · min (∆pea , 0) /p◦ ; ∆pea = pa − pea ; pa = pH2 + paH2 O ;
p

pea = pia − ∆prc

a
▲

Problem 10. Anode hydrogen generation.

Consider the term ṅgH2 in Problem 8. With ν = (νH2 , νO2 , νH2 O ), what is the correct
expression for ṅgH2 for a single fuel cell ?

11
 Problem 10: Which of the following expressions for hydrogen Choice
generation ṅgH2 is correct?
A: ṅgH2 = νH2 · i/F
B: ṅgH2 = −νH2 · i/F
C: ṅgH2 = νH2 · i/ (F νe− )
D: ṅgH2 = νH2 · i/F
▲

Problem 11. Anode diffusion.

There is a possible gas pressure drop from the anode volume to the surface of the
catalyst/membrane due to diffusion resistance through the Gas Diffusion Layer (Fig. 5),
etc. With only water vapor and hydrogen in the anode volume, a possible expression for
the resulting pressure drop in hydrogen is

pas pa − paH2 O · exp (fa (T ) · i′′ )

H2
= (20)
paH2 pa − paH2 O

where superscripts a and as refer to bulk anode volume and anode surface (i.e., at the cata-
H2 O (T ), fa (T ) > 0 is a known function of temperature
lyst/membrane surface), paH2 O = 21 psat
(involving diffusion coefficients, etc.), while i′′ is the current density in the membrane.
What modifications to the model is required if this diffusion term in Eq. 20 is impor-
tant?
Problem 11: Which of the following statements about the effect of Choice
anode diffusion is correct?
A: Diffusion drop in pressure has no effect on the model: a reduction in
hydrogen pressure results in an increase in the water pressure, and
these changes cancel out.
B: The anode volume model is as found above in Problems 8–10. However,
the pressure pasH2 should be used in the electric model to compute the
fuel cell voltage.
C: The partial pressure of hydrogen in Problem 8 must be changed to pas H2 .
D: It is not necessary to modify the model because the controller for
hydrogen feed will cancel out the effect.
▲

Problem 12. Cathode oxygen balance.

Assume that the well mixed cathode volume Vc contains H2 O + O2 + N2 , and that
the water content is fixed to be pcH2 O = psat
H2 O (T ) where the saturation pressure is given
by Antoine’s model as in Problem. 3.
Assuming ideal gas and constant temperature, and known influent flow rate V̇ci of
atmospheric air at pressure patm (containing oxygen and nitrogen; mole fraction xair
O2 ≈ 0.2)
controlled by the fan/compressor, the oxygen species balance leads to the following model
for oxygen pressure:

12
 Problem 12: Dynamic model for oxygen pressure under given Choice
conditions — which model is correct?
dpO2 xair
O patm pO2 e
A: dt
= 2
Vc
V̇ci − Vc c
V̇ − RT g
ṅ
Vc O 2
dpO2 pO2 i xair p
B:
atm RT g
dt
= Vc c
V̇ − O2Vc V̇ce − ṅ
Vc O 2
dpO2 xair
O2 patm p
C: dt
= Vc
V̇ci − VOc2 V̇ce + RT g
ṅ
Vc O2
dpO2 xair
O2 patm p
D: dt
= Vc
V̇ci − VOc2 V̇ce + RT
Vc
(ṅH2 O − ṅN2 )
▲
Problem 13. Cathode nitrogen balance.
With the assumptions in Problem 12, the nitrogen species balance leads to the follow-
ing model for nitrogen pressure:
Problem 13: Dynamic model for nitrogen pressure under given Choice
conditions — which model is simplest and correct?
dpN2 xair
O patm p N2 e
A: dt
= 2
Vc
V̇ci − V̇
Vc c
dpN2 xair
O patm p
B: V̇ci − VNc2 V̇ce − RT ṅgH2 O − ṅgO2


dt
= 2
Vc Vc
dpN2 (1−xair
O2 )patm p
C: dt
= Vc
V̇ci − VNc2 V̇ce
dpN2 (1−xair
O2 )patm p
D: dt
= Vc
V̇ci − VNc2 V̇ce + RT ṅg
Vc N 2
▲
Problem 14. Cathode effluent flow rate.
We will assume that V̇ci is a known function of time. From Bernoulli’s steady state
energy balance, a typical valve model for the effluent flow rate would be
s
∆pec
V̇ce = Kc · . (21)
p◦
Consider the term V̇ee in Problems 12–13. Modify the model in Eq. 21 so that (i) the
flow unidirectional (i.e., only can flow out), and (ii) give expressions for ∆pec .
Problem 14: Modified cathode effluent flow rate V̇ca — which set of Choice
expressions is correct?
A: V̇ce = Kc · pmax (∆pec , 0) /p◦ ; ∆pec = pc − patm ; pc = pO2 + pN2 + pcH2 O
p

B: V̇ce = Kc · pmin (∆pec , 0) /p◦ ; ∆pec = pc − patm ; pc = pO2 + pN2 + pcH2 O

C: V̇ce = Kc · pmax (∆pec , 0) /p◦ ; ∆pec = pc − patm ; pc = pO2 + pN2
D: V̇ce = Kc · min (∆pec , 0) /p◦ ; ∆pec = pc − patm ; pc = pO2 + pN2
▲
Problem 15. Cathode generation.
Consider the term ṅgO2 in Problem 12. With ν = (νH2 , νO2 , νH2 O ), what is the correct
expression for ṅgO2 for a single fuel cell ?.
Problem 15: Which of the following expressions for oxygen generation Choice
ṅgO2 is correct?
A: ṅgO2 = νO2 · i/ (F νe− )
B: ṅgO2 = −νO2 · i/ (F νe− )
C: ṅgO2 = νO2 · i/F
D: ṅgO2 = −νO2 · i/F

13
 ▲

Problem 16. Cathode diffusion/current limitation.

There is a possible gas pressure drop from the cathode volume to the surface of the
catalyst/membrane due to diffusion resistance through the Gas Diffusion Layer (Fig. 5),
etc. With water vapor, oxygen, and nitrogen in the cathode volume, a possible expression
for the resulting pressure drop in oxygen pressure is
pcs pc − pcH2 O − pcN2 exp (fc (T ) · i′′ )
O2
= (22)
pcO2 pc − pcH2 O − pcN2

where superscripts c and cs refer to bulk cathode volume and surface (i.e., at the cata-
lyst/membrane surface), pcH2 O = psatH2 O (T ), fc (T ) > 0 is a known function of temperature
(involving diffusion coefficients, etc.), while i′′ is the current density in the membrane.
In the literature, it is common to write this reduction in partial pressure as
pcs
O2 i′′
≈ 1 −
pcO2 i′′ℓ

where i′′ℓ is a limiting current density.

Find an expression for i′′ℓ based on (linearization of) Eq. 22.
Problem 16: Which of the following expressions for i′′ℓ is consistent Choice
with Eq. 22?
pcN
A: i′′ℓ = pcO
2
· fc (T )
2
pcN
B: i′′ℓ = pcO
2
· 1
fc (T )
2
pcO
C: i′′ℓ = pcN
2
· fc (T )
2
pcO
D: i′′ℓ = pcN
2
· 1
fc (T )
2
▲

Problem 17. Maximal Fuel Cell efficiency.

Fuel Cell efficiency η is typically found as the produced electric power divided by the
enthalpy of reaction,
u·i
η= .
∆r H̃ · ṅgH2 O
Here, current i is given as
i = −F · ṅe−
where ṅe− is the molar flow rate of electrons, and the generation of product ṅgH2 O is relatd
to ṅe− as
ṅe
ṅgH2 O = .
νe −
In the limit i → 0+ , u = un where Nernst voltage un is as given in Problem 2. Denote the
resulting maximal efficiency by η0 [zero current],
un · i
η0 = .
∆r H̃ · ṅgH2 O

14
Figure 6: Typical external voltage of fuel cell: (i) using Butler-Volmer activation voltage,
(ii)-(iii) using a more realistic activation voltage due to Wang, at 1 and 5 bar, respectively,
(iv) with zero current/Nernst voltage (blue, dashed) .

For simplicity, we consider a reference temperature of 25 ◦ C, which gives

∆r G̃ = ∆r G̃◦ = 1 · ∆f G̃H2 O = −228.572 kJ/mol

while
∆r H̃ ◦ = −241.785 kJ/mol.

Based upon the above considerations, what is the theoretical maximal efficiency of
a fuel cell at standard state (i.e., pressures pj = p◦ ) and with (extremely) low current
throughput i, when we assume that both reactants and product are in gas phase?
Problem 17: Approximate fuel cell efficiency under the above Choice
considerations — which maximal efficiency η0 is correct?
A: η0 ≈ 0.4
B: η0 ≈ 0.95
C: η0 ≈ 0.83
D: η0 ≈ 0.28
▲

Problem 18. Practical Fuel Cell efficiency.

Figure 6 shows how the external voltage u in the efficiency expression of Problem 17
varies with current density i′′ (and thus with i).
Based on the above, how much is the practical efficiency ηi′′ of a fuel cell reduced
compared to the maximal efficiency, say at i′′ = 0.5 A/cm2 ?

15
 Problem 18: Approximate fuel cell efficiency under the above Choice
considerations — which efficiency η0.5 is correct?
A: η0.5 ≈ η0
B: η0.5 ≈ 0.49 · η0
C: η0.5 ≈ 0.61 · η0
D: η0.5 ≈ 0.83 · η0
▲

Problem 19. Consequence of efficiency loss.

In Problems 17–18, we have seen that a fuel cell does not have 100% conversion of
reactants into electric power.
What is the consequence of imperfect efficiency η?
Problem 19: Which of the following statements is correct? Choice
A: Both with i′′ → 0+ (η0 ) and for finite i′′ (ηi′′ ) will the efficiency loss lead
to heating of the fuel cell. It is necessary to actively control the cooling
of the fuel cell to (i) avoid overheating, and (ii) manage the humidity of
the membrane.
B: The efficiency loss with i′′ → 0+ (η0 ) is an inherent loss, that has no
effect on the temperature of the system; there can be no loss when
there is no current.
C: The efficiency loss for finite i′′ > 0 (ηi′′ ) leads to heating of the fuel cell,
and some cooling is necessary. However, this does not influence the
humidity of the system.
D: The efficiency loss for finite i′′ > 0 (ηi′′ ) leads to heating of the fuel cell,
but the heating is automatically compensated for by the
humidity/water balance.
▲

Problem 20. Choice of operating conditions.

As seen from the expression for pressure correction to the standard state equilibrium
voltage, increasing the pressure (if possible) should give a higher equilibrium voltage, and
hence more power out of the fuel cell.
The Butler-Volmer expression for activation voltage used in the group project was
pressure independent; however, a more accurate (and complex) expression for the cathode
activation, Fig. 7, shows that the voltage loss ua for a given current density i′′ decreases
with pressure.
In summary: the equilibrium voltage increases with higher electrode pressures, and
the activation losses are reduced when current starts to flow. This means that there is a
double advantage in operating at higher pressure.
Which of the following arguments for choice of operating pressures has potential safety
issues?

16
 Figure 7: Pressure dependence of activation voltage loss relation.

Problem 20: Choice of operating pressures for fuel cell — which Choice
argument has safety issues?
A: Operating at a high pressure requires power to compress the reactants
(hydrogen, oxygen). The chosen pressure conditions must balance the
increased electric power produced vs. the increased power needed to
compress the reactants.
B: Increased pressures in the electrode volumes must be balanced against
the possibility of gas permeation through the membrane; higher
pressure may require thicker membrane.
C: Operating at a high pressure may require more expensive equipment
(thicker membrane to avoid gas permeation, also leading to higher
resistive voltage loss ur ), more solid build, etc., and the increased
design/purchase cost must be balanced against increased power
production.
D: The best way to increase the power is to increase the Nernst voltage by
feeding oxygen to the cathode instead of air. This way, the partial
pressure of oxygen can be made much higher while still having the same
electrode pressures, pa = pc .
▲

— END OF EXAM —

17