Optimizing the steam reformers at AkzoNobel

Master of Science Thesis within the

Innovative and Sustainable Chemical Engineering Programme

YLVA LINDQVIST
Department of Chemical and Biological Engineering
Division of Chemical Engineering
CHALMERS UNIVERSITY OF TECHNOLOGY
Gteborg, Sweden, 2013

Master of Science thesis

Optimizing the Steam reforms at AkzoNobel

YLVA LINDQVIST
SUPERVISOR
Ulf Andersson
AkzoNobel Pulp and Performance Chemicals
EXAMINER
Professor Derek Creaser
Chemical Reaction Engineering
Chalmers University of Technology

Department of Chemical and Biological Engineering

CHALMERS UNIVERSITY OF TECHNOLOGY
Gteborg, Sweden 2013

Optimizing the steam reformers at AkzoNobel

Master of Science thesis within the Innovative and Sustainable Chemical
Engineering Programme
YLVA LINDQVIST

YLVA LINDQVIST, 2013

Department of Chemical and Biological Engineering

Chalmers University of Technology
SE-412 96 Gteborg
Sweden
Telephone + 46 (0)31-772 1000

Gteborg, Sweden 2013

Optimizing the steam reformers at AkzoNobel

YLVA LINDQVIST
Department of Chemical and Biological Engineering
Chalmers University of Technology

ABSTRACT
This thesis aims at optimizing the energy efficiency at two of AkzoNobel Pulp and
Performance Chemicals steam reformers. The steam reformers produce hydrogen and
steam from a hydrocarbon feedstock. Since hydrogen, and not steam, is the primary
product for AkzoNobel the optimized process must maintain the same level of hydrogen
production but with less hydrocarbon feedstock consumption.
The thesis is divided into three sections; the first examines how the plants has been
operated historically, the second reports the results of experiments performed at the
plants to test possible strategies to improve energy efficiency and the third investigates
the possibility to preheat the air to the burners with the flue gases.
The most important operating parameters in a steam reformer are the steam-to-carbon
ratio and the temperature and pressure in the reformer. Therefore experiments have been
performed with different temperature and steam-to-carbon ratio. It is possible to
decrease the hydrocarbon feedstock consumption with up to 3% by changing the
operating parameters and still maintain the same hydrogen production. However under
these conditions the steam production decreases.
The air to the burners could be preheated with a heat exchanger with heat recovered
from the flue gases. This could save 2 % of the hydrocarbon feedstock with a payback
time of 1.3 years for the heat exchanger.

TABLE OF CONTENTS
1.

INTRODUCTION ................................................................................................................. 1
1.1 Aim...................................................................................................................................... 1
1.2 Structure of the thesis work ................................................................................................. 1
1.3 Delimitations ....................................................................................................................... 1
1.4 Short description of the reforming process ......................................................................... 1
1.4 Background ......................................................................................................................... 2
1.4.1 Hydrogen peroxide ....................................................................................................... 2
1.4.2 Important years for the steam reformers ...................................................................... 3

2.

THE STEAM REFORMING PROCESS .............................................................................. 4

2.1 The reforming reactions ...................................................................................................... 4
2.2 The design of the steam reformer plants at Bohus and Alby .............................................. 5
2.3 Four challenges for nickel steam-reforming catalysts ....................................................... 10
2.4 Creep damage to the tubes ................................................................................................ 11
2.5 The variation of natural gas ............................................................................................... 11

THE DIFFERENCES BETWEEN THE STEAM REFORMERS ...................................... 13

3.1 How to measure the consumption ratio ............................................................................. 13
3.1.1 Bohus.......................................................................................................................... 13
3.1.2 Alby ............................................................................................................................ 17
3.2 How the supplied energy is consumed .............................................................................. 21
3.2.1 Bohus.......................................................................................................................... 21
3.2.2 Alby ............................................................................................................................ 23
3.3 The relationship between consumption ratio and production level ................................... 25
3.3.1 Bohus.......................................................................................................................... 25
3.3.2 Alby ............................................................................................................................ 28
3.4 The relationship between steam to carbon ratio and consumption ratio ........................... 31
3.4.1 Bohus.......................................................................................................................... 32
3.4.2 Alby ............................................................................................................................ 32
3.5 The relationship between air residual in flue gases and consumption ratio ...................... 33
3.5.1 Bohus.......................................................................................................................... 34
3.5.2 Alby ............................................................................................................................ 35

EXPERIMENTAL TESTS ................................................................................................. 37

4.1 Furnace temperature in Bohus........................................................................................... 37
4.1.1 Predicted results ......................................................................................................... 37
4.1.2 Results ........................................................................................................................ 37
4.2 Steam-to-carbon ratio in Bohus......................................................................................... 39

II

4.2.1 Carbon formation ....................................................................................................... 39

4.2.2 Predicted results ......................................................................................................... 39
4.2.3 Results ........................................................................................................................ 40
4.3 Steam to carbon ratio in Alby ........................................................................................... 42
4.3.1 Carbon formation ....................................................................................................... 43
4.3.2 Predicted results ......................................................................................................... 43
4.3.3 Results ........................................................................................................................ 43
5.

PREHEATING OF THE AIR TO THE BURNERS........................................................... 44

6.

CONCLUSIONS ................................................................................................................. 45

7.

FUTURE WORK ................................................................................................................ 46

8.

REFERENCES .................................................................................................................... 47

APPENDIX I: THE DATES OF THE EXPERIMENTS ............................................................ 48

APPENDIX II: NOTATIONS .................................................................................................... 49
APPENDIX III: ALBY ............................................................................................................... 50
APPENDIX IV: BOHUS ............................................................................................................ 51

III

1. INTRODUCTION
AkzoNobel Pulp and Performance Chemicals produces hydrogen as a feedstock for the
production of hydrogen peroxide. The most economical way of producing hydrogen is
by steam reforming (Rostrup-Nielsen, 2002). At three of AkzoNobels plants the
hydrogen is produced by steam reforming. This thesis will focus on optimizing two of
the steam reformers, one in Bohus and one in Alby, both located in Sweden. The Bohus
and Alby steam reformers have capacities of producing 6000 Nm3 hydrogen/h and 3000
Nm3 hydrogen/h respectively. The hydrocarbon feedstock is different for the plants;
Bohus uses natural gas (NG) and Alby uses liquefied petroleum gas (LPG). The
reformers also produce steam which is used for heating other processes at the site and
for district heating. During most of the year all of the produced steam in Bohus is not
utilized since heating demands are too low. Consequently, the Bohus reformer would
ideally lower its steam production but keep the same level of hydrogen production with
lower natural gas consumption.
The hydrocarbon feedstock cost for hydrogen production through steam reforming is
normally about two thirds of the total production cost (Rostrup-Nielsen, 2002). With
todays increasing energy prices it is therefore of importance to have a very efficient
process and there are large economic incentives for optimization.

1.1 Aim
The aim of this study is to investigate the possibilities for AkzoNobel Pulp and
Performance Chemicals of optimizing the two steam reformers in Alby and Bohus to
decrease the feedstock consumptions and keep the same level of hydrogen production.

1.2 Structure of the thesis work

This report will focus on three different ways to optimize the steam reformers;
1) Differences between the steam reformers; Investigate how the plants have been
operated historically and try to find the best operation settings. This part also
includes comparing the two plants to find the most optimal operating conditions.
2) Experimental tests; Find the most important parameters of the plants and
perform experiments in the plants by varying them.
3) Preheating of air to the burners; Investigate the possibility of preheating the air
to the burners at the Bohus plant with the flue gases.

1.3 Delimitations
This report will focus mainly on the Bohus steam reformer since it is at the Bohus plant
the study will be performed. The Alby steam reformer will also be considered but not to
the same extent. The report will not focus on any other uses of excess steam, as for
example electricity generation, since this already has been investigated by AkzoNobel
Pulp and Performance Chemicals.

1.4 Short description of the reforming process

This section contains a short description of the reforming process (Fel! Hittar inte
referensklla.). The feedstock (NG in Bohus and LPG in Alby) is pretreated before
entering the steam reformer which means that the LPG is evaporated and sulfur is
removed from both the natural gas and the LPG. The feedstock is mixed with steam and

is fed to the reformer where it is heated in tubes and the following reaction mainly
occurs:
(1)
The reaction is endothermic and therefore needs heat to be sustained. This is provided
by a burner with NG (in Bohus) or LPG (in Alby) as fuel.
The heat from the reformer production stream and from the flue gases from the burners
that leave the reformer are used to produce steam. The production stream is then fed to a
high temperature shift converter (HTS) and carbon monoxide and steam is converted to
hydrogen and carbon dioxide:
(2)
After the HTS the product stream is fed to a pressure swing absorption unit (PSA)
where it is purified to 99.99% hydrogen.
Flue gases

H2
Water

Steam

PSA

Reformer

Hydrocarbon
feedstock

HTS
Steam
Pre-treatment

Figure 1: An overview of the steam reforming process at Alby and Bohus.

1.4 Background
AkzoNobel Pulp and Performance Chemicals (former Eka Chemicals) is a leading
global force in pulp bleaching technologies, paper chemicals, colloidal silica and
expandable microspheres. One of the products is hydrogen peroxide.

1.4.1 Hydrogen peroxide

The hydrogen produced in Bohus and Alby is used at the same site as a feedstock for
hydrogen peroxide production. Hydrogen peroxide has strong oxidizing properties, low
toxicity and is easy to handle and is therefore widely used as a bleaching chemical. In
ambient conditions it spontaneously decomposes to water and oxygen and the growing
environmental awareness the last decades have led to an increased use since other

bleaching agents tend to be more environmentally unfriendly. The most important

application of hydrogen peroxide is for bleaching of chemical and mechanical pulp. It is
also used for bleaching recycled paper, textile bleaching, as a raw material in the
chemical industry, in the metal industry, for disinfection, laundry bleaching, hair
bleaching and chemicals analysis (Eka).
AkzoNobel Pulp and Performance Chemicals manufactures hydrogen peroxide in five
plants located in Sweden, Norway, the USA and Venezuela. There are two plants in
Sweden, one in Alby and one in Bohus. The hydrogen peroxide is produced by oxidized
hydrogen (Eka):
(3)
Hydrogen is produced in different ways in AkzoNobels different plants. In the USA
and Alby the hydrogen is a byproduct from other processes. In Alby this does not fully
cover the need of hydrogen which is also produced from a steam reformer. The Bohus
and Venezuela plants produce hydrogen only from steam reforming. In Norway the
hydrogen is produced by water electrolysis.

1.4.2 Important years for the steam reformers

To get a better overview of the steam reformers the most important changes are
presented below:
-1990 The first Bohus steam reformer with LPG as feedstock was built in Bohus.
-1997 The steam reformer in Bohus was replaced with a reformer with larger capacity.
-1997 The Alby steam reformer was built.
- 2004 Bohus started to use natural gas as feedstock instead of LPG.
-2005 A second reformer furnace was built in Bohus in order to double the production
capacity.

2. THE STEAM REFORMING PROCESS

This section contains the theoretical information needed to understand the steam
reforming process. It will explain the design of the reformer, the reactions, and the most
common hazards for the equipment.

2.1 The reforming reactions

The basic reforming reaction for a hydrocarbon that occurs in the reformer tubes can be
written as in equation 1. For natural gas methane reacts with steam in the presence of a
supported nickel catalyst and by a reaction called steam methane reforming reaction
(SMR) (Liu, Song, & Subramani, 2010) and is written below.
(4)
However, in reality the SMR is not just one reaction but involves contributions from
several different catalytic reactions called the water-gas shift reaction (WGS), reverse
water-gas shift reaction (RWGS), the Boudouard reaction and methane decomposition
reaction which are presented below (Liu, Song, & Subramani, 2010).
(5)
(6)

(7)

(8)
The reforming reaction (4) normally reaches equilibrium over an active catalyst at high
temperatures. Since the reaction is endothermic it is favored by high temperatures. The
water shift reaction (5) is however exothermic and favored by lower temperatures. The
reforming reaction is favored by low pressures because volume expansion occurs while
the water shift reaction is independent of pressure (Liu J. A., 2006).
In general steam reforming of hydrocarbons is catalyzed by group VIII metals, with
Nickel as the most cost effective catalyst. Steam reforming of methane is strongly
endothermic and reversible. The equilibrium of reaction (4) depends on temperature,
pressure and steam to carbon ratio. In order to have a high conversion of methane it is
necessary to operate at high temperatures, low pressures and relatively high steam to
carbon ratios (Joensen & Rostrup-Nielsen, March 2002). At too high temperatures
however, (over 920 C) the hydrocarbons may react in a parallel reaction by thermal
cracking causing the formation of coke. The steam to carbon ratio is typically 3:1 in
steam reformers (Liu J. A., 2006). To avoid carbon formation at the catalyst (coking),
the steam-to-carbon ratio must be over 1:1 or greater (Cooper, 1998). The coking could
deactivate the catalyst and it is therefore important to maintain a ratio that is sufficiently
high. The relationship between conversion of methane, temperature, pressure and
steam-to-carbon ratio is presented in Figure 2.

Figure 2: The equilibrium conversion of methane in steam reforming against temperature, pressure and
steam-to-carbon (S/C) ratio

In tubular steam reformers, heat is supplied to the reaction by putting the tubes in a fired
furnace. In a typical steam reformer only 50% of the supplied heat is transferred to the
reforming process (Joensen & Rostrup-Nielsen, March 2002). The overall product gas
from the reformer tubes is a mixture between hydrogen, carbon monoxide, carbon
dioxide and unconverted methane and steam. The composition depends on the
temperature of the reactor, the operating pressure, the composition of the feed gas and
the proportion of steam fed to the reactor (Liu J. A., 2006). The amount of carbon
monoxide from the reformer is quite high since equation 4 is favored by high
temperatures. Due to this, a downstream process is necessary to reduce the carbon
monoxide concentration in the product stream and increase the yield of hydrogen. In
Bohus and Alby the process gas from the reformer tubes is fed to a high temperature
shift converter to reduce the carbon monoxide concentration.

2.2 The design of the steam reformer plants at Bohus and Alby
The steam reformers at Alby and Bohus are built with the same design but with minor
differences. The largest difference is that the feedstock to the Bohus reformer is natural
gas and the feedstock to the Alby reformer is liquefied petroleum gas (LPG). The
capacity of the Bohus reformer (6000 Nm3/h) is also twice as large as the Alby reformer
(3000 Nm3/h). To manage this higher capacity the Bohus plant has two reformer
furnaces instead of one. The following section contains a longer description of the
components in the steam reformers. Figure 3 shows the Bohus reformer and Figure 4
shows the Alby reformers.

Atm
3033
ABC

3013

3032

3075
3077

3074
3006
3015
3082

3073
AB

Steam

3014

3030
AB

3072

3076

H2
3004

3071
AB

3005
3003A

3034
AB

NG

3003B

3031
A

3070

3001

3002

3031
B

Figure 3: A detailed overview of the Bohus steam reformer.

Atm
3033
ABC

3013

3032

3075
3077

3074
3006
3015

3073
AB

3082
3030
AB

Steam

3014

3072

3076

H2
3004

3071
AB

3005
3003A

3034
AB

LPG

3079

3070

3001

3002

3031
A

Figure 4: A detailed overview of the Alby steam reformer.

LPG Vaporizer (3079)

The LPG is vaporized by using steam. Since the natural gas is already vaporized, this
heat exchanger is only available at the Alby plant. In Alby most of the LPG is fed to the
heat exchanger 3070 but some is fed to the burners as fuel.
Hydrogen recycle
Pure hydrogen from the finished product is added to the LPG/NG stream to be used
downstream in the process.
Feed gas heat exchanger (3070)
The natural gas /LPG to the reformer tubes are heated to approximately 400 C with the
product gas from the reformer tubes.
Co/Mo Unit (3001)
The NG/LPG contains olefinic hydrocarbons which should be converted to a saturated
hydrocarbon before entering the reformer tubes. The LPG/NG stream is fed to a vessel
containing a cobalt-molybdate catalyst. In the presence of the catalyst the free hydrogen
reacts with the olefinic hydrocarbons which are converted to a saturated hydrocarbon.
Two examples:
(9)
(10)
The sulfur compounds in the natural gas and LPG, reacts in the presence of the Co/Mocatalyst, with the free hydrogen to hydrogen sulfide. For example:
(11)
The reactions are all exothermic which results in increased temperature of the LPG/NG
stream after the Co/Mo unit.
Zinc Oxide Desulfurizers (3002)
After the Co/Mo unit the stream of LPG/NG is fed to a zinc oxide desulfurizer. The zinc
oxide catalyst will absorb the hydrogen sulfide. The operating variables which affect the
performance of the zinc oxide desulfurization catalyst are the flow rate of feed gas, type
and total amount of sulfur in the feed gas, pressure and temperature (Hydro-ChemProcessing, 1991). The sulfur is removed since it is a severe poison for the reformer
catalyst downstream in the process.
Mixture of steam and LPG/NG
The LPG/NG is mixed with steam before it is fed to the reformer tubes. The ratio
between the mixed steam and LPG/NG is called steam-to-carbon ratio.
The reformer (3003)
The reformer is a vessel with tubes heated by burners. It is in these tubes the reforming
reactions occur. The gases in the tubes are never mixed with the gases in the furnace
from the burners. Therefore two streams are leaving the reformer, the production stream
from the tubes and the flue gases from the burners.

The steam reforming reactions 4 occur in catalytic tubes heated by burners. The tubes
are placed in a circle inside the reformer furnace. The tubes are fed with NG/LPG and
overheated steam. The tubes can be seen in Figure 5. Reaction 4 almost goes to
equilibrium and the process gas effluent will be a mixture of CH4, CO2, CO, H2 and
H2O. The major variables in the operation of the reformer are as previously mentioned;
the reformer outlet temperature, the reformer pressure and the steam-to-carbon ratio
(Hydro-Chem-Processing, 1991).

Figure 5: Picture taken inside the Bohus reformer furnace. The tubes are arranged in a circle inside the walls.
The outlet for the flue gases is at the top of the reformer.

Since the steam reformer reaction 4 is endothermic the reaction needs heat to be
sustained. The heat is provided by three burners in each reformer furnace. Most of the
fuel requirement to the burners is supplied by the purge or vent gas from the PSA
Hydrogen Purification unit. The rest of the fuel requirements are supplied from pure
NG/LPG. Air is also fed to the burner to maintain the combustion. Besides maintaining
the reformer reaction, the burners also heat the gases in the tubes and furnace. The
burner should be operated to provide a long and slender flame and the flame should not
impinge on the tubes. Ten percent excess air is normally required for satisfactory
combustion of the fuel. If possible, however, it could be desirable to operate the furnace
excess air percentage below those specified. (Hydro-Chem-Processing, 1991) It is in
other words, desirable to operate the burner with as low air feed as possible since the air
takes up heat (energy) which just produces more flue gases.
Reformer effluent process boiler (3071)
The process boiler utilizes the heat available in the effluent gases from the reformer
tubes to increase steam production. The outlet temperature stream of the effluent gas
can be controlled by this boiler.

3004 High temperature shift converter (HTS)

In the high temperature shift converter most of the carbon monoxide in the effluent gas
from the reformer is converted to carbon dioxide and water in the water gas shift
reaction (WGS):

This reaction is facilitated by a chromium promoted iron oxide catalyst. The conversion
is dependent on temperature, pressure, steam/carbon ration and flow rate. Temperature
is an important parameter and can be used to optimize the performance of the high
temperature shift converter. The higher the temperature, the faster the reaction rate, but
the lower equilibrium CO content (Hydro-Chem-Processing, 1991). In order to achieve
low CO without having significant excess steam it is desirable that the reaction is run at
low temperatures. However, at low temperatures, reaction rates diminish and the
process becomes kinetically controlled (Liu, Song, & Subramani, 2010). Hence, there is
an optimal temperature for a given temperature activity. When the catalyst gets older
this optimum temperature is often increased due to lower activity of the catalyst. The
shift converter is designed to operate with an inlet temperature of 343 C to 379 C. The
maximum allowable temperature is 426 C on the outlet of the high shift temperature
converter. Approximately 80% of the carbon monoxide is converted to carbon dioxide
in the shift converter. The catalyst is sensitive to sulfur poisoning. (Hydro-ChemProcessing, 1991)
Process boiler exchanger (3072)
The heat from the process gas from the HTS converter is used to produce steam in the
process boiler exchanger.
Waste heat boiler feed water exchanger (3073)
The heat from the process stream from the process boiler exchanger is used to preheat
the water from the deaerator.
Deaerator feed water exchanger (3074)
The heat from the process gas stream from the heat boiler feed water exchanger is used
to heat the water to the deaerator.
Process gas cooler (3075)
The process gas is further cooled so that the temperature is lowered and condensate
dumped.
Condensate drums (3082)
The condensate in the process stream is separated by the condensate drums and fed to
the deaerator. The process stream is fed to the PSA unit.
Pressure swing absorption units (3005)
The PSA system purifies the process stream from the high temperature shift converter to
99.99% pure hydrogen. The vent gas is fed to the burners in the reformer as fuel. From
the purified hydrogen stream some hydrogen is taken and fed to the LPG/NG stream
before the desulfurization unit.

Steam pre-heat coil (3076)

The heat in the reformer furnace stack gases is used to preheat the steam which should
be mixed with the LPG/NG.
Waste heat boiler (3006)
The waste heat boiler recovers heat from the exhaust gases of combustion to produce
steam.
Boiler feed water economizer (3077)
The economizer heats the feed water to the steam drum coming from the deaerator.
Deaerator (3013)
The deaerator removes carbon dioxide in the condensate water from the condensate
drum.
Steam drum (3014)
A separate steam drum is used to provide adequate space for steam separation and aid
natural circulation (Hydro-Chem-Processing, 1991).
Water buffer tank (3015)
The fresh water to the reformer is taken from the water buffer tank.

2.3 Four challenges for nickel steam-reforming catalysts

One of the largest producers of catalysts for steam reforming, Haldor Topse, claims
that there are four challenges for nickel steam-reforming catalysts. They include
activity, sulphur poisoning, carbon formation and sintering and the relationship between
them are described in Figure 6 (Sehested, 2006).

Figure 6: The relationship between carbon formation, activity, sintering and sulphur poisoning.

Activity-the catalyst must have sufficient activity to equilibrate the reaction mixture in
the design catalyst volume. In a primary reformer the main reason for low activity is
sulphur poisoning and sintering resulting in higher tube temperatures and even carbon
formation.
Sulphur poisoning- Sulphur is a common and severe poison for nickel catalysts that
blocks the active nickel sites. The sulphur-cointaining compounds in the feed react with
hydrogen to hydrogen sulfide at reforming conditions. The sulphur atom in the
hydrogen sulfide adsorbs strongly at the nickel active sites and blocks it as shown in
equation 12.
(12)

10

Carbon formation-Carbon formation at the tubes can increase the pressure drop, crush
the catalytic pellet, block the active nickel surface and decrease the heat transfer.
Sintering-Nickel particles in the catalyst may grow by fusion of smaller particles
causing a reduction of the catalytic surface area.

2.4 Creep damage to the tubes

On account of the tough operating conditions in a steam reformer the tubes in the
reformer furnace run close to their metallurgical limits and suffer so-called creep
damage over time. Since local overheating occurs in the reformer, this damage does not
occur with equal rate all over the tubes. The typical life time of reformer tubes is 100,
000 hours (11.4 years if the reformer is operated constantly). Even under good
conditions the tubes will undergo creep damage. There is however some things that can
be avoided to prolong the life time of the tubes. Whenever reformer tubes are heated
and cooled stresses are temporarily increased and this leads to an acceleration of creep
damage. Therefore, every shut-down that can be avoided prolongs the lifetime of the
tubes. Another way to prolong the tube lifetime is to control the catalyst charging in the
tubes. The catalyst charging has a significant impact on the reformers life time in two
ways. Firstly, uneven catalyst loading of the tubes results in different pressure drops in
different tubes and that, in turn, results in different gas compositions and a variation of
wall temperature. The different wall temperatures of the tubes results in different life
times. Secondly, packing within a tube must be uniform, otherwise there will be spots
with no reaction, resulting in hot spots and decreased lifetime of the tubes. (Brightling,
2002)

2.5 The variation of natural gas

The raw material in the steam reformer at Bohus is natural gas. Natural gas is an
odorless and colorless naturally occurring mixture of hydrocarbon and non-hydrocarbon
gases found beneath the earths surface, often together with petroleum or coal (Liu,
Song, & Subramani, 2010). Natural gas consists mostly of methane but the composition
of the gas can vary regionally. The natural gas is supplied to Bohus by a pipeline from
Denmark (Swedegas, 2013). The composition of the natural gas is not constant and the
following tables are monthly mean values for the natural gas used in Bohus provided by
the supplier. It can be seen in these tables that both the energy content in the natural gas
and the ratio between hydrogen and carbon can vary slightly. It is unknown how these
variations affect the efficiency in the Bohus steam reformer.

11

Table 1. The composition of natural gas during 2012 in mol%.

CH4

C2H6

C3H8

inC4H10 C4H10
0,37
0,54

iC5H12
0,13

nC5H12
0,08

C6+

CO2

N2

12jan
12feb
12mar
12apr
12maj
12jun
12jul
12aug
12sep
12okt
12nov
12dec

88,87

6,09

2,43

0,06

1,08

0,36

88,94

6,07

2,47

0,38

0,55

0,13

0,09

0,06

0,32

89,06

6,02

2,4

0,38

0,54

0,13

0,09

0,06

0,32

88,57

6,24

2,57

0,38

0,56

0,13

0,09

0,06

1,1

0,32

88,46

6,27

2,59

0,38

0,56

0,13

0,09

0,06

1,14

0,32

88,25

6,36

2,65

0,39

0,57

0,12

0,08

0,06

1,18

0,32

88,97

6,08

2,45

0,38

0,55

0,13

0,08

0,06

0,97

0,32

89,27

2,34

0,37

0,52

0,13

0,08

0,05

0,92

0,31

89,07

6,06

2,39

0,37

0,53

0,13

0,08

0,06

0,32

88,86

6,14

2,48

0,38

0,55

0,13

0,08

0,05

1,02

0,31

89,2

5,93

2,31

0,36

0,51

0,12

0,07

0,05

0,45

89,25

5,91

2,19

0,33

0,47

0,11

0,07

0,05

1,1

0,53

Mean
value

88,9

6,1

2,44

0,37

0,54

0,13

0,08

0,06

1,04

0,35

Table 2 The heating value and density during 2012

12-jan
12-feb
12-mar
12-apr
11-may
12-jun
12-jul
12-aug
12-sep
12-okt
12-nov
12-dec
MV

Higher heating value

MJ/Nm3
kWh/Nm3
43,72
12,15
43,79
12,16
43,74
12,15
43,86
12,18
43,87
12,19
43,94
12,21
43,78
12,16
43,66
12,13
43,73
12,15
43,8
12,17
43,51
12,09
43,29
12,02
43,72

12,17

Lower heating value

MJ/Nm3
kWh/Nm3
39,55
10,99
39,6
11
39,56
10,99
39,68
11,02
39,7
11,03
39,75
11,04
39,61
11
39,5
10,97
39,53
10,98
39,61
11
39,35
10,93
39,14
10,87
39,55

11,01

Density
kg/Nm3
0,83
0,83
0,83
0,83
0,83
0,83
0,83
0,82
0,82
0,83
0,82
0,82
0,83

12

3 THE DIFFERENCES BETWEEN THE STEAM

REFORMERS
The differences between the Bohus and Alby steam reformer were investigated. This
was done by using and comparing data available at AkzoNobel from the plants.

3.1 How to measure the consumption ratio

The consumption ratio between the consumed NG/LPG and the produced hydrogen is a
measure of how efficiently the steam reformer operates. A lower consumption
corresponds to higher efficiency. In the operating manuals the design ratios of steam
reformers with natural gas and LPG can be found. The design consumption ratios have
been compared with the actual consumption ratios of the reformers.

3.1.1 Bohus
The consumption ratio of the Bohus steam reformer is calculated by dividing the natural
gas consumption with the hydrogen production. The natural gas consumption is the
natural gas to the tubes and the burners for both the A and B furnace. The produced
hydrogen is the hydrogen leaving the reformer, the recycled hydrogen and the hydrogen
to the PSA is not included here. This relationship is described in equation 13.
(13)

This can be compared with the design ratio given in the operating manual (HydroChem, 1996) which is 0.49. The reformer also produces steam and the consumption
ratio of the steam production can be measured as:
(14)

The steam is measured is the steam leaving the reformer to be used in other processes.
The last three years (2010, 2011 and 2012) and 2006 are investigated and are shown in
Figure 7Figure 10. 2006 is included since it is the first year with two furnaces and
natural gas as feedstock. The consumption ratio from 1998 (Figure 11) is also included
since the steam reformer was operated with liquefied petroleum gas during that time
period and it is interesting to compare this with the consumption ratio in Alby.

13

0,60

0,52

NG/H2 (Nm /Nm )

0,56

0,48

0,44

0,40

2006
Figure 7: The consumption ratio in Bohus 2006.

0,60

0,52

NG/H2 (Nm /Nm )

0,56

0,48

0,44

0,40

2010
Figure 8: The consumption ratio in Bohus 2010.

14

0,60

0,52

NG/H2 (Nm /Nm )

0,56

0,48

0,44

0,40

2011
Figure 9: The consumption ratio in Bohus 2011.

0,60

0,52

NG/H2 (Nm /Nm )

0,56

0,48

0,44

0,40

2012
Figure 10: The consumption ratio in Bohus 2012.

15

0,30

0,26

LPG/H2 (Nm /Nm )

0,28

0,24

0,22

0,20

0,18

1998

Figure 11: The consumption ratio in Bohus 1998.

It can be seen in the figures that no significant change of the consumption ratio has
occurred the last seven years. The consumption ratio is mostly lower than the design
consumption ratio which indicates that the performance of the reformer is quite good.
No connection can be seen between the consumption ratio and summer and winter, the
performance of the reformer seems to be independent of the seasons.
The steam consumption ratio for 2010, 2011 and 2012 are shown below:
0,30

NG/Steam (Nm /kg)

0,28

0,26

0,24

0,22

0,20

2010

Figure 12: The steam consumption ratio in Bohus 2010.

16

0,30

NG/Steam (Nm3/kg)

0,28

0,26

0,24

0,22

0,20

2011

Figure 13: The steam consumption ratio in Bohus 2011.

0,30

NG/Steam (Nm /kg)

0,28

0,26

0,24

0,22

0,20

2012

Figure 14: The steam consumption ratio in Bohus 2012.

In the graphs it can be seen that the steam consumption ratio varies substantially over
time.

3.1.2 Alby
The consumption ratio of the Alby steam reformer is calculated by dividing the
liquefied petroleum gas consumption with the hydrogen production:
(15)

17

This can be compared with the value given in the design operating manual (HydroChem, 1996) which is 0.21. The steam consumption ratio in Alby is calculated as:
.
(16)

The last three years (2010, 2011 and 2012) were investigated for both hydrogen and
steam consumption ratio and the graphs are shown below:

0,30

0,26

LPG/H2 (Nm /Nm )

0,28

0,24

0,22

0,20

0,18

2010

Figure 15: The consumption ratio in Alby 2010

0,30

0,26

LPG/H2 (Nm /Nm )

0,28

0,24

0,22

0,20

0,18

2011

Figure 16: The consumption ratio in Alby 2011.

18

0,30

0,26

LPG/H2 (Nm /Nm )

0,28

0,24

0,22

0,20

0,18

2012

Figure 17: The consumption ratio in Alby 2012.

0,20

LPG/Steam (Nm /kg)

0,18

0,16

0,14

0,12

0,10

2010

Figure 18: The steam consumption ratio in Alby 2010.

19

0,20

LPG/Steam (Nm /kg)

0,18

0,16

0,14

0,12

0,10

2011

Figure 19: The steam consumption ratio in Alby 2011.

0,20

LPG/Steam (Nm /kg)

0,18

0,16

0,14

0,12

0,10

2012

Figure 20: The steam consumption ratio in Alby 2012.

As in Bohus the consumption ratios for the Alby site varies significantly during
different time periods. It can also be seen that the consumption ratio is higher than the
design consumption ratio. It is also higher than the consumption ratio in Bohus 1998
(when Bohus used LPG as feedstock). Since the consumption ratio is higher than both
the design ratio and the ratio in Bohus the performance of the Alby steam reformer has a
potential to be better. As in Bohus no connection between the seasons of the year and
the consumption ratio can be seen.
The following sections will examine different theories why the consumption ratio
varies.

20

3.2 How the supplied energy is consumed

Since the reforming reaction 4 is endothermic heat must be supplied to the reformer
furnaces by burners. The fuels for these burners are NG/LPG and also vent gas from the
PSA hydrogen purification unit. The heat value of the NG and LPG is quite constant
and known, but the heat value of the flue gases varies and the composition can only be
estimated.

3.2.1 Bohus
To calculate how the energy is used in the steam reformers in Bohus the following
assumptions are made:

The natural gas has the same properties as methane.

The only reaction occurring in the tubes at is
100% conversion of natural gas in the reformer tubes.
The vent gas from the PSA is assumed to consist of 50% H2 and 50% CO2
(Krans, 2011).

First, the total energy available is calculated. The energy available is the energy content
in the fuel fed to the burners. The fuel consists of both natural gas and vent gas from the
PSA-unit. In the control system the vent gas energy content is recalculated as the energy
content of liquefied petroleum gas. This is done based on the ratio between NG and air
fed to be burners,
.The maximum energy to the furnaces can therefore be
calculated as:
(17)

Secondly, the energy demand for the endothermic reaction is estimated as:
(18)

Thirdly, some energy is used to heat the gases in the reformer. The gases in the reformer
tubes, NG and steam, requires the energy for heating:
(19)
(20)
(21)
(22)
The gases in the furnace, NG, air and vent gas also require energy for heating:
(23)

21

(24)
(25)
(26)

(27)

(28)

The energy that is not used for sustaining the reaction is considered to be a loss. Losses
could be due to incomplete combustion of the fuel or heat losses. The total loss is
calculated in equation 29.

(29)

Two examples of how the energy is distributed are shown below (Figure 21and Figure
22). The first one is for maximum production and the second is for minimum
production.

11,08%
30,83%

39,92%
1,8%

NG tubes
Steam
NG burners
Air
Vent gas
Reaction
Losses

4,26%
2,15%
9,95%

Figure 21: How the energy is utilized in the reformer in Bohus with maximum hydrogen production.

22

NG tubes
Steam
NG burners
Air
Vent gas
Reaction
Losses

29,2%
10,99%

14,43%
32,49%
1,54%

2,03%
9,33%

Figure 22: How the energy is utilized in the reformer in Bohus with minimum hydrogen production.

3.2.2 Alby
To calculate how the energy is used in the steam reformers in Alby the following
assumptions are made:

The LGP has the same properties as propane.

The only reaction occurring in the tubes is
100% conversion of the steam reforming reaction in the reformer tubes.
The vent gas from the PSA is assumed to consist of 50% H2 and 50% CO2
(Krans, 2011).

The calculations are similar to those for Bohus. The maximum available energy is:

(30)

Secondly, the energy demand for the endothermic reaction can be estimated as:

(31)
Thirdly, some energy is used to heat the gases in the furnaces. The gases in the tubes,
natural gas and steam, are heated as:

(32)

23

(33)
The gases in the furnace, liquefied petroleum gas, air and vent gas are heated as:
(34)

(35)

(36)

The total loss in the Alby steam reformer is calculated as:

(37)
Two examples of how the energy is distributed are shown below. The first one is for
maximum production and the second is for minimum production.
LPG, tubes
Steam
LPG, burners
Air
Vent gas
Reaction
Losses

28,11%
5,86%
21,45%

34,52%

1,68%
6,4%
1,97%

Figure 23: How the energy is utilized in the reformer in Alby with maximum hydrogen production.

24

29,02%
27,76%

LPG, tubes
Steam
LPG, burners
Air
Vent gas
Reaction
Losses

6,64%

26,17%

1,83%
7,02%
1,55%

Figure 24: How the energy is utilized in the reformer in Alby with maximum hydrogen production.

For both Alby and Bohus a considerable amount of energy (26-40%) is used to heat the
air. The air is needed for the combustion in the burners. If this could be decreased fuel
could be saved. One possibility is to preheat the air with the flue gases (see chapter 5).
Another possibility is to decrease the air flow to the burners. The most economical air
flow is investigated in section 3.5.
The possibilities of preheating the steam more is complicated since the steam already
has an inlet temperature of 400-450 C, so it would require more expensive materials
for the piping. The plants are also built in a way which makes implementing a heat
exchanger to the steam more complicated then to the air. The possibility of minimizing
the steam flow is however examined in section 4.2 for Bohus and 4.3 for Alby.
The flows of NG/LPG to the tubes and burners are considered to be too small for any
preheating to be economical.

3.3 The relationship between consumption ratio and production

level
Whether the efficiency of the steam reformers varies with the production level has been
investigated. The consumption ratio of natural gas to hydrogen should not vary with
different production levels according to the operating manual (Hydro-Chem-Processing,
1991). The steam production consumption ratio is also of interest since the steam is
used in other processes.

3.3.1 Bohus
The production level the last three years (2010-2012) are plotted against the
consumption ratio (
) in Figure 25:

25

0,60
0,58
0,56

NG/H2 (Nm3/Nm3)

0,54
0,52
0,50
0,48
0,46
0,44
0,42
0,40
3000

4000

5000

6000

H2 Production (Nm /h)

Figure 25: The hydrogen production in Bohus 2010 plotted against the consumption ratio.

0,60
0,58
0,56

0,52

NG/H2 (Nm /Nm )

0,54

0,50
0,48
0,46
0,44
0,42
0,40
3000

4000

5000

6000

H2 Production (Nm /h)

Figure 26: The hydrogen production in Bohus 2011 plotted against the consumption ratio.

26

0,60
0,58
0,56
0,54

NG/H2

0,52
0,50
0,48
0,46
0,44
0,42
0,40
3000

4000

5000

6000

H2 Production (Nm /kg)

Figure 27: The hydrogen production in Bohus 2012 plotted against the consumption ratio.

No trend can be seen that the production level has any impact on the consumption ratio
which it should not according to the design. The same is done with the production level
against the consumption ratio of consumed natural gas to produced steam (
)
in Figure 28Figure 30: .

0,30

NG/Steam (Nm /kg)

0,28

0,26

0,24

0,22

0,20
3000

4000

5000

6000

H2 Production (Nm /h)

Figure 28: The hydrogen production in Bohus 2010 plotted against the steam consumption ratio.

27

0,30

NG/Steam (Nm3/kg)

0,28

0,26

0,24

0,22

0,20
3000

4000

5000

6000

H2 Production (Nm /h)

Figure 29: The hydrogen production in Bohus 2011 plotted against the steam consumption ratio.

0,30

NG/Steam (Nm /kg)

0,28

0,26

0,24

0,22

0,20
3000

4000

5000

6000

H2 Production (Nm /kg)

Figure 30: The hydrogen production in Bohus 2012 plotted against the steam consumption ratio.

It can be seen in Figure 30 that the consumption ratio of natural gas to steam is much
lower at higher production levels. In 2010 (Figure 28) and 2011 (Figure 29) this trend
cannot be seen; this might depend on that that in 2010 and 2011 the hydrogen
production was pretty constant.

3.3.2 Alby
The same is done for the last three years in Alby. First the production level is plotted
against the consumption ratio
in Figure 31Figure 33.

28

0,30

LPG/H2 (Nm /Nm )

0,28

0,26

0,24

0,22

0,20
1500

2000

2500

3000

3500

H2 Production (Nm /h)

Figure 31: The hydrogen production in Alby 2010 plotted against the consumption ratio.

0,30

LPG/H2 (Nm /Nm )

0,28

0,26

0,24

0,22

0,20
1500

2000

2500

3000

3500

H2 Production (Nm /h)

Figure 32: The hydrogen production in Alby 2011 plotted against the consumption ratio.

29

0,30

LPG/H2 (Nm /Nm )

0,28

0,26

0,24

0,22

0,20
1500

2000

2500

3000

3500

H2 Production (Nm /h)

Figure 33: The hydrogen production in Alby 2012 plotted against the consumption ratio.

Neither in Alby can a connection between production level and consumption ratio be
found. For the steam consumption ratio of liquefied petroleum gas the same trend can
be seen as in Bohus with lower consumption ratio at higher production levels as can be
seen in Figure 34Figure 36.
0,20

LPG/Steam (Nm /kg)

0,18

0,16

0,14

0,12

0,10
1500

2000

2500

3000

3500

H2 Production (Nm /h)

Figure 34: The hydrogen production in Alby 2010 plotted against the steam consumption ratio.

30

0,20

LPG/Steam (Nm /kg)

0,18

0,16

0,14

0,12

0,10
1500

2000

2500

3000

3500

H2 Production (Nm /h)

Figure 35: The hydrogen production in Alby 2011 plotted against the steam consumption ratio.

0,20

LPG/Steam (Nm /kg)

0,18

0,16

0,14

0,12

0,10
1500

2000

2500

3000

3500

H2 Production (Nm /h)

Figure 36: The hydrogen production in Alby 2012 plotted against the steam consumption ratio.

For the hydrogen production there is no connection between production level and
consumption ratio. The differences in consumption ratio must involve other causes.
However, the steam production is much more efficient at higher production levels.

3.4 The relationship between steam to carbon ratio and

consumption ratio
One of the most important parameters in steam reformers is the steam to carbon ratio
(Hydro-Chem-Processing, 1991) (Liu, Song, & Subramani, 2010) (Liu J. A., 2006). It is
therefore of interest to examine if any connection between the efficiency and steam to
carbon ratio can be found.

31

3.4.1 Bohus
In the Bohus steam reformer the steam to carbon ratio has been constant (3.5 kg steam
/Nm3 natural gas) since natural gas was introduced in 2004. Since the ratio has been this
constant over the years it is hard to decide the ratio which is the most optimal.

3.4.2 Alby
In the Alby steam reformer the ratio varies much more and the steam-to-carbon ratio for
last three years is plotted against the consumption ratio in Figures 37, 38 and 39 below.

0,30

LPG/H2 (Nm /Nm )

0,28

0,26

0,24

0,22

0,20
7

10

11

12

S/C (Nm /kg)

Figure 37: The steam-to-carbon ratio in Alby 2010 plotted against the consumption ratio.

0,30

LPG/H2 (Nm /Nm )

0,28

0,26

0,24

0,22

0,20
7

10

11

12

S/C (Nm /kg)

Figure 38: The steam-to-carbon ratio in Alby 2011 plotted against the consumption ratio.

32

0,30

LPG/H2 (Nm /Nm )

0,28

0,26

0,24

0,22

0,20
7

10

11

12

S/C (Nm /kg)

Figure 39: The steam-to-carbon ratio in Alby 2012 plotted against the consumption ratio.

In the graphs a trend of lower consumption ratio can be seen with lower steam-tocarbon ratio for all three years. If the steam-to-carbon ratio is 9 kg/Nm3 instead of 10.5
kg/Nm3 the efficiency is increased with 2-6%. To confirm this hypothesized
relationship, an experiment with lower steam-to-carbon ratio was performed in Alby,
see section 4.3.

3.5 The relationship between air residual in flue gases and

consumption ratio
The flue gas from the burners in the reformer contains some residual oxygen. In theory
the reformer should perform better if the oxygen concentration is lower. This is due to
the fact that if there is a smaller air flow to the burners, less fuel is needed to heat the
air.

33

3.5.1 Bohus

0,60

NG/H2 (Nm3/Nm3)

0,56

0,52

0,48

0,44

0,40
1,0

1,2

1,4

1,6

1,8

2,0

2,2

O2 (%)

Figure 40: Oxygen percentage in flue gases 2010 in Bohus plotted against the consumption ratio.

0,60

0,52

NG/H2 (Nm /Nm )

0,56

0,48

0,44

0,40
1,0

1,2

1,4

1,6

1,8

2,0

2,2

O2 (%)

Figure 41: Oxygen percentage in flue gases 2011 in Bohus plotted against the consumption ratio.

34

0,60

0,56

NG/H2

0,52

0,48

0,44

0,40
1,0

1,2

1,4

1,6

1,8

2,0

2,2

O2 (%)

Figure 42: Oxygen percentage in flue gases 2010 in Bohus plotted against the consumption ratio.

In Figure 40-Figure 42 a connection between low consumption ratio and low oxygen
content can be observed. However, with the present control system the air is as low as
possible in Bohus already. If the control system can be improved a considerable amount
of fuel (natural gas) could be saved.

3.5.2 Alby
In Alby it is difficult to see a connection between the oxygen content in the flue gases
and the consumption ratio. An experiment with lower oxygen is recommended but has
not been performed in this thesis.
0,30

LPG/H2 (Nm /Nm )

0,28

0,26

0,24

0,22

0,20
2

O2 (%)

Figure 43: Oxygen percentage in flue gases 2010 in Alby plotted against the consumption ratio.

35

0,30

LPG/H2 (Nm /Nm )

0,28

0,26

0,24

0,22

0,20
2

O2 (%)

Figure 44: Oxygen percentage in flue gases 2011 in Alby plotted against the consumption ratio.

0,30

LPG/H2 (Nm /Nm )

0,28

0,26

0,24

0,22

0,20
2

O2 (%)

Figure 45: Oxygen percentage in flue gases 2012 in Alby plotted against the consumption ratio.

36

4 EXPERIMENTAL TESTS
This section contains the results and discussion of the experiments performed in the
steam reformers at Bohus and Alby in the spring of 2013. The dates of the experiments
are presented in Appendix I.

4.1 Furnace temperature in Bohus

The temperature is one of the most important parameters for the steam reforming
reaction 4 (Liu, Song, & Subramani, 2010) (Hydro-Chem-Processing, 1991). The outlet
stream from the reforming tubes has a temperature of 820C. With a lower temperature
the steam reforming reaction 4 will have a lower conversion of methane as explained in
section 2.2. However with a lower temperature less fuel is needed to heat the tubes and
therefore an optimal outlet temperature exists. An experiment done by Eka in 2006
indicated. The equilibrium conversion of methane in steam reforming against
temperature, pressure and steam-to-carbon (S/C) ratio that 810C gives a better
consumption ratio between natural gas and hydrogen production than 820C (Andersson
& Thorn, 2006). This experiment lasted only for six hours and the results could be due
to natural variations in the steam reformer. Therefore a new experiment with a
temperature of 810C was performed during a test period of one week.

4.1.1 Predicted results

This is the predicted consequences of the experiment before it started:
1. Since the temperature is lowered, the conversion of natural gas should be
decreased and more natural gas is needed to the tubes in order to have the same
hydrogen production.
2. The vent gas after the PSA should contain more unconverted methane and will
therefore be more energy rich. Because of this more energy-rich vent gas, less
fuel in the form of natural gas is needed in the burners.
3. Since the temperature in the reformers is decreased, less fuel (natural gas) in the
burners.
4. Since the outlet temperature of the product gas and flue gases is decreased, the
steam production will also decrease.
If the experiment is successful the natural gas to the burners will decrease more than the
required increase in natural gas to the tubes.

4.1.2 Results
During the first days the results of the experiments went as hoped, thee natural gas
consumption to the burners decreased more than the natural gas consumption to the
tubes increased. However after three days the consumption ratio increased again as can
be seen in Figure 46.

37

Figure 46: The consumption ratio and the tubes outlet temperature during the experiment. It should be
Natural gas/Hydrogen instead of Hydrogen/Natural gas for FY5000.

Due to this unknown increase of the consumption ratio the results are divided into three
groups; before the experiment, after the experiment but before the increase of
consumption ratio and after the increase of consumption ratio. The results are shown in
table 3:
Table 3. The most important data before and during the experiment.
19/3-21/3
21/3-23/3
26/3-28
R
0.476
0.468
0.477
H2 production
4528 Nm3/h 4505 Nm3/h 4483 Nm3/h
NG to tubes A
760 Nm3/h 773 Nm3/h 783 Nm3/h
NG to tubes B
760 Nm3/h 773 Nm3/h 783 Nm3/h
NG to burner A
337 Nm3/h 302 Nm3/h 306 Nm3/h
NG to burner B
297 Nm3/h 260 Nm3/h 266 Nm3/h
Steam production
7.62 t/h
7.2 t/h
7.3 t/h
Tubes out temp A
820 C
810 C
810 C
Tubes out temp B
820 C
810 C
810 C
It can be seen in table 3 that the experiment went as predicted during the first three days.
During these days the natural gas consumption decreased with 2.1 % with unchanged
hydrogen production. However, after this the consumption ratio started to increase
again. It is unknown why; it could be due to uncontrolled variations such as a temporary
change of the heating value of the natural gas. The experiment should be performed
again to see if the results are similar. It is important to investigate if the increased
natural gas consumption after three days was a coincidence or a part of the experiment.
The settings used in the experiment could potentially saving 2.1 % natural gas
consumption without reducing the hydrogen production, but the steam production
would decrease by 5.2 %. Before implementing this idea in the plant it should be
considered if the savings in natural gas are worth the loss of steam.

38

4.2 Steam-to-carbon ratio in Bohus

Another important parameter in steam reforming is the steam to-carbon-ratio (section
2.2) to the tubes. In the Bohus steam reformer the ratio has always been 3.5 kg steam to
1 Nm3 natural gas. Therefore it is of interest to test another ratio. A high steam-tocarbon ratio gives a high conversion of the steam reforming reaction (4). Excess steam
is also needed to avoid carbon formation on the catalyst. However the heating of the
steam requires extra fuel to the burners and therefore an optimal steam-to-carbon ratio
exists in terms of energy consumption and stable operation of the catalyst.

4.2.1 Carbon formation

Although the stoichiometry for reaction 4 suggests that only 1 mol of H2O is required
for 1 mol of CH4, the reaction in practice is being performed using high steam-carbonratio, typically in the range 2.5-3 mol/mol in order to reduce the risk of carbon
deposition on the catalyst surface (Liu, Song, & Subramani, 2010). The steam to carbon
ratio used in Bohus plant (3.5 kg/Nm3 equivalent to 3.9 mol/mol) matches the
recommended ratio from the manufacturer of the reformer, Hydro-Chem. However, this
ratio is much higher than the recommended values found in most literature (around 3
mol/mol) and could probably be decreased without causing carbon formation on the
catalyst. The contact person for the supplier of the catalyst, Johnson Matthey was
therefore asked about the minimum steam-to-carbon ratio for the Bohus catalyst and he
writes that the minimum steam-to-carbon ratio will depend on feedstock, reformer
outlet temperature, catalyst age etc. However, as a rule of thumb, for natural gas feeds,
we would recommend a steam to carbon ratio of 2.8 mol/mol as a minimum. It is
possible to run lower than this and there are plants that do this. However, there is a
possibility to over-reduce the HTS catalyst if the steam to carbon ratio is reduced too
far, as you will have too strong a reducing environment in the HTS. This will reduced
the iron HTS catalyst to metallic iron and promote Fischer-Tropsch reactions, where
higher hydrocarbons will be made over the catalyst bed.
Based on this information it is considered safe to do an experiment with a steam to
carbon ratio of 3.0 kg/Nm3.

4.2.2 Predicted results

This is the predicted consequences of the experiment before it started:
1. The conversion of natural gas in the reformer tubes as explained in section 2.1
will decrease due to less steam. To maintain the hydrogen production more
natural gas to the tubes is needed.
2. The fuel (natural gas) to the burners will decrease due to less heating demand.
3. The temperature of the steam to the tubes will have a higher temperature due to
a smaller flow of steam in the steam pre-heat coil (3076). It is important that the
temperature is lower than 450 C due to the properties of materials in the piping.
4. The condensate water flow from condensate drums (3082) will decrease due to
less water in the system.
5. The steam production is believed to decrease due to lower flow of the
production gas from the reformer tubes.
The experiment is successful if the decrease of natural gas to the burners is larger than
the increase of natural gas to the tubes.

39

4.2.3 Results
The experiment with decreasing the steam-to-carbon ratio to 3.0 kg/ Nm3 from 3.5 kg/
Nm3 has been performed three times.
Experiment 1
The first experiment was performed just after the experiment with lower temperature. It
was supposed to be one week but an unplanned shut-down of the reformer ended the
experiment after just three days. Until then the results of the experiment were as
predicted. In Figure 47 the consumption ratio before and during the experiment can be
seen.

Figure 47: The consumption ratio and the steam-to-carbon ratio during the first experiment. It should be
Natural gas/Hydrogen instead of Hydrogen/Natural gas for FY5000.

Since the experiment was performed just after the experiment with lower temperature
the results have to be compared with conditions before any experiment. The results are
presented in table 5.
Table 5. The most important data before and during the experiment.
19/3-21/3
29/3-31/3
R
0.476
0.468
H2 production
4528 Nm3/h 4459 Nm3/h
NG to tubes A
760 Nm3/h 774 Nm3/h
NG to tubes B
760 Nm3/h 773 Nm3/h
NG to burner A
337 Nm3/h 282 Nm3/h
NG to burner B
297 Nm3/h 258 Nm3/h
Steam production
7.62 t/h
7.44 t/h
3
S/C A
3.5 kg/ Nm 3.1 kg/ Nm3
S/C B
3.5 kg/ Nm3 3.1 kg/ Nm3
Steam temp A
391 C
402 C
Steam temp B
391 C
402 C
Condensate water flow 3.00 Nm3/h 2.44 Nm3/h
The natural gas consumption is decreased with 1.7 % according to this experiment. The
steam production is decreased with 0.8 %.

40

Experiment 2
Due to the fact that experiment 1 was too short another experiment was performed. This
experiment also started out with good results but after a few days something happened
with the steam reformer which caused an increasing consumption ratio. This is believed
to be unrelated to the experiment. This can be seen in Figure 48.

Figure 48: The consumption ratio and the steam-to-carbon ratio during the second experiment.

If the assumption is made that the increasing consumption of natural gas during the last
days was due to something else (it became even worse when the experiment was
finished) then the results are presented in table 6. If this assumption is not made, the
results in Table 6 would be different.
Table 6. The most important data before and during the experiment.
4/5-5/5
7/5-310/5
R
0.489
0.476
H2 production
4979Nm3/h 4956 Nm3/h
NG to tubes A
874 Nm3/h 902 Nm3/h
NG to tubes B
874 Nm3/h 902 Nm3/h
NG to burner A
353 Nm3/h 291 Nm3/h
NG to burner B
333 Nm3/h 264 Nm3/h
Steam production
9.15 t/h
8.66 t/h
3
S/C A
3.5 kg/ Nm 3.0 kg/ Nm3
S/C B
3.5 kg/ Nm3 3.0 kg/ Nm3
Steam temp A
387 C
399 C
Steam temp B
387 C
399 C
The natural gas consumption is decreased by 2.6 % with the same hydrogen production
and the steam production is decreased by 2.4 %.
Experiment 3
Since neither experiment 1 nor 2 went as predicted all days, a third experiment was
performed. In this experiment all went well as can be seen in Figure 49 and table 7.

41

Figure 49: The consumption ratio and the steam-to-carbon ratio during the third experiment.

Table 7. The most important data before and during the experiment.
24/5-26/5
28/5-3/6
R
0.502
0.487
3
H2 production
5009 Nm /h 5002 Nm3/h
NG to tubes A
902 Nm3/h 930 Nm3/h
NG to tubes B
902 Nm3/h 930 Nm3/h
NG to burner A
362 Nm3/h 295 Nm3/h
NG to burner B
346 Nm3/h 277 Nm3/h
Steam production
9.3 t/h
9.07 t/h
3
S/C A
3.5 kg/ Nm 3.0 kg/ Nm3
S/C B
3.5 kg/ Nm3 3.0 kg/ Nm3
Steam temp A
388 C
400 C
Steam temp B
388 C
399 C
The results confirmed the findings in experiment 1 and 2. The natural gas consumption
decreased with 3.0% and the steam production decreased with 2.5%. This indicated that
a steam-to-carbon ratio of 3.0 is to be preferred and the steam reformer should be
operated at this ratio in the future.

4.3 Steam to carbon ratio in Alby

Since Alby is using LPG instead of natural gas the results from Bohus cannot be
directly implemented there. However, as in Bohus, an optimal steam-to-carbon ratio
exits.
A steam-to-carbon ratio experiment has already been performed in 2008 by Alby, which
indicated 10.5 kg/Nm3 to be the optimal ratio. However, Figure 37, Figure 38 and
Figure 39 indicates that the last three years a lower steam-to-carbon ratio is to be
preferred.

42

4.3.1 Carbon formation

The operating manual recommends a steam-to-carbon ratio of 8.5 kg/Nm3 and this has
been tested before in both steam reformers in Alby and Bohus, so there is no risk of
carbon formation on the catalyst due to lower ratio at this level.

4.3.2 Predicted results

The predicted results are the same as in section 4.2.2 except that the saved fuel is
liquefied petroleum gas instead of natural gas.

4.3.3 Results
The temperature of the steam to the tubes increased from 439 C to 448 C when
decreasing the steam-to-carbon ratio from 10.42 kg/Nm3 to 9.9 kg/Nm3. Since the
temperature should not exceed 450 C the steam-to-carbon ratio could not be further
decreased. The experiment lasted for a week and the results are presented in table 8.
Table 8. The most important data before and during the experiment.
13/3-25/4
2/5-6/5
R
0.242
0.239
H2 production
1430 Nm3/h 1418 Nm3/h
LPG to tubes
234 Nm3/h
234 Nm3/h
LPG to burner
113 Nm3/h
105 Nm3/h
Steam production
1431 kg/h
1417 kg/h
S/C
10.41 kg/ Nm3 9.9 kg/ Nm3
Steam temp
439 C
448 C
3
Condensate water flow
1.41 Nm /h
1.27 Nm3/h
The liquefied petroleum gas consumption decreased by 1.2% with the same hydrogen
production and the steam consumption decreased by 1.3 %. Since the steam-to-carbon
ratio could not be decreased to less than 9.9 kg/Nm3, larger effect cannot be seen. It
seems however that it is more efficient for the Alby steam reformer to operate at a lower
steam-to-carbon ratio. It is important to consider if the saved LPG is enough to
compensate for the lost steam.

43

5. PREHEATING OF THE AIR TO THE BURNERS

The air to the burners is taken from inside the steam reforming building. It has therefore
the same temperature as the indoor temperature and has a small variation during the
year, depending on the outside temperature. In Bohus the two furnaces have different air
inlet temperature depending on the fact that A furnace has an air-intake higher from the
ground and the air is therefore warmer. Inside the furnace the air is heated to around
1000 C and as can be seen in Figure 21and Figure 22 the heating of air accounts for a
large part of the natural gas consumption. If the air was warmer when entering the
burner it would need less energy for heating. The idea is that heat recovered from the
flue gases could be used to preheat the air.
The savings in natural gas with a certain preheating temperature
as:

can be calculated

(38)

The flue gas leaving the steam reformer to the atmosphere has a temperature of 200 C.
This heat could be used in a heat exchanger to preheat the air to the burners. If a
maximum temperature difference of 10 C is chosen the preheating temperature can
be calculated as:

(39)

Since the flow of the flue gases is not measured in the plant it has to be calculated.
During the time the master thesis was written, an energy saving project was performed
at AkzoNobel Pulp and Performance Chemicals. Both the flow and composition of the
flue gases was calculated in this project. The results indicated that the heat exchanger
could preheat the air to 127 C, saving 4400 MWh or 155 000 euro each year.
A heat exchanger with these properties has an investment cost of around 200 000 euro
with equipment, installation and engineering included. The design details regarding this
heat exchanger are unfortunately confidential.
This gives a payback period of 1.3 years which could be small enough for AkzoNobel to
do an investment. However, since it is not known in detail which equipment and piping
are included in the investment cost (it is not for example certain that new fans are
included in this cost) the payback period could increase.

44

6. CONCLUSIONS
The following conclusions were obtained in the thesis:

It does not matter for the efficiency of the reformer whether it is summer time or
winter time.

The consumption ratio of natural gas to hydrogen is independent of the

production level of hydrogen.

The steam consumption ratio is lower at higher production levels of hydrogen.

The steam-to-carbon ratio in Bohus should be decreased from 3.5 kg/Nm3 to 3.0
kg/Nm3 to improve the efficiency of the steam reformer. With this ratio it is
probably no risk of carbon formation.

The steam-to-carbon ratio in Alby should be decreased from 10.4 kg/Nm3 to 9.9
kg/Nm3 to improve efficiency of the steam reformer. With this new ratio it is no
risk of carbon formation.

A lower reformer temperature in Bohus has the possibility of making the

reformer more efficient. The experiments to test this theory were however
inconclusive and additional experiments are needed.

The air to the burners could be preheated with heat recovered from flue gases
with an estimated payback period of 1.3 years for required process revisions.

45

7. FUTURE WORK
Since the master thesis is limited to a certain time period there was insufficient time to
do as many experiments as wanted. Some suggestions on future experiments are listed
below:
-Test other steam-to-carbon ratios in the reformers to find the optimal ratio.
-Test different pressures in the reformer tubes. A lower pressure could give a higher
conversion of the reformer reaction in the tubes. The pressure could be varied by
varying the pressure after the natural gas compressors. Beware however, that the PSAunit might not be able to handle lower pressures.
-The experiment with lower temperature in the reformer furnace should be repeated
since the result was so unexpected. A similar experiment could also be done in the Alby
steam reformer.
-An experiment with higher temperatures in the furnace could be performed in both
Alby and Bohus.
-An experiment with higher temperature before the HTS could be performed in Alby.
-If the composition of the product gas after the steam reformer and HTS could be
investigated in both Alby and Bohus it would add much knowledge on how to optimize
the reformers.
All experiments would ideally be performed with both high and low production rates.

46

8. REFERENCES
Andersson, K.-E., & Thorn, M. (2006). Optimization test Steam Reformer 7th to9th June 2006.
Bohus: Eka Chemicals.
Brightling, J. (2002). Managing steam reformer tubes. Nitrogen & Methanol (256), 29.
Cooper. (1998). Patent No. 5741440. United States.
Eka. A handbook about hydrogen peroxide and how to handle it. Bohus: Eka Chemicals.
Hydro-Chem. (1996). Operating manual, Hydrogen generating plant. Alby: Eka Nobel.
Hydro-Chem-Processing. (1991). Operating manual hydrogen generating plant. Bohus: Eka
Nobel.
Joensen, F., & Rostrup-Nielsen, J. R. (March 2002). Conversion of hydrocarbons and alcohols
for fuel cells. Journal of Power Sources , 195-201.
JohnsonMattheyCatalysts. (2005). Retrieved from DYCAT 890 Steam Reforming Catalyst:
www.jmcatalysts.com
Liu, J. A. (2006). Kinetics, catalysis and mechanism of methane steam reforming. Worcester
Polytechnic Institute, Chemical Engineering Department.
Liu, K., Song, C., & Subramani, V. (2010). Hydrogen and Syngas Production and Purification
Technologies. Hoboken: John Wiley & Sons.
Mrstedt, S.-E., & Hellsten, G. (2005). Data och digram-Energi och kemitekniska tabeller.
Malm: Daleke Grafiska AB.
Rostrup-Nielsen, T. (2002). Manufacture of hydrogen. Catalysis Today , 106 (1-4), 293-296.
Sehested, J. (2006). Four challenges for nickel steam-reforming catalysts. Catalysis Today ,
103-110.
Swedegas. (2013). Gaskvalitet i Sverige. Retrieved April 3, 2013, from Swedegas:
www.swedegas.se

47

APPENDIX I: THE DATES OF THE EXPERIMENTS

Experiment
Bohus: Temp from 820 C to 810 C
Bohus: S/C ratio from 3.5 kg/Nm3 to 3.1 kg/Nm3
Bohus: S/C ratio from 3.5 kg/Nm3 to 3.0 kg/Nm3
Bohus: S/C ratio from 3.5 kg/Nm3 to 3.0 kg/Nm3
Alby: S/C ratio from 10.4 kg/Nm3 to 9.9 kg/Nm3

Start date
21 Mar 2013
28 Mar 2013
6 May 2013
27 May 2013
2 May 2013

Finish date
28 Mar 2013
31 Mar 2013
13 May 2013
3 Jun 2013
7 May 2013

48

APPENDIX II: NOTATIONS

Liquefied petroleum gas
(Mrstedt & Hellsten,
2005)
(Mrstedt & Hellsten,
2005)
(Mrstedt & Hellsten,
2005)

Density of LPG
Molar mass of LPG
Specific heat of LPG
Lower heat value for LPG

(Swedegas, 2013)

Natural gas
Lower heat value for NG

(Swedegas, 2013)

Water
(Mrstedt & Hellsten,
2005)
(Mrstedt & Hellsten,
2005)

Specific heat of water (liquid)

Specific heat of steam
Air
Specific heat of air
Density of air

1.28

(Mrstedt & Hellsten,

2005)
(Mrstedt & Hellsten,
2005)

Hydrogen
Specific heat of hydrogen
Density of hydrogen

(Mrstedt & Hellsten,

2005)
(Mrstedt & Hellsten,
2005)

Carbon dioxide
Specific heat of carbon dioxide
Density of carbon dioxide

(Mrstedt & Hellsten,

2005)
(Mrstedt & Hellsten,
2005)

Data for the SMR

Enthalpy of formation SMR for LPG at 825C
Enthalpy of formation SMR for LPG at 825C

(Liu, Song, &

Subramani, 2010)
(Liu, Song, &
Subramani, 2010)

Others
Molar volume

49

APPENDIX III: ALBY

Notation

Tag

Unit

FC5410A
TC5120

FC5430A
Estimated

FC5100A
TI5102

FC5200A
TI5202A

TI5301A

TI45410A

FF5430A
FI5210
HC5000
FC5420A
Estimated

Min
production
18/10-19/10
2012
116

Max
production
10/10-12/10
2012
244

100

100

Flow of air to burner

Temperature of air to
burner
Flow of LPG to tubes
Temperature of LPG
to tubes
Mass flow of steam to
tubes
Temperature of steam
to tubes
Temperature of
product gas from tubes
Temperature of flue
gases from furnace

4317

9491

40

40

234

437

327

350

2484

4578

436

404

819

817

951

1047

Ratio air/fuel to burner

20

23

2374

5467

1436

3029

1276

2202

20

20

Description
Flow of LPG to burner
Temperature of LPG
to burner

Mass flow of total

product steam
Flow of product
hydrogen
Flow of vent gas to
burners
Temperature of vent
gas to burners

50

APPENDIX IV: BOHUS

Notation

Tag

Unit

Flow of NG to
burner A
Flow of NG to
burner B

FC5410A
FC5410B
TI0113
TI0213
TI0313

FC5430A
FC5430B
TI5433A

TI5433B

FC5100A
FC5100B
TI5102

FC5200A
FC5200B
TI5202A

TI5202B

TC5300A

TC5300B

TI45410A

TI45410B

FF5430A
FF5430B
HC5000
FC5420A
FC5420B
Estimated

Description

Temperature of NG
to burners
Flow of air to
burner A
Flow of air to
burner B
Temperature of air
to burner A
Temperature of air
to burner B
Flow of NG to
tubes A
Flow of NG to
tubes B
Temperature of NG
to tubes
Flow of steam to
tubes A
Flow of steam to
tubes B
Temperature of
steam to tubes A
Temperature of
steam to tubes B
Temperature of
product gas from
tubes A
Temperature of
product gas from
tubes B
Temperature of flue
gases from furnace
A
Temperature of flue
gases from furnace
B
Ratio air/fuel to
burner A
Ratio air/fuel to
burner B
Flow of product
hydrogen
Flow of vent gas to
A burners
Flow of vent gas to
B burners
Temperature of
vent gas to burners

Min
production
18/8-20/8
2012

Max
production
2/9-4/9
2012

212

362

217

335

82

86

5167

9276

5284

8568

48

48

37

37

550

873

550

873

372

371

1926

3058

1925

3059

400

393

401

395

820

820

820

820

951

1024

961

1038

11

13

11

12

3246

5184

1336

1874

1336

1874

40

40

51