Preprints of the 8th IFAC Symposium on Advanced Control of Chemical Processes

The International Federation of Automatic Control

Furama Riverfront, Singapore, July 10-13, 2012

Optimization of a methane autothermal reforming-based hydrogen

production system with low CO2 emissions
Wei Wu*,a and Chutima Tungpanututhb
a

Department of Chemical Engineering, National Cheng Kung University, Tainan 70101, Taiwan, R.O.C.
Department of Chemical and Materials Engineering, National Yunlin University of Science and Technology, Douliou, Yunlin
64002, Taiwan, R.O.C.
Abstract: The aim of this work is to develop a hydrogen production system with respect to saving energy
and reducing carbon dioxide. Methane and carbon dioxide are major feedstocks of an autothermal
reformer (ATR) and a CO2 reformer, respectively. The waste heat recovery technique is applied to build a
stand-alone ideal heat-integrated system. The proposed system configuration and optimal operating
conditions are verified with the Aspen HYSYS simulator.
Keywords: Autothermal reforming; CO2 reformer; Hydrogen production; Heat recovery; Optimization

Ersoz et al. [13] provided a simulation study for a fuel cell

system combined with a methane ATR-based hydrogen
production process. Furthermore, the optimization of the
methane ATR process with regard to the conflict between the
increase in hydrogen yield and the reduction in carbon
monoxide has also been investigated [14, 15]. Using a similar
approach, Silva et al. [16] undertook the simulation and
optimization of a glycerol ATR-based hydrogen production
process with the assistance of Aspen HYSYS.

1. Introduction
About 95% of the hydrogen that was produced in the U.S.
used a reforming process with natural gas as the feedstock.
The conventional process primarily consists of two steps.
First, the reformation of the feedstock with high temperature
steam to obtain a syngas and, second, a water gas shift (WGS)
reaction that produces hydrogen, carbon dioxide, and a few
residual materials [1]. As for fuel cell applications, the
pressure swing adsorption (PSA) process is connected to the
WGS reactor such that the purity of hydrogen can reach
almost 99.99+%, and the waste gas of PSA may flow into a
preferential oxidation (PROX) reactor to produce the high
temperature flue gas [2]. Moreover, the modeling, simulation
and optimization of a class of hydrogen production processes
using the Aspen HYSYS process simulator have been
studied [3, 4].

2. Description of Major Process Units

There are five major process units in our methane
autothermal reforming-based hydrogen production process.
Three two-phase streams of methane, water and air (oxygen)
are well mixed at room temperature with two molar ratios of
water-to-methane (H 2 O/CH 4 ) and oxygen-to-methane

(O 2 /CH 4 ) . The mixed stream is pre-heated by a heater

equipment and flows into the autothermal reforming (ATR)
reactor at the prescribed inlet temperature, TATR,in . Assuming

In general, methane reforming processes constantly generate

a large amount of carbon dioxide from the steam methane
reformer (SMR), WGS reactor and PROX reactor. The CO2
capture technologies including MEA absorption and
membrane processes can be devoted to reducing greenhouse
gas emissions [5, 6], but the capture and storage of CO2 need
additional electricity. Nord et al. [7] used the pre-combustion
CO2 capture technique to effectively reduce greenhouse gas
emissions. Recently, Fan et al. [8] utilized greenhouse gases
through the CO2 reforming process with specific catalysts to
improve hourly space velocity and hydrogen production, and
Fan et al. [9] studied the optimization of hydrogen
production from CO2 reforming of methane via experiments.
With regard to methane reforming-based hydrogen
production processes, a combination of partial oxidation and
steam reforming of methane in an autothermal reforming
(ATR) reactor has recently become more popular. The
modeling, kinetic analysis, and simulation of ATR of
methane for prescribed reactors have been studied [10-12].
IFAC, 2012. All rights reserved.

that the system is an adiabatic reactor with packed catalysts,

the major reactions are described as follows.
CH + 2O CO + 2H O ( ), H10 =-802.3kJ/mol (1)
CH + H O CO + 3H ( ), H 0 = 206.2 kJ/mol
2

(2)

CO + H O CO + H ( ),

(3)

H30= - 41.2 kJ/mol

CH + 2H O CO + 4H ( ), H 40 =165.0 kJ/mol (4)

and the kinetic rate equations for

661

are expressed as [10],

8th IFAC Symposium on Advanced Control of Chemical Processes

Furama Riverfront, Singapore, July 10-13, 2012

where
=

.
.

&

!"# !#

!"$ !#

&
'

*+

'&

*+

and

K e5 = exp(30782/ T + 42.97) . The CO2 reformer produces

syngas which is similar to the steam reforming of methane,
i.e. the large amount of CO in the outlet stream of the CO2
reformer is reproduced. Moreover, another high temperature
water gas shift (HTWGS) reactor with packed catalyst is
connected in order to convert components of CO and H2O in
the stream into CO2 and H2. The 600K operating temperature
of the HTWGS reactor is lower than that of the CO2 reformer,
so another heat exchanger for heat removal (i.e. a cooler) is
added to adjust the inlet temperature of HTWGS reactor,

(5)

k5 = 8.71 102 exp(23.7/ RT )

(6)

TWGS ,in . Moreover, the kinetic model of HTWGS is

(7)

expressed by,
PCO2 PH2
95 1.1 0.53
rWGS = 7.79 1014 exp
PCO PH2O 1

K e2 PCO PH2O
RT

=
where
/#

&.

!"# !#

'

*+

(8)

(11)

,- = 1 + /"$ !"$ + /# !# + /"# !"# +

Afterward the outlet stream of the HTWGS reactor flows into

the pressure swing adsorption (PSA) unit to produce high
purity hydrogen (99.95+%) by using a solid adsorbent, e.g.,
activated carbon can be separated. Assuming that the PSA
process is an isothermal one, the outlet composition of the
PSA is expressed as follows

. The reaction rate constant 01 = 021 exp(61 /89),

j=1, ,4, the adsorption constant for the combustion of

"
methane /:" = /2:
exp(<:" /89) , i=CH4, O2, and the
adsorption constant of species i /: = /2: exp(<: /89),
i=CO, H2, CH4, H2O. Notably, the rate equation is based on
the
temperature
dependence
of
the
Langmuir-Hinshelwood-Hougen-Watson (LHHW) model.
All kinetic data including kinetic parameters ( 021 , 61 ) ,
equilibrium constants ( /> , /> & , /> ) and adsorption
constants (/21 , <: ) are taken from the reference cited [10].
Since the outlet temperature of the ATR reactor could
be over 850K and the stream contains a high percentage of
carbon dioxide, a CO2 reformer aims to reduce the
greenhouse gas emission and increase hydrogen production
accordingly. In our approach, the outlet stream of the ATR
reactor needs to be mixed with the second feed flow of
methane ?CH |ABBC at room temperature, and the well-mixed
stream flows into the CO2 reformer at the prescribed inlet
temperature, TCO R,in . Assuming that the system is a

yH2 ,PSA ,out = S PSA

y i ,PSA ,out = (1 S PSA )

i = H2, CH4, CO, CO2, H2O

H G

(13)

(14)

where S PSA (= 99.95%) represents the hydrogen purity from

the PSA. Moreover, the mass flow rates of H2 and CO2 at two
exits of the PSA unit are,

(9)

with the corresponding kinetic model

= 0D F"# F"$

i ,PSA ,in

and the outlet composition of PSA waste gas is shown by

y i ,PSA ,in
y i ,PSA ,out =
yi ,PSA ,in + (1 S PSA ) y H2 ,PSA ,in

i = CH4, CO, CO2, H2O

nonisothermal tubular reactor with packed catalysts, the

reversible endothermic reaction proceeds in a CO2 reformer
[9],
CH + CO 2CO + 2H ( D ), HDE = 205 kJ/mol

y i ,PSA ,in

(12)

(10)

662

8th IFAC Symposium on Advanced Control of Chemical Processes

Furama Riverfront, Singapore, July 10-13, 2012

CH4 |LIJJK
298K, 1atm

CH4 |MIJJK

298K, 1atm

Figure 1 Methane autothermal reforming-based hydrogen production system

FH2 = y H2 ,PSA ,out FPSA ,in

methane ?CH |ABBC and inlet temperature of the CO2

(15)

reformer , TCO R ,in , are adjustable variables of the CO2

2

FCO2 = y CO2 ,PSA ,out FPSA ,in

(16)

reformer, and inlet temperature of the HTWGS reactor ,

TWGS ,in , is an adjustable variable of the HTWGS reactor.

When the waste gas of the PSA flows into a preferential

oxidation (PROX) reactor, the high temperature flue gas is
produced. The kinetics of the PROX reactor are simplified as
1
Cat
C
(17)
CO+ O2
CO2 , HCO
= 282.99 kJ/mol
2

Moreover, the operating conditions are set by the ATR with

H2O/CH4 =1.5 , O2 /CH 4 = 0.55 and TATR ,in = 773K , the CO2

reformer with ?CH |ABBC = 9.92 kg/h at 25C, and the

HTWGS reactor with TWGS ,in = 664K . Moreover, Fig. 2(a)

and the corresponding reaction rate is shown by

0.5 0.1
O2 CO

rCO = 3.528 10 exp( 33092 / RT )P P

2

(18)

demonstrates that the proposed hydrogen production system

with the aid of the CO2 reformer can ensure a higher

Notably, the built-in property database in the Aspen HYSYS

hydrogen flow rate and lower carbon dioxide flow rate than

provides accurate thermodynamic data, which are calculated

with the conventional design in Fig. 2(b). Obviously, the CO2

with the Peng-Robinson equation of state.

reformer can contribute to increased hydrogen production by

Demonstration: In Fig. 1, the water-to-methane molar ratio

consuming carbon dioxide. However, the CO2 reformer in

(H2O/CH4 ) , oxygen-to-methane molar ratio (O 2 /CH 4 ) , inlet

this process design needs an additional methane feed flow

due to a very low composition of methane in the outlet of

temperature of the ATR reactor, TATR ,in , are denoted as

ATR.

adjustable variables of the ATR unit, the second feed flow of

(a)

663

8th IFAC Symposium on Advanced Control of Chemical Processes

Furama Riverfront, Singapore, July 10-13, 2012

(b)

Figure 2 Methane ATR-based hydrogen production system: (a) CO2 reformer with the second feed of methane (b)
without CO2 reformer
To achieve reduce the energy needed and the carbon dioxide
produced, a heat recovery method, shown in Fig. 3, is
developed. The flue gas produced from the PROX reactor
goes through two heat exchangers to rapidly heat the inlet
flow of the ATR reactor, and the recirculating streams go

Figure 3

through another two heat exchangers to cool the outlet flow

of the CO2 reformer. Finally, the outlet stream of the fourth
heat exchanger (HX4) reaches a temperature of 774 K and
this can be treated as the heat source for the CO2 reformer.

Methane ATR-based hydrogen production system with waste heat recovery

I, the optimization algorithm for maximizing hydrogen
selectivity is described as follows.
(20)
max S H

Notably, the original devices of heater and cooler shown in

Fig. 2(a) are replaced by four heat exchangers. Using the
Aspen HYSYS simulator, although the waste heat recovery
design may reduce the hydrogen yield by about 2.5%, but the
total benefits with regard to saving energy make this process
worthwhile.

ui

subject to
573 K

3. Process Optimization
Maximizing the hydrogen yield and minimizing carbon
dioxide emissions are the optimization goals for the hydrogen
production system; presented in this work. Two optimization
strategies are introduced. First, the hydrogen selectivity in
terms of the flow rates leaving the reactor is defined as
T2UV- WU2X # 2YZ

RS = (T2UV- WU2X 2W 1)

\]^

, j = CH4, H2O, CO, CO2, O2 (19)

TATR ,in 823 K

0.3

O2 /CH4

H2O/CH4

0.7
2

803 K

TCO2R ,in 983 K

573 K

TWGS ,in 773 K

(21)

and the corresponding model equations have been built in the

Aspen HYSYS. Notably,

This shows that if the hydrogen product increases or the

carbon dioxide decreases in the outlet flow of the HTWGS
reactor, then the value of RS definitely increases. For Case

ui TATR ,in ,O2 /CH4 ,H2O/CH4 ,TCO2R ,in ,CH4 Feed ,TWGS ,in
664

8th IFAC Symposium on Advanced Control of Chemical Processes

Furama Riverfront, Singapore, July 10-13, 2012

represents one of decision variables. The upper and lower

limits of physical constraints are based on the previous study
of operating conditions for each reactor.

4. Conclusions
The configuration proposed in this work is an energy-saving
process, since the ATR reactor reduces energy consumption
by up to 57.4%, and the waste heat recovery design can
completely take over functions of heater and cooler. The
proposed system can also reduce carbon dioxide emissions,
because the CO2 reformer can reduce these by 42.54% and
increase hydrogen yield by over 13.60%. It has been verified
that the ATR reactor plus CO2 reformer can provide greater
hydrogen produce than the conventional SMR process.
However, both HTWGS and PROX reactors may generate a
large amount of CO2 if the waste heat is not enough and the
concentration of CO is too high.

Remark 2: The hydrogen selectivity is denoted as a sole

objective function, but the maximization of hydrogen
selectivity could identify the operating scenarios for
maximizing the hydrogen yield and minimizing carbon
dioxide emissions simultaneously. Seven decision variables
are determined by solving this constrained optimization
problem. The optimal operating conditions are verified with
the Aspen HYSYS simulator. Second, the carbon dioxide
selectivity for the evaluation of the effect of carbon dioxide
reduction is shown by
T2UV- WU2X "$ 2YZ

R_` = (T2UV- WU2X 2W

)\]^

, k = CH4, H2, H2O, CO, O2 (22)

Acknowledgment

Notably, the total carbon dioxide reduction is relevant to

decrease the value of R_` . Thereby, Case II for the
optimization algorithm in regard to the minimization of R_`
is described as
(23)
min S CO
ui

The authors would like to thank the National Science Council

of the Republic of China for financially supporting this
research under Contract No. NSC 100-2211-E-006-264.

[1]

subject to the same constraints in Eq. (21).

Table 1 Comparisons of process before and after
optimization
Component/
before
Case I for
Case II for
selectivity
optimization
maximizing
minimizing
H2
CH4
H 2O
CO
CO2
O2

(kg/h)

SH2 (kg/h)

SCO2 (kg/h)

23.53
0.09
46.58
65.86
87.46
12.81

22.71
0.99
18.46
82.70
58.49
10.92

21.12
5.44
21.87
79.45
51.40
9.15

[2]

[3]

[4]

[5]

Remark 3: Similarly, the carbon dioxide selectivity is

another objective function. The minimization of carbon
dioxide selectivity could identify the operating scenarios for
maximizing the hydrogen yield and minimizing carbon
dioxide emissions simultaneously. The optimal operating
conditions can be found by solving the minimization
algorithm according to seven decision variables. Similarly,
the systems performance with regard to the reduction in
carbon dioxide emissions can be evaluated by the Aspen
HYSYS simulator. Furthermore, a comparison of the system
before and after using optimization strategies is shown in
Table 1. Obviously, both optimization cases may induce a
lower hydrogen yield than the original design with the
emissions by 38.4%~45.9%. Moreover, if the waste heat
recovery design is employed, then the hydrogen yield for the
system using Cases I and II usually decrease by 2.5%.
However, it is truly compensated by saving energy and
reducing carbon dioxide emissions. he mass flow of CO2 falls
from 87.46 kg/h to 58.49 kg/h, and the optimization of Case
II can achieve up to a 41.23% reduction in carbon dioxide
emissions.

[6]

[7]

[8]

[9]

665

References
Xu J, Froment GF. Methane steam reforming,
methanation and water-gas shift: I. intrinsic
kinetics. AIChE J. 1989;35: pp.88-96.
Liu K, Song C, Subramani V. Hydrogen and
Syngas Production and Purification Technologies,
A John Wiley & Sons, Inc., 2010.
Posada A, Manousiouthakis V. Heat and power
integration of methane reforming based hydrogen
production. Industrial & Engineering Chemistry
Research 2005;44:9113-19.
Zamaniyan A, Behroozsarand A, Ebrahimi H.
Modeling and simulation of large scale hydrogen
production. Journal of Natural Gas Science and
Engineering 2010;2: 293-301.
Aresta M. Carbon Dioxide Recovery and
Utilization,
Kluwer
Academic
Publishers,
Netherlands, 2003.
Tarun CB, Croiset E, Douglas PL, Gupta M,
Chowdhury MHM. Techno-economic study of CO2
capture from natural gas based hydrogen plants.
Journal of Greenhouse Gas Control 2007;1:55-61.
Nord LO, Anantharaman R, Bolland O. Design and
off-design analyses of a pre-combustion CO2
capture process in a natural gas combined cycle
power plant. International Journal of Greenhouse
Gas Control 2009;3:385-92.
Fan MS, Abdullah AZ, Bhatia S. Utilization of
greenhouse gases through carbon dioxide reforming
of methane over NieCo/MgOeZrO2: preparation,
characterization and activity studies. Appl Catal B
2010;100:365-77.
Fan MS, Abdullah AZ, Bhatia S. Hydrogen
production from carbon dioxide reforming of
methane over NieCo/MgOeZrO2 catalyst: Process
optimization. International Journal of Hydrogen
Energy 2011;36: 4875-86.

8th IFAC Symposium on Advanced Control of Chemical Processes

Furama Riverfront, Singapore, July 10-13, 2012

[10]

[11]

[12]

[13]

[14]

[15]

[16]

Hoang DL, Chan SH. Modeling of a catalytic

autothermal methane reformer for cell applications.
Applied Catalysis A: General 2004; 268:207-16.
Zahedi nezhad M, Rowshanzamir S, Eikani MH.
Autothermal reforming of methane to synthesis gas:
Modeling and simulation. International Journal of
Hydrogen Energy 2009;34: 1292-1300.
Halabi MH, de Croon MHJM, van der Schaaf J,
Cobden PD, Schouten JC. Modeling and analysis of
autothermal reforming of methane to hydrogen in a
fixed bed reformer. Chemical Engineering Journal
2008; 137: 56878.
Ersoz A, Olgun H, Ozdogan S. Simulation study of
a proton exchange membrane (PEM) fuel cell
system with autothermal reforming. Energy
2006;36:1490-1500.
Behroozsarand A, Ebrahimi H, Zamaniyan A.
Multiobjective
optimization
of
industrial
autothermal reformer for syngas production using
nonsorting genetic algorithm II. Ind. Eng. Chem.
Res., 2009;48:7529-39.
Akbari MH, Sharafian Ardakani AH, Andisheh
Tadbir M. A microreactor modeling, analysis and
optimization for methane autothermal reforming in
fuel cell applications. Chemical Engineering
Journal 2011; 166:1116-25.
Silva GF, Fereira ALO, Cartaxo SJM, Fernandes
FAN. Simulation and optimization of H2
production by autothermal reforming of glycerol.
Computer
Aided
Chemical
Engineering
2009:27:987-92.

666