See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/352878967

Design of Ammonia Reactor

Technical Report · July 2021

DOI: 10.13140/RG.2.2.23762.89289

CITATIONS READS

0 2,210

1 author:

Talal Mohammed
University of Birmingham
1 PUBLICATION   0 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Design of Ammonia Reactor View project

All content following this page was uploaded by Talal Mohammed on 01 July 2021.

The user has requested enhancement of the downloaded file.

 Design of Ammonia Reactor

Talal Mohammed Jassim

Third Year Chemical Engineer

University of Birmingham

Abstract: This report detailed the introductory units of an electricity powered ammonia plant. It began by
describing how nitrogen is formed through pressure-swing adsorption. Following that, the report detailed
how hydrogen is produced my first using a mechanical vapour compression unit to provide the electrolysis
fresh desalinated water. The plant decided to maintain de-coupled units type structure where nitrogen and
hydrogen are produced during off-peak hours. As a result, they are compressed and stored in tanks
accordingly. On the other, ammonia synthesis runs continuously and hence operates 24/7. Ammonia is
formed inside a series of three fixed bed reactors where it converts syngas (hydrogen and nitrogen) into
ammonia. Finally, ammonia is separated using two separators: a high pressure and low pressure one. More
specifically, this report detailed the extensive design of the three fixed bed reactors. It began by solving the
mass balance, which yielded the conversion profile along desired bed length. Then, the energy balance was
solved where temperature profile was determined as a function of length. These two balances helped
determine the structure of the reactor where it will be filled with small cylindrical tubes carrying the catalyst.
After that, reactor accessories such as head and skirt thickness were calculated in addition to shell thickness
and maximum compressive strength. Finally, the reactor’s control philosophy was introduced in addition to
the sizing of one pump and one valve, before the whole reactor fixed cost was calculated:
Key Words: Ammonia, fixed bed reactors, syngas, bed length, skirt thickness, pump, catalyst

0
Table of Contents 6 Process Control for Reactor ........................34

1 Introduction ................................................ 2 7 Pump and Valve Sizing ...............................35

1.1 Ammonia .................................................. 2 7.1 Pump Sizing .............................................35

1.1.1 Chemical Structure and Properties ................. 2 7.2 Valve Selection ........................................37
1.1.2 Historical Background & Synthesis .................. 2
1.1.3 Applications ..................................................... 3 8 Costing ......................................................38
1.1.4 Market for Ammonia and Major Players......... 3
8.1 Method 1.................................................38
1.2 HaberVentus Ammonia Production ............ 4
8.2 Method 2.................................................38
1.2.1 Overall Operation ............................................ 4
1.2.2 Specific Nodes/Units Involved ........................ 5 8.3 Method 3.................................................38
1.2.3 Environmental Considerations & Economics .. 8
1.2.4 Safety .............................................................. 9 8.4 Method of choice .....................................38

2 Individual Reactor Design Balances ............ 11 9 References .................................................39

2.1 Ammonia Synthesis Mass Balance............ 11 10 Appendix................................................40

2.2 Ammonia Synthesis Block Diagram .......... 11
2.3 Ammonia Synthesis Energy Balance ......... 12
2.4 Introduction to Reactor Design ................ 13
2.4.1 Reaction Chemistry + Rate ............................ 13
2.4.2 Relationship between Parameters ................ 13
2.5 Introduction to Mass and Energy Balance . 15
2.5.1 Mass Balance (dX/dL) .................................... 15
2.5.2 Energy Balance (dT/dL) ................................. 16
2.5.3 Calculating unknown variables of Mass/Energy
Balance 17
2.6 Solving Mass and Energy balances ........... 20
3 Reactor Sizing ............................................ 22
3.1 Quantifying amount of catalyst for the
reactors............................................................... 22
3.2 Fixed Bed structure.................................. 22
3.3 Design of Catalyst Tubes .......................... 23
3.4 Shell Diameter......................................... 23
3.5 Total Length of Reactor............................ 24
3.6 Volume of Catalyst .................................. 25
3.7 Diameter of the bed ................................ 25
3.8 Total Reactor Volume .............................. 25
4 Reactor accessories calculations ................ 27
4.1 Shell thickness + Wall Thickness of Vessel 27
4.2 Head Thickness Calculation ...................... 28
4.3 Skirt Support Thickness............................ 28
4.3.1 Dead Weight Stress ....................................... 29
4.3.2 Seismic Load Stress ....................................... 29
4.3.3 Final Skirt Thickness Calculation ................... 29

5 Extra Specifications for Reactor ................. 31

5.1 Catalyst Lifetime...................................... 31
5.2 Pressure Drop Across Reactors ................. 32
5.3 Quench System Design ............................ 33
1
1 Introduction 1.1 Ammonia
In the last couple of years humanity has made 1.1.1 Chemical Structure and Properties
significant strides towards life beyond fossil fuels. Ammonia mainly exists as an odourless gas; it is
Both oil and fossil fuels are nearing the end of their the result of covalently bonding one nitrogen atom
lifetimes as it projected that reserves would run out to three hydrogen atoms. For large amounts it is
by the 2060’s. This intensifies the quest of manufactured synthetically, however, it does exist
developing renewable energy as means of a naturally both through plants and animals, in
substitute, mainly to these fuels which have been addition to the human body after consuming
primary energy sources for centuries now. The protein-rich foods (protein is broken down into
transition is proceeding slower than expected and ammonia and amino acids). For storage
complications are arising as a result. However, it considerations, it must be converted into liquid
remains the only route of achieving sustainability ammonia. In addition, when stored it must avoid
for this generation and future ones. The main contact with air as that triggers it to return to its
sources of renewable energy are: Biomass, Solar, gaseous form.
Geothermal, Wind, and Hydropower energy. In the 1.1.2 Historical Background & Synthesis
USA, widely considered the largest individual First isolated and introduced almost 240 years ago,
market in the world, around 88% of the energy ammonia oversaw plenty of minor adjustments and
consumed in 2019 was powered by fossil fuels as low scale methods of production from scientists
show by figure 1 below [24]. On the other hand, such as Priestly, Bartholet, and Nikodem. However,
renewables constituted the remaining 12% with in the early 20th century, the production of ammonia
Biomass energy making up 43% of that share. took a turning point and significantly improved into
a large-scale operation. It was Fritz Haber who
initiated this movement as he produced ammonia in
a low quantity by mixing hydrogen and nitrogen.
Another scientist, Walter Nernst actually proposed
an entire model, more specifically he synthesized
ammonia by mixing H2 and N2 over an iron catalyst
at reaction conditions of 1000°C and 75 bar.
However, given the lack of technology within the
field of materials and construction at that particular
time, it was difficult to accommodate those
conditions in 1900’s pressure vessels. Eventually,
mainly due to the work of Haber, the first ammonia
plant surfaced by 1913, it was situated in Oppau,
Figure 1: USA energy consumption for 2019 Germany as shown below. The German company
BASF were the ones operating this plant with a
Governments are trying to incentivize transition capacity 30 m.t/d of ammonia [25].
to renewable energy both for producers (through
carbon/emission tax) and consumers (diesel tax). In
addition, a significant number of green-oriented
companies are entering the energy market which is
a positive sign of improving and mitigating the
problem of over-reliance on fossil fuels.

As a consequence, this report overviews a project

of producing ammonia in a plant that is powered by
wasted electricity taken from a wind farm. The
primary aim of the project was to both utilize Figure 2: First Ammonia Plant PFD
wasted electrical energy and produce 30 m.t/d of
Ammonia all while being carbon neutral. More The working mechanism of the BASF plant was to
specifically, it considers the detailed design of the compress the H2 and N2 mixture (Syngas) and filter
reactor used to produce ammonia inside the plant. it. Then it enters the reactor where the feed is heat
The reactor will be design focus of this report and heat-exchanged against the products. Also, a gas-
will be assessed and developed extensively. purge is introduced to modify purification and then
it is condensed into liquid ammonia. Once the
2
product is cooled and purified it sent into a further growth is attributed towards humanity’s progression
separator where one stream is final product, and the into the Circular Economy model that conveys
other stream is further distilled and recycled. produce and re-use as opposed to the traditional
method of take-make-waste. In addition, the idea of
In the modern times much of the working principle the model is limit/eradicate process wastes either by
is the same. H2 and N2 are passed over a catalyst of finding a suitable use for them beyond the process
choice through reactor conditions of temperature or regenerate them. As a result, given that ammonia
around 350°C – 550°C and pressure of 150 – 300 can be regarded as a source of green energy (mainly
bar. The main improvement of this process is the because manufacture could yield no carbon
conversion. Through upgraded catalyst which production) it is expected to be in high demand.
extend beyond iron, and also increase in recycle to
the reactor, modern day plant can achieve
conversion up to 20% per pass. In addition, the
reactor technology has also seen significant
upgrades in terms of both axial and radial
technologies in fixed beds. Finally, because of
major advancements in process control and overall
technology, the parameters can be optimized to best
suit a specific plant and its needs. These parameters
include temperature, pressure, bed length, volume
of catalyst…etc.
1.1.3 Applications Figure 4: Ammonia Global Market Revenue
Ammonia has a wide scope of applications in many
industries such as chemicals, textiles, plastics, and The figure above entails the global ammonia market
pesticides. However, more than the majority (82%) revenue, again most of the market will be
of the ammonia produced worldwide is used for dominated by fertilizers (red). Additionally, what
agriculture and more specifically as a fertiliser as can be noticed is that as we near the 2030’s the
shown in figure 3 [26]. revenues are going to increase significantly hence
implying a huge increase in quantity demanded.
MarketResearchFuture® are projecting that the
ammonia global revenues will reach 68.4 billion
dollars by the 2028. For completion purposes, the
current ammonia global revenues for 2020 are
around 30 billion dollars [27]. Ultimately, this
means that the market for ammonia could more than
double in the next 10 years.

With regards to regional analysis, the Asia-Pacific

region has the market share (around 18%) and this
is can be justified by the growing demand for
fertilizer production in Indonesia, China, Thailand,
and India. Finally, the major players globally
Figure 3: Ammonia Application Usage include world renowned companies such as SABIC,
Koch Fertilizer, CF industries Holdings Inc., and
The explanation of the major consumption of Acron. However, the companies mentioned above
ammonia within the fertilizer industry is because it are major players in terms of production quantity.
very essential in preserving global food production As a result, their methods of processing entail
rates. Ammonia is used to create ammonium nitrate heavy amounts of CO2 produced and also other
(fertilizer). The fertilizer’s function is to administer emissions. Given that the world is transitioning into
nitrogen and that then serves as a key ingredient in green energy that does not endanger the
plants growing. Finally, ammonia can also be used atmosphere, it is important to consider green
as a refrigerant due to its excellent heat transfer ammonia producing companies such as Haldor
properties. Topsoe, KBR, ThyssenKrupp, and
1.1.4 Market for Ammonia and Major Players AmmoniaCasale. The aforementioned companies
The market for ammonia is set to grow specialize in carbon-friendly ammonia production.
considerably over the next couple of decades. This
3
1.2 HaberVentus Ammonia Production Nitrogen will be produced by the pressure swing
1.2.1 Overall Operation adsorption method (PSA). On the other hand,
The goal is to produce 30 m.t/d of NH3 and the hydrogen is produced by alkaline electrolysis (AE).
energy requirements will solely be met by electrical However, since the electrolysis unit requires fresh
energy produced by the Whitelee Farm, all while desalinated water, a mechanical vapour
being carbon neutral. In addition to being carbon compression unit (MVC) is installed prior to the
neutral, there will be no waste emissions that could electrolysis unit. The plant operates in a two
endanger both the environment and atmosphere. decoupled sections manner where H2 and N2 are
The plant will be situated 27 kilometres to the both produced in the night-time using off-peak
south-west of the Whitlee Wind Farm. This location electricity, then they are kept in storage tanks and
lies in Stevenston, Scotland. The proposed plant ready for when required. On the other hand, the
location is shown below. The plant will look to run ammonia synthesis loop runs continuously at 24
for 330 days/yr. and ammonia production will run hours a day. This stunt will help reduce plant costs
for 24 hours a day. The total output of the plant is mainly energy that is bought from the farm. The
11,000 m.t/d. In addition, Whitelee Farm will ammonia synthesis loop first takes and compresses
supply only 30% of the wasted energy at no cost, the Syngas mixture before heat-exchanging this
the other 70% will have to be purchased from the feed against first run products. Then, the reactants
farm directly. are taken to the ammonia reactors. The structure of
the reactors is going to be 3 individual fixed bed
reactors operating in series. As a result, in total
there are going to be 3 reactors with three separate
beds. More information will be disclosed later.
Once the ammonia is formed, it is then separated in
a two-stage flash separation process (one is high
pressure and the other is low pressure). Before
entering the separators, it is heat-exchanged against
the incoming feed (as mentioned earlier) and is
cooled down to require flash separator temperature.
After passing the separators, the ammonia product
Figure 5: HaberVentus Ammonia Plant Site is separated, one stream is purged and recycled,
meanwhile the other is stored and hence can be
transported later.
NH3

Air N2 NH3
PSA Synthesis Loop Separators

Desal.
Seawater Seawater H2
MVC Elec. NH3

NH3
Transport
Figure 6: HaberVentus Plant Block Storage
Diagram

4
1.2.2 Specific Nodes/Units Involved
1.2.2.1 Process Flow Diagram

Figure 7: HaberVentus Plant PFD

1.2.2.2 Node 1: Nitrogen Production used to remove impurities such as carbon

The chosen technology for nitrogen production is particulate matter. The now filtered and compressed
PSA as mentioned earlier. Generally-speaking, air gas enter the 4-bed PSA configuration, the air goes
flows into the first column where carbon molecular through the first vessel which is the absorption
sieves are first used to separate nitrogen and vessel then followed by the desorption vessel, this
oxygen. Next, an adsorbent is introduced and this area is titled PSA1-01 in the PFD above. Finally,
binds to oxygen much faster than hydrogen. the oxygen-argon mixture separated from the
Subsequently, the free nitrogen flows out as absorbent in the second vessel flows out as a
product, meanwhile the adsorbed oxygen flows into separate stream (OA1-1). As for the nitrogen gas, it
the desorption unit. In the desorption unit, the is once again cooled (COL1-02) and filtered (FIL1-
vessel is depressurized to remove the oxygen and 02) to reach a purity of 99.9%. It is now ready for
regenerate the adsorbent. Ultimately, the either use or storage.
regenerated adsorbent can flow back to the first 1.2.2.3 Node 2: Desalinated Water Production
vessel (adsorber) and once again it can process new This node details the operation of the MVC unit
incoming air feed. More specific to our design specifically. As mentioned in section 1.2.1,
(referring to PFD above), air at standard electrolysis requires a large amount of desalinated
temperature and pressure of 298 K and 1bar first seawater for its particular operation. As a result, this
enters a compressor (COM1-01) where pressure unit was introduced to handle that requirement. The
increases to 23 bar. Next, the compressed air is operation of MVC entails two inputs being fresh
passed through a cooler (COL1-01) followed by a seawater and a stream of vapour. What happens
filter (FIL1-01) to purify the air from any unwanted next is that through the installation of several
components. The filter is a cyclone separator that is evaporation chambers and tubes, the two inlets
5
come into contact. Moreover, the vapour stream is permeability towards both hydrogen and oxygen,
mechanically compressed, and this significantly they either dissolve in the electrolyte or form gas
increases both temperature and pressure. This now bubbles. As a result, the diaphragm is able to draw
heated vapour begins to warm up the tubes installed other components such as saline potentially and
in each evaporation chamber. Subsequently, the isolate those two product gases (hydrogen and
fresh seawater is then met with these tubes and the oxygen) away from all this. It is worth noting that
heat causes the water to start evaporating. The mixing of gases of avoided, and hence it easy to
outcome is now hot water vapour that is taken to the separate the gases and obtain hydrogen.
subsequent evaporation chamber and is utilized as
the heat source for the tubes. When the next portion At HaberVentus the design of this unit initially
of sea water enters the chamber, the hot water oversaw the product stream from the MVC being
vapour condenses in the process. Finally, the the desalinated water into a centrifugal pump titled
condensed water is taken as ready product for CEN3-01 in the PFD. This pump unit raised both
electrolysis unit and the water that did not convert the temperature and pressure of the water to 348 K
is returned to sea as “high salinity” brine. A typical and 30 bar respectively. What followed was entry to
MVC set-up is shown below in figure 8. [28] 6-A50 electrolyser stacks (AE3-01) where
subsequent reactions and transfer of electrons
occur. The hydrogen produced is taken to a
compressor (COM3-01) where the electrolysis
process begins. A current (DC) is applied to the
stack, subsequently electrodes begin to travel from
DC source to the cathode. The selected choices for
the cathode and anode were nickel on steel and
nickel cobaltite. The reason for the cathode and
anode choice is because the metals aforementioned
possess excellent mechanical strength, superb
resistance to corrosion, and finally they are good
conductors of electricity. The electrolyte of choice
is potassium hydroxide, the function of the
electrolyte is to allow the ions to travel between the
Figure 8: MVC Unit Set-up electrodes. Then, the electrodes are placed far apart,
and they are separated by a porous PTFE
With regards to HaberVentus design of this node diaphragm. The function of the diaphragm is to
more specifically, fresh seawater comes from the separate the hydrogen and oxygen as it exhibits low
river near the plant as stream SE2-01 at a permeability to both of them. The hydrogen and gas
temperature of 283 K and pressure of 1 bar. Next, either slightly dissolve or exist in gas bubble
the seawater enters the MVC unit titled MVC2-01 configurations, either way the diaphragm exhibits
in the PFD where the process operationg described low permeability. Given the 84.8% stack efficiency,
above occurs. Ultimately, desalinated seaweater the product gas produced consists of 89% hydrogen
(WA2-03) leaves at a much higher temperature of and 11% water, this mixture leaves the electrolyser
323 K while maintaing the same condition. The at 390 K.
brine is recovered in the same conditions, however,
it is assumed that it undergoes further processing Following the electrolyser unit, the products then
before returning to the river (not part of proceed to a shell & tube heat-exchanger. The
HaberVentus operation). hydrogen enters the shell-side port at 390 K and 1
1.2.2.4 Node 3: Hydrogen Production bar, the shell is made of stainless stell. As for the
The selected technology for hydrogen production water, it is chilled and desalinated, it enters the
was Alkaline Electrolysis (AK). In addition, the tube-side port at a temperature of 278 K. The
working principle of AK is the desalinated water hydrogen stream flows counter-currently to the
enters into a set-up where an electrolyte, anode, water. As a result, hydrogen is cooled down to 310
cathode, and diaphragm simultaneously exist. K, meanwhile the present water is condensed and
Generally-speaking, the anode and cathode hence purity of the hydrogen increases to 99%. On
submersed into an electrolyte of choice will enable the other hand, the coolant stream is heated to 303
transfer of ions back and forth. The diaphragm is K. The resultant hydrogen out of the heat-exchanger
placed in the centre. The function of this diaphragm is then compressed to 70 bar, then proceeds to enter
is to separate product gases. It exhibits low a horizontal gas-liquid separator. Due to the effect
6
of gravity, the entrained water droplets present in The ammonia reactor synthesis system is composed
the hydrogen will fall out and collected at the of 3 axially mixed fixed bed reactors in series.
bottom of the vessel. Next, the hydrogen gas Syngas (nitrogen + hydrogen mixture) enters the
contained with the water enter an inlet diverted first reactor at the reactant auto-ignition temperature
which separates the two. Finally, in preparation for of 400°C (673 K) and pressure 200 bar. Reactor 1
storage, the water is completely removed from the specification is 5 metres long and achieves a
hydrogen. Subsequently, the hydrogen is sent to a conversion of 11%. This reactor utilizes 2572.8 𝑘𝑔
compressor where it is compressed to 200 bar, then of catalyst dispersed over 200 small tubes each with
it is stored in gas form inside a Type-IV carbon- dimensions of 2 m length and 0.15 m diameter. The
fibre tank. rate of the reaction in the first reactor is
1.2.2.5 Node 4: Nitrogen and Hydrogen Storage 11.6 𝑚𝑚𝑜𝑙 𝑔−1 ℎ𝑟 −1 . In addition, reactor 1
Nitrogen leaving the PSA unit is stored inside of a reached a temperature of 702 K within the length of
storage tank GST4-01 at 6 bar and 96 K. This bed, however this is cooled down to 688 K by the
nitrogen stream NI4-02 then enters a cryogenic introduction of an unconverted feed quench stream.
pump CRY4-01 where nitrogen gas is trapped by In addition, it is worth noting that pressure drop
being condensed on a cold surface. Then for occurs across all reactors and specifically 0.72 atm
hydrogen storage, hydrogen enters the node at 57 for the first reactor. As a result, the stream leaving
bar and 273 K (HY4-03), this stream enters the reactor 1 enters reactor 2 at 688 K and 199.28 atm.
storage tank GST4-02 where it then leaves at 200 In reactor 2, once again conversion remains at 11%,
bar and 333 thanks to an expander. however reactor length increases to 6 m. The
reactor utilizes 3210.1 𝑘𝑔 of catalyst which is more
than reactor 1 as the rate of the reaction drops to
9.3 𝑚𝑚𝑜𝑙 𝑔−1 ℎ𝑟 −1. The catalyst is packed in 2.5
m long tubes with 0.15 m diameter, there are 250
tubes. Finally, the temperature profile of reactor 2 is
it reaches 722 K across the bed length of the
reactor; however, this is cooled down to 710 K with
the help of the quench design. Also, a pressure drop
of 0.64 atm occurs in this reactor. Lastly, in reactor
3 a length of 6.4 m is designed with 2.7 m long
tubes of diameter 0.15 m, there 270 tubes total for
this reactor. The conversion stayed constant at 11%
but the rate decreases once again to
Figure 9: Example of Hydrogen Storage Tank
8.99 𝑚𝑚𝑜𝑙 𝑔−1 ℎ𝑟 −1 and so catalyst utilization
1.2.2.6 Node 5: Ammonia Synthesis increases to 3315.4 𝑘𝑔. Given the length of reactor,
In order to heat the syngas mixture (nitrogen and the temperature profile reached an outlet value of
hydrogen) up to the ignition temperature of the first 748 K, while a pressure drop of 0.55 atm occurred.
reactor, the mixture is passed into a heat-exchanger. The total output of ammonia from this reactor
This heat-exchanger is a plate-fin type specifically system is 38 m.t./d.
made for high operating pressures; in addition, they
are mainly used for high temperature profiles. The
plate-fine HEX also offers a unique compact
design. The task is to provide thermal energy such
that the reactants are raised to a temperature of
400°C. Furthermore, the products are then cooled
down to a temperature of 266°C. For the product
stream at the separation stage, coolant water is used
to further bring down the temperature of the
products to 59°C. The source of water is the MVC
unit as it provides desalinated water. The water
from the MVC is at 50°C, and so after heat-
exchanger operation the products temperature is
decreased down to 30°. The second heat-exchanger
will use seawater as the cold fluid. The sizing of the
compartment for the first HEX is 214 m2, as for the Figure 10: Typical Axially Mixed Ammonia
second one it is 156 m2. Reactor
7
1.2.2.7 Node 6: Ammonia Separation and Storage dioxide emissions. The future is now. Hence, this is
This node was designed for the purification of the what drives HaberVentus to operate with
ammonia produced, the ammonia is separated from sustainability at the back of our minds, in addition
the large amounts of nitrogen and hydrogen. In to limiting carbon footprint.
terms of structural context, this section follows the
HEX5-01 which uses the heat of the products
formed in the last reactor REB5-03 to heat up the
initial feed into the first reactor REB5-01. The
products then enter an expansion valve PRV6-01
where the pressured is lowered from 200 bar to 50
bar. The temperature at this stage is 332 K. After
this step, the products progress to the first
distillation column (DIS6-01). In this first separator,
a large chunk of nitrogen and hydrogen, in addition
to some ammonia and inerts, leave the top of the
column and hence be used for recycling. This
stream is titled HN6-07 and is sent to the first
Figure 11: Temperature profile. vs Pressure
reactor REB5-01. Furthermore, a purge stream
relationship for Ammonia Reactors
(HN6-10) is taken from HN6-07. Moving on, the
bottom of the separator sees the bulk of the
The figure above details how the GDP (blue) and
ammonia leave, this ammonia (AM6-03) is then
carbon footprint (red) compare for the UK over the
passed over an expansion valve PRV6-02 to relieve
last 20 years. There has been steady progress with
some of the pressure. The next unit is the second
regards to carbon production, however the annual
distillation column, the low-pressure column.
value for the UK still remains high compared to
Stream AM6-03 enters the distillation column
other values. Hence, more green powered projects
DIS6-02 where again the bulk of the ammonia
should be considered and developed. [29]
leaves through the bottom as stream AM6-05,
which is then taken to storage unit GST6-01. The
The important design feature of this process is
top products of the low-pressure column
that it is making use of wasted energy. When wind
(nitrogen+hydrogen+inerts) will be recycled back to
farms cannot operate due to low electricity
the first reactor and this is HN6-08. However,
demands, huge sums of electricity go to waste and
before it is sent back to the reactor, it undergoes
hence here is the significance of this project, it is
another purge, where HN6-09 is the purge stream.
efficient. The product, ammonia, is a clean and
Finally, whenever product ammonia is required for
renewable resource that is an alternative to fossil
delivery or transport say for a customer or another
fuels in many processes such as manufacture of
plant, the ammonia gas is taken from the storage
plastics, pharmaceuticals, and other products. The
tank onto the company van and hence it can be
main condition of this process was that ammonia
delivered. The conditions inside the van storage will
produced must be clean meaning renewable and
replicate that of the storage contained.
zero/low CO2 emissions. The raw materials in this
1.2.3 Environmental Considerations & Economics process are hydrogen and nitrogen. The hydrogen is
The modern is rapidly transitioning away from the produced from the electrolysis of desalinated water.
conventional resources such as fossil fuels and This process is completely carbon-free, and no
natural gas. For centuries, these resources have wastes are created, hence it complies with the clean
exhausted the environment through the release of requirement. As for nitrogen, it is produced through
gargantuan carbon dioxide amounts, this has air separation via pressure swing adsorption. This
negatively impacted the earth and atmosphere. As a results in the formation of nitrogen and oxygen.
result, climate is accelerating exponentially with all This unit of the process has no wastes produced and
sorts of negative implications. Industries answer by is also carbon-free once again complying with the
resorting to renewable resources that signify the carbon constraint. Finally, when both are reacted
need for sustainable development (the ability to together into the reactor (Haber-Bosch) at very high
provide for the current generation without temperatures (380-550℃), usually propane-powered
comprising the future generation’s ability to meet furnace burners are used to provide this heat.
their needs). These renewable resources mean However, for this process a train of heat-exchangers
“clean” and that explicitly means that little to no are used to achieve those temperatures. As a result,
waste are associated with its use and no carbon no CO2 is produced. The one issue with this process
8
is that the back-up grid is used just in case demand well-being of the employees. Also, especially if the
for electricity increases suddenly and the process business is successful, this will lure other
needs to be powered alternatively. The companies that require ammonia such as fertilizer
implementation of the back-up grid introduces CO2 producers to possibly move into the area and hence
footprint. However, the amount produced will be eliminate significant transportation costs. Lastly,
minor compared to the large CO2 amounts produced the increase in economic growth results in more
when fossil fuels are used to power the process. room for government projects and expansions
However, relative to other ammonia producing within the town of Stevenson (build more roads,
companies, our carbon figures are almost open new establishments, expand operations).
inconsiderable. Table 1 details carbon dioxide 1.2.4 Safety
produced as a result of ammonia processes, there Lengthy and premeditated safety considerations
seems to be a steady rate. Note however that these must be made as this plant has a considerable
numbers are for the early 2000’s. It would be number of hazards associated with it. Firstly, some
interesting to see what the late 2000’s presented in of the gases in the plant such as hydrogen are
terms of carbon production; however, it is a unified flammable and hence there is always a risk of a fire
goal of all companies to work to reduce that. starting. Also, a pressure surge could occur as there
are many pipes involved in the plant. This pressure
surge mainly happens due to the outlet of the pipe
being closed; hence the fluid is trapped, and
pressure starts building up. The most dangerous
Table 1: Carbon Dioxide production from fluid to be trapped is ammonia gas, so if ammonia
Ammonia Production Processes gas faces a pressure while being transported across
a pipe, if it absorbs too much while standing still it
In terms of assessing the economic effect of our could explode. On the topic of ammonia, exposure
plant, Ammonia is a very highly sought-after to ammonia is very irritating and could be
feedstock (especially for fertilizers), hence the plan damaging for the eyes. In addition, given that this is
is to produce it and sell it to other plants that require a plant producing 30 m.t./d of ammonia, there is
ammonia. This means that less costs are incurred on going to ammonia all over the plant in reactors,
the buying companies as ammonia is cheaper than storage tanks, pipes, and vans for transport. Another
having to employ fossil fuels within their own possible hazard for the human body specifically is
respective plants (no emissions as well). oxygen toxicity, especially near the PSA unit. If
Furthermore, our company can use these profits for oxygen is inhaled in supplemental amounts, then it
R&D and start looking into further technologies could be quite toxic for the lungs. Moving on, there
that could make us stand-out. As for the town of is also the threat of asphyxiation especially if the
Steventson and more generally Scotland, this chemicals are disposed of in an incorrect way, the
introduction of this plant producing ammonia air around could be filled with irritant chemicals
results in the influx of capital into the economy. that could deprive a person from actual oxygen.
There will be a significant number of jobs created With regards to the electrolyser unit, the potassium
meaning better futures for the families of the oxide used as electrolyser is quite toxic and so if
employees and also future generations. The influx inhaled it could pose as a safety hazard. Then there
of money especially through the foreign workers are general hazards such as working with highly
will result in more government revenue. The pressurized gases, technicians working at large
unemployment rate will significantly drop in the heights, and facing a shock when using electrical
area. All these factors mentioned above result in the appliances.
increase in the GDP of Scotland since exports go up
significantly. Overall, due to the aforementioned In order to avoid all these hazards, safety practicing
chain effect mentioned above, the economic growth will be ordered on all workers on-site. The workers
within the area significantly rises. The downsides and operators will be advised on how to safely use
are that foreign workers will present job rivalry. chemicals and maintain equipment. Moreover,
Also, the consumption increases due to the influx of technicians will be taught to handle high pressure
new workers and hence more resources will be gases and how to operate around them. Back to the
utilized that expected. operators, they will be equipped with high level
personal protective equipment such as protective
As for the social development, the quality of life glasses, safety helmets, fire-resistant suits, gloves,
within the town of Steventson will evolve highly protected boots, respirators, and protection
significantly, the booming business improves the
9
for the ears. With regards to equipment, regular The below table presents a couple of very serious
maintenance will occur to make sure the units are incidents that occurred inside ammonia plants.
up to speed and not faulty. In addition, detailed and Therefore, it is apparent that the safety protocol is
thorough inspections will occur regularly to make to be taken considerably, and also be applied
sure the equipment is not damaged. With regards to rigorously. This is because a major accident
major hazards that cannot be stopped, a well set-out occurring inside the plant’s doors would not only
safety procedure will be implemented. The well- and mainly risk lives, but potentially end the
known safety protocol is to be comprehend by each business. The company will then have to pay fines
member of staff. There must be several assembly in damages and face lawsuits from the people of
points in the case fire spreads at an already known Steventson and as well as plant workers. The plant
point. In addition, regular evacuation procedures would also be heavily under the microscope by
will be done to keep the protocol freshly engrained national safety boards. The long-term effects would
in workers. As for around the plant, in order to find be detrimental. Therefore, safety must be applied in
out where ear protection should be offered, noise the best way possible and no room for error is
measurements will be taken regularly. Yellow tapes allowed. One of the good safety features of this
enclosed on perimeters will detail that is area is plant is that there are no fuel burners are the
inaccessible due to possible hazards. electricity produced by the plant is used to convert
unit’s needs. However, it seems that the most
As for process control, each unit and node apparent danger stems from chemical exposure and
described will have some form of process control the reactor. For chemical exposure, it is just a case
implemented. Every pipe in the plant will have of pipes, vessels, and storage containers being
some form of pressure relief valves in case pressure regularly maintained. On the other hand, for the
builds in of them. Quench designs are readily reactor, the process control must be diligent and
available for the highly exothermic units. The controlled, mainly controlling temperature and
material of vessels is chosen wisely enough to avoid pressure. If those two areas are met with a great
breakdown of equipment. Composition analysis will deal of focus, then the overall plant can be
be done to see any unwanted pollutant or toxins. potentially on the safe side. It all comes down to
Pressure and temperature will be assessed so that process design, safety implementation, and finally
they do not exceed desired levels. thoughtful engineering.

Table 2: Ammonia Plant

Incident Review
10
2 Individual Reactor Design Balances
2.1 Ammonia Synthesis Mass Balance

stream (kmol/hr) 1 1.1 1.2 2

N2 135.54 120.63 107.36 95.55

H2 406.62 361.89 322.08 286.65

I 0.12 0.12 0.12 0.12

NH3 13.6 43.42 69.96 93.58

total (kmol/hr) 555.87 526.05 499.51 475.89

Table 3: Ammonia Synthesis Mass

Balance in kmol/hr

stream (kg/hr) 1 1.1 1.2 2

N2 3795.12 3377.64 3006.08 2675.4

H2 813.24 723.78 644.16 573.3

I 4.8 4.8 4.8 4.8

NH3 231.2 738.14 1189.32 1590.86

total (kg/hr) 4844.36 4844.36 4844.36 4844.36

Table 4: Ammonia Synthesis Mass

Balance in kg/hr

2.2 Ammonia Synthesis Block Diagram

1 1.1 1.2 2
Reactor 1 Reactor 2 Reactor 3
(BED 1) (BED 2) (BED 3)

Figure 12: Ammonia Synthesis Block

Flow Diagram

At this stage of the design, all that was readily Finally, Reactor 3 produces 9.6 m.t/d of ammonia.
available was the mass balance for the reactor loop Also, what can be calculated from this information
as shown above. In addition, neither temperature is the conversion. All of the three reactors have the
profile nor pressure of reactor were known. On the same conversion of 11%. In addition, the total
other hand, stream content into and out of each bed single-run conversion is 29.5% with recycle
was known. What can be calculated is the ammonia included. In the next couple of sections, primary
production from each reactor/bed. For Reactor 1, it parameters for the reactor will be configured, and
produces around 12.2 m.t/d of ammonia. As for then from there actual design of the physical
Reactor 2, the production of ammonia is 10.8 m.t/d. properties of the reactor will begin.

11
2.3 Ammonia Synthesis Energy Balance

Reactants at Products at
673 K 748 K

∆𝐻1 ∆𝐻3

∆𝐻2
Reactants at Products at
298 K 298 K

Figure 13: Ammonia Synthesis Energy

Balance Flow Diagram

With regards to the energy balance around the

reactor system, it treats the system as 1 reactor. Up
to this point of the design, the inlet and outlet
temperature of the whole reactor system (all three
reactors) is unknown, however it was calculated in
later sections. For the sake of the chronological
order of this report, the values were used here but
later it is revealed how to calculate them. In order to
calculate enthalpy values, information from mass
balance is used in addition to specific heat vs
Figure 14: Specific Heat as a function of
temperature charts such as figure 14.
temperature for nitrogen.

Enthalpy Value

∆𝐻1 −5999375.4 𝑘𝐽 ℎ𝑟 −1

∆𝐻2 −3729033.6 𝑘𝐽 ℎ𝑟 −1

∆𝐻3 3374023.6 𝑘𝐽 ℎ𝑟 −1

∆𝐻 6354384.5 𝑘𝐽 ℎ𝑟 −1

Table 5: Ammonia Synthesis Energy

Balance Enthalpy Values 12
2.4 Introduction to Reactor Design With regards to pressure, Le Chatelier’s principle
In order to initiate design of the reactors, what indicates that if the pressure is increased then the
firstly needs to be decided is the selection of equilibrium will shift towards the side with less
process parameters such as temperature and moles. In the particular case of the Haber-Bosch
pressure for example. However, the selection of process, the right side (products) contain less moles
these integral parameters depends on other and, as a result, an increase in pressure will lead to
parameters such as inert concentration, feed the formation of more products and this is desired.
amount, percentage of ammonia in feed, conversion However, increasing the pressure excessively could
of reactor…etc. Consequently, before starting the in effect increase temperature inside reactor and
actual designing process, the relationship between hence that could damage the reactor. This will be
all these parameters will discussed and assessed. shown in the next section.
This parameter analysis will lead to selection of 2.4.2 Relationship between Parameters
optimal parameters for HaberVentus specific 2.4.2.1 Reactor Type and Mode of Operation
design. Finally, the reactors can then be sized, Now that the influence of temperature, pressure,
accessorized, supported and configured accordingly and catalyst on rate of reaction is understood, this
[1]. next section will work to address the relationship
2.4.1 Reaction Chemistry + Rate between these parameters and others. Firstly, the
The Haber-Bosch reaction, as mentioned earlier, type of reactor must be identified to generate such a
involves mixing nitrogen and hydrogen over a relationship of parameters through mass/energy
catalyst of choice at moderately high pressures and balances. The reactor type is going to be a fixed bed
temperatures. The reaction is shown below by reactor operating auto-thermally. Additionally, this
equation 1: means that because the reaction itself is highly
exothermic, then the heat produced by the reaction
will be used to heat up the feed to the ignition
temperature. This heat integration is done through
𝑁2(𝑔) + 3𝐻2(𝑔) ↔ 2𝑁𝐻2(𝑔)
the heat-exchangers in the previous section. It must
(1) be noted that autothermal reactors are only limited
−1
∆𝐻 = −92 𝑘𝐽 𝑚𝑜𝑙 to reaction that occur exothermically. Also, the
word “auto-thermal” implies once again that no
The reaction is both reversible and exothermic. In heat is added of removed during stationary
order to consider how each of temperature, operation. If the reaction is going to proceed at
pressure, and catalyst affect the reaction rate then maximum possible conversion (subject to company
Le Chatelier’s principle must be considered. As a design), then equation 2 below describes the
result, in order to maximize the yield of ammonia temperature profile inside the reactor [2]:
which is the goal of this entire process then the
position of the equilibrium must be shifted to the ∆𝑇𝑎𝑑 = 𝑇𝑜𝑢𝑡 − 𝑇𝑖𝑛 (2)
right as much as possible. For temperature, since
the reaction is exothermic, to shift the equilibrium The term ∆𝑇𝑎𝑑 is known as adiabatic temperature
to the right the temperature must be lowered. This increase, it characterizes the amount of heat
drop in temperature effect will favour the forward released in the reaction. Furthermore, since the
reaction (production of Ammonia). This analysis reactor is autothermal, the heat generated can be
justifies why ammonia reactors in general operate at measured by the temperature profile across the
industrially defined “low” reaction temperatures of reactor (in vs out). Now that the reactor type and
400°C to 450°C. Designing for these “low” mode of operation have been selected, the
temperatures affects the rate in a negative as the temperature profile across the reactor will be
rate of the reaction will drop significantly. selected. Moreover, this entails the difference
Additionally, this is not ideal as the industrial world between the outlet and inlet temperature as
is driven by contract and customer demands for discussed above in equation 2.
products to be delivered at a certain date. As a
2.4.2.2 Temperature Profile Optimization
result, catalyst is employed to essentially recuperate
What yields the relationship between temperature
the loss of reaction rate. The more catalyst used
and length of bed is a mass balance done on a small
results in higher reaction rate. Moreover, this point
portion of the fixed bed reactor. On the other hand,
addresses how important is the catalyst to this
what yields the relationship between conversion and
commercial process.
length is the mass balance. However, since the
length of bed is not known at this point, these two
13
equations can be related by equal and cancelling 2.4.2.3 Pressure of Reactors Optimization
bed length. It must be noted that at this stage mass It is assumed that pressure remains constant
and energy balance were not done individually due across each of the reactors. However, realistically
to the current ambiguity of parameter conditions. speaking this is not factual as some pressure drop
However, a design equation (merging the two) can occurs due to contact of reactant gas with reactor
be used in this preliminary stage to help design the walls and hence friction is created. Through
temperature profile. The design equation is as navigating different literature, it is concluded that
follows below. pressure inside the reactor can be calculated as a
function of outlet reactor temperature (in this case
2 ∙ 𝑛𝐴,𝑖𝑛 ∙ ∆ℎ𝑅 reactor 3). Given the outlet temperature determined
∆𝑇 = − ∙ 𝑋𝐴
𝜐𝐴 ∙ 𝑚𝐴,𝑖𝑛 ∙ 𝐶𝑝𝐴 (3) in the last section was 748 K, figure 15 can be used
to determine the pressure value of the three
reactors. The pressure value is found to be exactly
Where: 200 bar. Once again, it must be reiterated that
pressure is presumed to remain constant across the
• Molar feed rate: 𝑛𝐴,𝑖𝑛 three reactors as the operation is considered to be
• Heat of the reaction: ∆ℎ𝑅 isobaric. As a result, in theory, it is assumed that
• Stoichiometric coefficient: 𝜐𝐴 this value of pressure prevails for all three reactors.
• Mass feed flowrate: 𝑚𝐴,𝑖𝑛 However, the pressure drop across each reactor can
• Specific heat capacity of the mixture: 𝐶𝑝𝐴 be quantified once more technical specifications are
• Required Conversion: 𝑋𝐴 found and this is done in the later sections of the
design.
Both the molar and mass rate can be taken from the
mass balances listed in the previous section.
Required conversion signals the total conversion
across the three reactors and hence it is 0.295. For
heat of the reaction, it is calculated by the equation
below [3]:

∆ℎ𝑅 = 𝐹𝑁2 ∙ 𝑋𝐴 ∙ ∆𝐻 (4)

𝐹𝑁2 signals the molar flowrate of nitrogen into the

first reactor, 𝑋𝐴 is the single pass conversion once
again. Lastly, ∆𝐻 is the reaction enthalpy at 298 K
which was referenced in equation 1. With regards to
specific heat capacity of the mixture, knowing that
the reactants enter the reactor at the ignition
temperature of 400°C, subsequently the specific Figure 15: Outlet Temp. vs Pressure relationship
heat values of each of nitrogen, hydrogen, and for Ammonia Reactors
ammonia were calculated at this temperature and
averaged together. The final outcome is ∆𝑇 = 76 as Although a pressure of 200 bar is always safe to
the total temperature profile. As a result, if the feed quote for this particular type of process, a thorough
ignition temperature is 400°C (673 K) and this is check of this value can be done through using the
the temperature of reactants coming into the reactor. pressure value to calculate temperature profile. This
By referring to equation 2 then, it shows that outlet is the reverse of the step done above using figure 15
temperature comes out to 475°C (748 K). It is worth [4]. As a result, figure 16 shows this exact
noting that this is the temperature profile across the relationship, it displays different pressure of values
entire reactor loop (3 reactors). The inlet of 194, 200, 213, and 235 bar against the
temperature describes the point where feed enters corresponding temperature profile. The aim of this
reactor 1. On the other hand, the outlet temperature is to display pressure range operability with 194 as
signifies the temperature of the product ammonia the low, 200 as the normal, and 235 as very high.
leaving the third reactor. Given that the entire Using ignition temperature (673 K) as the inlet,
temperature profile across the reactor has been outlet comes out to 750 K which is not far from
determined, now the accompanying pressure will be HaberVentus outlet temperature of 748 K. The
optimized. check is successful.
14
 2.5.1 Mass Balance (dX/dL)

|𝑢 𝐶 𝐴|𝑥

Reactor

|𝑢 𝐶 𝐴|𝑥+∆𝑥

Figure 16: Temperature profile. vs Pressure Figure 17: Mass Balance being done over a portion
relationship for Ammonia Reactors of the reactor

Now that the temperature profile and pressure value Where:

across the entire reactor set-up were found, it is • 𝑢 is velocity of flow (𝑚 ℎ−1 )
time to progress to the next stage of the design • 𝑐 is concentration of reactant (k𝑚𝑜𝑙 𝑚−3)
which is solving the mass and energy balances. In • 𝐴 is area of reactor (𝑚2 )
addition, there are a couple of more factors that • So: |𝑢 𝐶 𝐴|𝑥 = 𝑚 ℎ−1 ∙ 𝑘𝑚𝑜𝑙 𝑚−3 ∙ 𝑚2
affect both temperature and pressure profiles such
= 𝑘𝑚𝑜𝑙 ℎ−1
as total molar flowrate, percentage of argon in feed,
nitrogen-to-hydrogen rations in feed, and amount of
This represents the molar flowrate
ammonia in feed. However, analysis of these
aspects yielded the same exact result to what was
Since Input = Output + disappearance by reaction as
already determined and hence it was decided not
this is considered in steady state and hence no
include these diagrams here.
accumulation occurs:

2.5 Introduction to Mass and Energy Balance |𝑢 𝐶 𝐴|𝑥 = |𝑢 𝐶 𝐴|𝑥+∆𝑥 + 𝐴 ∙ ∆𝑥 ∙ 𝜐

The mass balance will aim to provide a ∙ 𝑅𝑁𝐻3 ∙ 𝜂 (5)
relationship between the conversion achieved
across the bed and the length of bed required for
each specific reactor. As for the energy balance, it Where:
illustrates the relationship between temperature
• ∆𝑥 = change of space/length (𝑚)
across the bed as the length increases. These
• 𝜐 = stoichiometric coefficient
equations provide a useful tool of mainly sizing the
bed in terms of length, and also help configure what • 𝑅𝑁𝐻3 = rate of reaction for ammonia
temperatures are achieved between the beds. In the production (𝑘𝑚𝑜𝑙 𝑚−3 𝑠 −1 )
previous section, it was disclosed how the total • 𝜂 is the effectiveness of the catalyst
temperature profile is calculated but not across each
bed/reactor. These equations developed by each
respective basis serve as the elementary basis of |𝑢 𝐶 𝐴|𝑥 − |𝑢 𝐶 𝐴|𝑥+∆𝑥 − 𝐴 ∙ ∆𝑥 ∙ 𝜐 ∙ 𝑅𝑁𝐻3
design of an ammonia reactor. It is very essential as ∙𝜂 =0 (6)
it provides the fundamental aspect of design.
Following this stage of design, more specific
considerations can be assessed and hence the design Divide everything by 𝐴∆𝑥
becomes more complicated. This is the intermediate
stage between parameter optimization and final |𝑢 𝐶 𝐴|𝑥 − |𝑢 𝐶 𝐴|𝑥+∆𝑥
design considerations. + 𝜐 ∙ 𝑅𝑁𝐻3 ∙ 𝜂 = 0 (7)
∆𝑥

15
 2.5.2 Energy Balance (dT/dL)
As ∆𝑥 → 0
|𝐹𝑡𝑜𝑡𝑎𝑙 |𝑥

𝑑𝐶
𝑢 = 𝜐 ∙ 𝑅𝑁𝐻3 ∙ 𝜂 (8)
𝑑𝑥

Then if the general conversion equation is

considered: Reactor
𝐶0 − 𝐶1
𝑋=
𝐶0 (9)

This can be restructured: |𝐹𝑡𝑜𝑡𝑎𝑙 |𝑥+∆𝑥

𝐶0 − 𝐶1 = 𝑋(𝐶0 )

∆𝐶 = 𝑋(𝐶0 ) Figure 18: Mass Balance being done over a portion

of the reactor
𝑑𝐶 = 𝑋(𝐶0 ) (10)
In this case, F is total flowrate of all reactants. Once
again, since Input = Output + disappearance by
reaction as this is considered in steady state and
Then equation 8 can be re-written after the equation
hence no accumulation occurs:
10 is substituted into it:

𝑑𝑋 |𝐴 𝐹𝑡𝑜𝑡𝑎𝑙 𝐶𝑝𝑚𝑖𝑥 𝑇|
𝑥
𝑢 𝐶0 = 𝜐 ∙ 𝑅𝑁𝐻3 ∙ 𝜂 (11) − |𝐴 𝐹𝑡𝑜𝑡𝑎𝑙 𝐶𝑝𝑚𝑖𝑥 𝑇| (15)
𝑑𝐿 𝑥+∆𝑥
= 𝜂 ∙ (−∆𝐻𝑅 ) ∙ 𝑅𝑁𝐻3 ∙ 𝐴 ∙ ∆𝑥
Also, notice that 𝑑𝑥 → 𝑑𝐿.
Where:
Then, since 𝑢 𝐶 𝐴 is equal to molar flowrate as • 𝐶𝑝𝑚𝑖𝑥 is the specific heat capacity of
shown below: reactant mixture (𝐽 𝑘𝑚𝑜𝑙 −1 𝑘𝑔−1 𝐾 −1)
• |𝐴 𝐹𝑡𝑜𝑡𝑎𝑙 𝐶𝑝𝑚𝑖𝑥 𝑇| = 𝑚2 ∙ 𝑘𝑚𝑜𝑙 ℎ−1 ∙
𝐹𝑁2 = 𝑢 𝐶 𝐴 (12) 𝑥
𝐽 𝑘𝑚𝑜𝑙 −1 𝑘𝑔−1 𝐾 −1 ∙ 𝐾 = 𝑘𝐽
This can be substituted into equation 11 and further
re-arranged to get: Divide everything by 𝐴∆𝑥

𝑑𝑋 𝜐 ∙ 𝑅𝑁𝐻3 ∙ 𝜂 ∙ 𝐴 |𝐴 𝐹𝑡𝑜𝑡𝑎𝑙 𝐶𝑝𝑚𝑖𝑥 𝑇| − |𝐴 𝐹𝑡𝑜𝑡𝑎𝑙 𝐶𝑝𝑚𝑖𝑥 𝑇|

𝑥 𝑥+∆𝑥
= (13)
𝑑𝐿 𝐹𝑁2 ∆𝑥 (16)
= 𝜂 ∙ (−∆𝐻𝑅 ) ∙ 𝑅𝑁𝐻3
Finally, because 1/2 a mole of nitrogen is required
to make one mole of ammonia then 𝜐 = 1/2. The After further rearrangement:
final mass balance equation is written below:
𝑑𝑇
𝐹𝑡𝑜𝑡𝑎𝑙 𝐶𝑝𝑚𝑖𝑥 = 𝜂 ∙ (−∆𝐻𝑅 ) ∙ 𝑅𝑁𝐻3 (17)
𝑑𝑋 𝑅𝑁𝐻3 ∙ 𝜂 ∙ 𝐴 𝑑𝐿
= (14)
𝑑𝐿 2 𝐹𝑁2
Finally, the energy balance comes out to:
In the next sub-section, the energy balance will be 𝑑𝑇 𝜂 ∙ (−∆𝐻𝑅 ) ∙ 𝑅𝑁𝐻3
done. This is because many of the same factors = (18)
present in the mass balance equation are also 𝑑𝐿 𝐹𝑡𝑜𝑡𝑎𝑙 𝐶𝑝𝑚𝑖𝑥
present in the energy balance.
16
2.5.3 Calculating unknown variables of Subsequently, the rate equation for this set of
Mass/Energy Balance particular reactions is:
Now that both the mass and energy balance
equations were found, in this section the unknown
variables will be calculated. The goal is to quantify (28)
these variables and then solve both the mass and
energy balances. To iterate what must be found,
both equations are listed again: Where:
𝑑𝑋 𝑅𝑁𝐻3 ∙ 𝜂 ∙ 𝐴 • 𝑘 is the rate constant for the reaction
= (19)
𝑑𝐿 2 𝐹𝑁2 • 𝑘𝑎 is the equilibrium constant
• 𝛼𝑥 is the activity for each component (x
signifies component)
𝑑𝑇 𝜂 ∙ (−∆𝐻𝑅 ) ∙ 𝑅𝑁𝐻3 • ∝ is the reaction constant (usually between
= (20)
𝑑𝐿 𝐹𝑡𝑜𝑡𝑎𝑙 𝐶𝑝𝑚𝑖𝑥 0.5 and 0.75)

Need to find: The activity terms are expressed by using the

activation coefficient as seen below [6]:
• 𝑅𝑁𝐻3
• 𝜂 𝑓𝑥
𝑎𝑥 = (29)
• 𝐶𝑝𝑚𝑖𝑥 𝑓𝑥0
• 𝐴
Where:
2.5.3.1 Rate Calculation
• 𝑓𝑥 is the individual fugacity for each
In order to calculate the rate of the reaction, the
component
chemistry and kinetics aspect of the reaction must
• 𝑓𝑥0 is the reference fugacity (take as 1 atm
be highlighted. As mentioned earlier, the overall
in this case)
reaction happening inside the reactor is:
Given the consideration that reference fugacity can
𝑁2(𝑔) + 3𝐻2(𝑔) ↔ 2𝑁𝐻2(𝑔) (21) be taken as 1 atm, equation 29 is re-modified:
The catalyst of choice is iron (as mentioned earlier), 𝑓𝑥 𝑓𝑥
and hence the rate equation can be modelled based 𝑎𝑥 = = = 𝑦𝑥 ∙ 𝜙𝑥 ∙ 𝑃 (30)
𝑓𝑥0 1
on reactant nitrogen actually absorbing onto the
iron catalyst surface. To further understand this,
Where:
below are the sequence of reactions that undertake
• 𝑦𝑥 is the mole fraction of each individual
and ultimately forms ammonia (reaction of
component x (gas)
reactants + products with catalyst):
• 𝜙𝑥 is the fugacity coefficient, again for each
𝑁2(𝑔) → 2𝑁(𝑎𝑑𝑠) component x
(22)
• 𝑃 is the pressure of the gas mixture entering
the reactor (found to be 200 in previous
𝐻2(𝑔) → 2𝐻(𝑎𝑑𝑠) (23) section)

𝑁(𝑎𝑑𝑠) + 𝐻(𝑎𝑑𝑠) → 𝑁𝐻(𝑎𝑑𝑠) (24) For 𝑦𝑥 , it represents the final mole fraction after the
reaction takes place. This variable depends on
𝑁𝐻(𝑎𝑑𝑠) + 𝐻(𝑎𝑑𝑠) = 𝑁𝐻2(𝑎𝑑𝑠) (25) initial mole fraction (𝑌𝑁2 ) and how much of it
actually converts. Additionally, it must also factor
in stoichiometric coefficients of the reactants
𝑁𝐻2(𝑎𝑑𝑠) + 𝐻(𝑎𝑑𝑠) = 𝑁𝐻3(𝑎𝑑𝑠) (26) involved in the reaction. This step is essentially
doing a mole balance over each reactor and
𝑁𝐻3(𝑎𝑑𝑠) → 𝑁𝐻3(𝑔) presenting them as mole percentage as shown in the
(27)
next page.

17
After doing a quick mole balance around the 𝜙𝑁2 = 0.94231827
reactor, the below equations are used to calculate + 0.2129547 × 10−3 𝑇
the final mole percentage of each component: + 28678509 × 10−3 𝑃
(34)
− 0.280826 × 10−6 𝑇 2
𝑌𝑁2 (1 − 𝑋) + 0.4886308 × 10−6 𝑃2
𝑦𝑁2 = (31)
1 − 2 𝑋 𝑌𝑁2
𝜙H2 = exp {exp (−3.7301 𝑇 0.125
𝑌𝐻2 − 3 𝑋 𝑌𝑁2 + 0.543) × 𝑃
𝑦𝐻2 = (32)
1 − 2 𝑋 𝑌𝑁2 − exp (−0.1374 𝑇 0.5
− 16.99) × 𝑃2 (35)
𝑌𝑁𝐻3 + 2 𝑋 𝑌𝑁2 + 314
𝑦𝑁𝐻3 = (33) × exp (−0.02291 𝑇
1 − 2 𝑋 𝑌𝑁2
− 5.943) × exp (−P
After inputting values from the mass balance, the /300)}}
table below shows the results for the mol fraction
values for each component throughout the three 𝜙𝑁𝐻3 = 0.1327886
beds: + 0.3038548 × 10−2 𝑇
− 0.4397572 × 10−3 𝑃 (36)
− 0.1233954 × 10−5 𝑇 2
+ 0.2851314 × 10−6 𝑃2
Comp. Reactor 1 Reactor 2 Reactor 3
For the rate constant, since it is a reversible
reaction, it is found through connection to the
𝒚𝑵𝟐 22.9% 21.5% 20.1% Arrhenius relationship as shown below:

𝐸𝑎
𝒚𝑯𝟐 68.7% 64.5% 60.2% 𝑘 = 𝑘0 exp ( ) (37)
𝑅𝑇
𝒚𝑵𝑯𝟑 0.08% 14.0% 19.7% Where:

Table 6: Final mole fraction of all components • 𝑘0 is Arrhenius coefficient = 8.85 × 1014
throughout the three reactors • 𝐸𝑎 is the activation energy =
170550 𝑘𝐽 𝑘𝑚𝑜𝑙 −1
What can be seen through this table that the fraction • 𝑅 is the universal gas constant = 8.314 kJ
of ammonia out of total mixture is steadily 𝑘𝑚𝑜𝑙 −1 𝐾 −1
increasing across each reactor. It also illustrates
how low the conversion is. Additionally, this is Finally, for the equilibrium constant 𝑘𝑎 , it is really
typical for ammonia reactors since the maximum dependent on the nature of the reaction. In this
conversion achieved in a single bed is realistically particular case, the Haber-Bosch reaction to form
20% (at HaberVentus it is 11%). Moving on, now ammonia is both reversible and exothermic as
the fugacity coefficient for each component must be mentioned previously. Given that it is reversible,
calculated. It was found through literature that the the backward reaction is possible if the conditions
equations (34-36) are normally used, they are warrant it, and it is ammonia decomposition. What
shown in the next column. also needs to be considered is the physical effects as
a consequence of the reaction. Additionally, when
Now that both the final mole fraction percentages the pressure and lowering temperature to favour
and fugacity coefficients of each component across ammonia production actually decreases the volume
all reactors were calculated, the activity is found by and subsequently the number of moles of gas. The
multiplying those two by the operational pressure. bigger picture is all these factors affect the
Also, after inspecting the values of the calculated equilibrium, then a mathematical model must be
activities, ∝ (reaction constant) is chosen at 0.5 for developed to represent this relationship. Given all
this process. Finally, both rate constant k and these considerations, an equation of the equilibrium
equilibrium constant 𝑘𝑎 will now be calculated [7]. constant is taken from literature:

18
 log10 𝑘𝑎 = −2.792312 log 𝑇 2.5.3.2 Effectiveness Factor Calculation
− 5.527463 × 10−5 𝑇 The effectiveness factor is a measure of the
+ 1.837742 × 10−5 𝑇 2 (38) functionality of the catalyst. Furthermore, ideally it
2002.4 is desired for the concentration and temperature
+ + 2.799 throughout the catalyst pellet to be the same as
𝑇
those at the surface. This means the catalyst
Now that the equilibrium constant is calculated for maintaining perfect temperature profile and
each of the reactors, the rate of reaction for each concertation/composition throughout the entirety of
reactor can be found. Once again, the rate equation: the reaction. However, this is just an “ideal”
situation, in reality there are many things that could
affect the catalyst-reactant encounter such as
catalyst poisoning, sintering, and deactivation. Back
(39)
to the point, effectiveness factor measures the rate
of the actual reaction against the rate if it
maintained “ideal” conditions. Before entering
The calculated values of the variables are shown further details about the calculation of the
below, this is only for the first reactor just to give effectiveness factor, some more properties of the
the reader an idea. The same exact equations apply catalyst are disclosed for better calculation
to the other two beds but what changes is mainly implementation.
mole fraction percentage, temperature, pressure due
to pressure drop, and flowrates of reactants. It is an iron-oxide catalyst (𝐹𝑒2 𝑂3 − 𝐹𝑒𝑂). The
manufacturing procedure is done through the fusion
Variable calculation for Reactor 1: of magnetite with the help of potassium, calcium,
and alumina oxides as promoters. The next step is
• 𝑘 = 51.1 reduction within situ. Oxygen is the removed, and
• 𝑘𝑎 = 0.32 hence the catalyst structure becomes porous iron
• 𝛼𝑁2 = 47.1 surrounded by promoters within the interphases
• 𝛼𝐻2 = 145.9 between the crystals and pore walls. Further details:
• 𝛼𝑁𝐻3 = 15.1
• Particle Size: 6-10 mm (Granules)
• ∝= 0.5 • Bulk Density: 2.65 kg/L
• Particle Density: 4.9 kg/𝑚3
Reaction Rate Across the Three
Typically-speaking, the rate equation presented in
Reactors
equation 39 is a summation of how temperature,
15 pressure, and composition affect the reaction for a
Reaction Rate

fine catalyst with no diffusion restriction. In

10
addition, it assumes that kinetic collisions are not
5
possible. However, as mentioned above this
assuming “ideal” conditions. More specific to this
0 process, since the catalyst is iron based and has a
Reactor 1 Reactor 2 Reactor 3 particle size in the range of 6-10 mm, this is not
true. Literature states that catalyst within that
particle size range is subject to restrictions of
Figure 19: Reaction Rate across ammonia reactors
diffusion mainly within pore structure. Hence, the
rate equation must be multiplied by the
The figure above shows how the reaction rate
effectiveness factor to account for these
changes as the reactants progress through the three
modifications. One source suggested equation (40)
reactors. It seems that there is a sharp drop from
as means to calculate effectiveness factor, but it is
reactor 1 to reactor 2 but then a shallower drop to
evaluated at reactor pressures of 160, 220, and 305
reactor 3. It is evident that the rate decreases
atm. The constants shown in the equation on the
throughout each reactor even if it marginal. The
next page are based on pressure. Since HaberVentus
only logical explanation for this as the reaction
reactor operates at 200 atm, the constants at 220 at,
proceeds in each reactor, the concentration of
were chosen and used in the calculation of the
reactants (nitrogen and hydrogen) drop. On the
effectiveness factor.
other, ammonia concentration increases [8].
19
 𝜂 = 𝑏0 + 𝑏1 𝑇 + 𝑏2 𝑋 + 𝑏3 𝑇 2 + 𝑏4 𝑋 2 2.6 Solving Mass and Energy balances
(40)
+ 𝑏5 𝑇 3 + 𝑏6 𝑋 3 Once again to reiterate, the equations are as
below:
Where: (constants are selected for 220 atm)
𝑑𝑋 𝑅𝑁𝐻3 ∙ 𝜂 ∙ 𝐴
• 𝑏𝑜 = −8.21 = (43)
𝑑𝐿 2 𝐹𝑁2
• 𝑏1 = 0.038
• 𝑏2 = 6.19
• 𝑏3 = −5.35 𝑑𝑇 𝜂 ∙ (−∆𝐻𝑅 ) ∙ 𝑅𝑁𝐻3
• 𝑏4 = −20.8 = (44)
𝑑𝐿 𝐹𝑡𝑜𝑡𝑎𝑙 𝐶𝑝𝑚𝑖𝑥
• 𝑏5 = 2.38 × 10−8
• 𝑏6 = 27.9 A MATLAB code was written to solve both
equations, the function “ode45” was used. This
The final outcome was that the effectiveness factor function is used to solve a differential equation.
ranged from 58-65% across the three reactors, with This yielded the respective relationships between
it decreasing after every reactor run. This decrease reactor conversion and reactor length. In addition,
in effectiveness factor can most certainly be the relationship between temperature across reactor
attributed to deactivation of catalyst as well as length.
potential catalyst poisoning. More of this will be
discussed in the catalyst deactivation section Firstly, the mass balance results for each reactor:
coming up later.
2.5.3.3 Specific Heat of the Mixture
This section entails calculating the specific heat Conversion vs. Length for Bed 1
capacity of the reactant mixture which includes 0.45
nitrogen, hydrogen, ammonia, and inert. The way 0.4
this done is through the equation below [9]: 0.35
Conversion

0.3
0.25
𝐶𝑝 = ∑ 𝑦𝑖 𝐶𝑝𝑖 (41) 0.2
0.15
0.1
Where 𝑦𝑖 represents mole fraction of each 0.05
component and 𝐶𝑝𝑖 is the specific heat of the 0
component. If all the terms are added in, a function 0 2 4 6 8 10 12
for the whole mixture can be computed [31]. This Length of Bed (m)
function is based on temperature and pressure of
each component and is shown below:
Figure 20: Conversion across length of bed 1
𝐶𝑝𝑚𝑖𝑥 = 35.31 + 0.02𝑇
+ 0.00000694 𝑇 2 (42)
Given that the achieved conversion in reactor 1 was
− 0.0056 𝑃 11%, the selected bed length is subsequently 2 m. It
+ 0.000014 𝑃 𝑇 is worth pointing out that this only represents length
of catalyst bed and not the whole reactor. The
2.5.3.4 Reactor Area Calculation operation of a fixed bed reactor overseas a vessel
The mass balance required a value for the area for it filled with catalyst pellets, there is space between
be solved. However, at this stage of the design, it is the vessel head and the catalyst bed, this is termed
difficult to assume an area value as the reactors bed voidage. The reactants begin to flow inside the
have not been sized yet. What was done was an reactor where it then meets the catalyst bed, this is
iteration method. Given that all the other values where the reaction between reactants and catalyst
within the mass balance were found, the iteration undergoes (adsorption first then desorption). Once
method inputted random values of A until the the reaction is complete, the products along with
numbers made sense. This concludes computation unconverted feed progress to the subsequent reactor
of variables that constitute both mass and energy where the same operation occurs again. The catalyst
balances, now the equations will be solved, and the bed can be set up in many ways such as powder,
outcomes will be displayed. pellets, or even small tubes containing catalyst.

20
 Moving on to the results of the energy balance.
Conversion vs. Length for Bed 2 Now that the desired bed length is known for each
0.3 reactor, the temperature profile can be determined.
0.25

0.2 Temperature vs. Length for Bed 1

Conversion

0.15 760
750
0.1 740

Temperature (K)
730
0.05
720
0 710
0 2 4 6 8 10 12 700
690
Length
680
670
660
Figure 21: Conversion across length of bed 2 0 2 4 6 8 10 12
Bed Length (m)
For this reactor (2), once again conversion was 11%
and hence bed length is 2.5 m.
Figure 23: Temperature profile across Bed 1
Conversion vs. Length for Bed 3 As a result, for the first bed, the feed enters at 673
0.4 K which is the auto-ignition temperature of the
0.35 reactants. Subsequently, if the bed length is 2 m
0.3 then the temperature reached will be 702 K.
Conversion

0.25 However, this is too high to enter reactor 2 and

0.2 hence it must be quenched. More will be covered
0.15 within the quench design section, but after
0.1 quenching 702 K decreases down to 688 K. The
0.05 quench comes in the form of cold unconverted fresh
0 feed of syngas.
0 2 4 6 8 10 12
Length of Bed (m)
Temperature vs. Length for Bed 2
760
Figure 22: Conversion across length of bed 3 750
Temperature (K)

740
Finally, for bed 3, the required height of the catalyst 730
bed is 2.7 m as conversion was 11%. 720
710
This analysis helps identify how long does the 700
690
catalyst bed for each reactor have to be. As the
680
reactant progresses through each reactor, the bed
0 2 4 6 8 10 12
length increases from 2 m to 2.5 m and then to 2.7
Bed Length (m)
m. In addition, this increase in bed length is down
to rate of reaction decreasing across the three
reactors. As the reaction progresses, the reactants
Figure 24: Temperature profile across Bed 2
are being used up more and more, hence the
concentration of reactants decreases significantly.
As mentioned above, outlet stream of reactor 1 is
Subsequently, this is why the rate of the reaction
originally 702 K but this is cooled down to 688 K.
drops across the three reactors. In addition, more
As a result, stream into reactor 2 is 688 K, and
catalyst is required to recuperate the reaction rate
given the bed length for reactor 2 was 2.5 m, this
and keep the reaction industrially feasible. The
corresponds to an outlet temperature of 722 K. This
amount of catalyst required for each reactor will be
temperature is cooled down to 710 K using quench
quantized in the next section [10].
design
21
 • 𝑅𝑁𝐻2 is the rate of the reaction
Temperature vs. Length for Bed 3 (𝑚𝑚𝑜𝑙 𝑔−1 ℎ𝑟 −1)
780
770 It can be seen that mass of catalyst depends on
760 amount of reactant, also how much of it converts. In
Temperature (K)

750 addition, mass of catalyst also depends on the rate

740
of the reaction. The less the rate of the reaction, the
more catalyst it requires. This equation is then
730
solved for each of the reactors and their respective
720
parameters, the results are shown below:
710
700 • Reactor 1: 2572.8 𝑘𝑔 𝑜𝑓 𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡
0 2 4 6 8 10 12
• Reactor 2: 3210.1 𝑘𝑔 𝑜𝑓 𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡
Length (m)
• Reactor 3: 3315.4 𝑘𝑔 𝑜𝑓 𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡

What can be noticed is that the need for catalyst

Figure 25: Temperature profile across Bed 3 increases more as the reaction progresses across the
reactor system. Conversion remains the same for all
Finally, for reactor 3, as mentioned previously three reactors and hence it not decisive. However,
reactor 2 outlet stream is at 710 K after cooling. both reaction rate and nitrogen feed flowrate
Moreover, this means the stream inlet to reactor 3 is decrease as the reaction progresses to the next
710 K. After the reaction occurs in bed 3, which is reactor/bed. Once again, this is because
2.7 m long, a temperature of 748 K is reached as concentration of the reactants drops meanwhile
shown above in the figure. This matches the overall products concentrations increase as they are being
outlet temperature of the reactor system calculated formed. Subsequently, the rate of the reaction drops
in section 2.3.1 which was done through a typical and hence it requires more catalyst to recuperate
design equation. The non-discrepancy is significant feasibility. However, it is interesting to compare
as it highlights that practical calculations match reactor 1 and 2 for example in terms of reaction and
theoretical analysis [11]. This concludes mass of catalyst. For reactor 1 the rate is
temperature profile modelling across the reactors. 11.6 𝑚𝑚𝑜𝑙 𝑔−1 ℎ𝑟 −1 and it requires
In addition, this is the conclusion of stage two of
2572.8 𝑘𝑔 𝑜𝑓 𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡. On the other hand, reactor
the design where parameters such as bed lengths
2 has a reaction rate of 9.3 𝑚𝑚𝑜𝑙 𝑔−1 ℎ𝑟 −1 ,
and temperature profile were calculated. Finally,
however it requires 637.3 kg more catalyst. This
these calculations can now be used for stage 3 of
even a small difference in reaction (∆=2.3)
the design which focuses on mechanical design of
accounts for a significant amount of more catalyst.
equipment and sizing.
3.2 Fixed Bed structure
3 Reactor Sizing The way of dispersing the catalyst in a fixed bed
can range from a number of different methods,
3.1 Quantifying amount of catalyst for the however for this design procedure a multi-tubular
reactors reactor structure is chosen as it best suited as the
It is important to quantify how much catalyst is catalyst amounts calculated in the previous section
needed as this will help decide the structure of the signify large amounts. The structure will be a good
fixed bed, and also how it delivers this catalyst. Fur handful of cylindrical tubes carrying catalyst
The design equation used to quantify amount of particles held up vertically in a vertical pressure
catalyst is shown below vessel. The overall procedure is that syngas feed
enters the reactor from the top (head section) and
𝐹𝑁2 𝑋𝑁2 flow down to these catalyst-filled tubes, then it
𝑚𝑐𝑎𝑡 = (45)
𝑅𝑁𝐻2 flows out from the bottom of the bed. What happens
after the first reactor is that it enters the next reactor
Where: from the top once again, then it flows down from
• 𝐹𝑁2 is the flowrate of nitrogen to each the bottom as product. This procedure is repeated
reactor (𝑘𝑚𝑜𝑙 ℎ−1 ) for the third final reactor. Figure 26 a is good
representation [12] of how the reactor should look
• 𝑋𝑁2 is how much nitrogen converts in each
like, more detailed sketches will be shown later.
reactor

22
 The reason why reactors 2 and 3 require more tubes
than reactor 1 is because they contain more mass of
catalyst. As a result, to maintain a unform
distribution of catalyst, more tubes are used. Below
is another sketch detailing the sizing of the catalyst
tubes, note that there will 200 tubes of this size for
example for reactor 1. Normally, these tubes are
surrounded by what is called the shell, it will be
sized in the next section.

Figure 26: Basic Sketch of the Reactor System

Now, given that the amount of catalyst has been

calculated and subsequently the fixed bed structure, L
it is time to quantify the number of catalyst tubes
and size them.
3.3 Design of Catalyst Tubes
The focus of this analysis will be how many tubes d
are required, and hence how small/large should the
tube be. Before considering how many tubes are
needed, the tubes must be sized first. From the
analysis on required bed length within the reactor, it
is known how long each bed should be. For
example, for reactor 1 the bed length was found to
be 2 m. As a result, this is the length of each
cylindrical tube. The catalyst tube diameter was
chosen to be 15 cm, and this is constant for all beds.
Now, the number of tubes can be calculated by the Figure 27: Sizing of Catalyst Tubes Inside Reactor
equation below [13]:
3.4 Shell Diameter
2
𝜋 𝑑𝑡𝑢𝑏𝑒 𝐿𝑡𝑢𝑏𝑒 The shell is the casing for the tubes, it is what
𝑛𝑡𝑢𝑏𝑒𝑠 = 𝜌𝑐𝑎𝑡 (46)
4 bundles them together and prevents them from
moving. The shell also aims to equally space out the
Where: tubes so that again they do not move randomly and
disrupt the process [29]. To calculate the diameter
• 𝑑𝑡𝑢𝑏𝑒 is the diameter of the tube of this shell, the below equation is used:
• 𝐿𝑡𝑢𝑏𝑒 is the length of the tube
• 𝜌𝑐𝑎𝑡 is the density of the catalyst. 𝑛𝑡𝑢𝑏𝑒𝑠
𝐷𝑠 = 𝑑𝑡𝑢𝑏𝑒 √ (47)
𝜙
Inputting all the necessary variables for each
reactor, the results are shown below.
Where:
• Reactor 1: 200 𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡 𝑡𝑢𝑏𝑒𝑠
• Reactor 2: 250 𝑐𝑎𝑡𝑙𝑦𝑠𝑡 𝑡𝑢𝑏𝑒𝑠 • 𝐷𝑠 is the shell diameter
• Reactor 3: 270 𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡 𝑡𝑢𝑏𝑒𝑠 • 𝜙 is the perforation fraction
23
The perforation fraction is representative of the 3.5 Total Length of Reactor
fraction of the tube that has holes in it to allow for The total length of the reactor is the length of the
better catalyst contact with reactant [25]. The higher whole unit including catalyst tubes, head section,
the perforation factor, the more holes are required. and shell. In order to calculate the length of the
Subsequently, the equation for perforation factor is reactor, the following equation used [14].
shown below.
𝜋
𝜙= (48) 𝐿𝑟𝑒𝑎𝑐𝑡𝑜𝑟 = 𝐿𝑡𝑢𝑏𝑒 + 𝐷𝑠 (49)
2𝑎2 √3

Where: The assumption is that adding the shell diameter to

the length of the existing tubes will constitute a fair
• 𝑎 𝑖𝑠 𝑡ℎ𝑒 𝑡𝑢𝑏𝑒 𝑠𝑝𝑎𝑐𝑖𝑛𝑔 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 approximation of the reactor. In order to visualize
this, a sketch is used once again.
The tube spacing coefficient is multiplied against
the diameter of the tube to determine the space
between centre of each tube. So, 𝑑𝑡𝑢𝑏𝑒 × 𝑎 is the
amount of spacing between each tube within the
bundle. The value for 𝑎 is determined to be 1.5. 𝐿2
Therefore, the space between each tube for the
reactors is 1.35 m. Using the value of tube spacing
coefficient, the perforation fraction is then
calculated. Finally, the shell diameter can then also
be calculated. Take reactor 1 as an example, tube
length is 2 m with diameter 0.15 m, in total there 𝐿1
are 200 tubes. The shell shown in red in the below
figure almost acts a wall for the tubes. Using the
equations above, the shell diameter is 3 m. 𝐿3

𝐷𝑆

Figure 29: Total Length of Reactor Illustration

Referring to figure 29, the total length of the reactor

𝐿3 is found by the summation of 𝐿1 (tube length)
and 𝐿2 (shell diameter). This method of finding the
total length of the reactor is reasonable, however
after all it is an estimate. The tubes are standing
vertically and hence there is no problem in using
tube length to size total length. Moreover, the
downside is the shell diameter as it extends
horizontally. When looking at figure 29, it seems by
using shell diameter and adding it to tube length,
the reactor volume might be slightly over-
Figure 28: Shell cast over Catalyst Tubes Inside estimated. However, this is not a significant
Reactor + Diameter problem as some space has to allowed for bed
voidage (empty space).
24
3.6 Volume of Catalyst
For the volume of the catalyst, it can be calculated Now that most of the physical components within
through the equation below. the reactor bed are calculated, a quick sketch of
only the reactor can be made to highlight design
𝑚𝑐𝑎𝑡 features up to this point.
𝑉𝑐𝑎𝑡 = (50)
𝜌𝑐𝑎𝑡

The mass of the catalyst was found a couple of

sections ago, as for the density of the catalyst it was
provided in the section for effectiveness factor
calculation. Overall, using both mass and density
values, the volume occupied by the catalyst packed
into the small tubes can be found. It is important to
iterate that this volume represents the entire catalyst
volume and not just for one tube.
3.7 Diameter of the bed
The diameter of the bed entails the combined 2m
diameter of the entire tube collection and can be
calculated by the equation below [15].
5m
4 𝑉𝑐𝑎𝑡 0.5
𝐷𝑏𝑒𝑑 =( ) (51)
𝜋 𝐿𝑏𝑒𝑑
a
Inserting all the variables, the diameter can be
calculated. For reference, if reactor 1 is taken as an
example, the diameter of the bed was 1.4 m. On the
other hand, the shell diameter was 3 m. It can be
seen that the shell diameter is twice as much the bed
diameter. This observation is important as the shell
encompasses all the tubes, and hence naturally it
should have a much bigger diameter. Also, one of
the important design considerations is that the tubes
will not be lined-up next to each other in a straight
line, they will be bundled together. In addition, 200
tubes with a total diameter of 3 m seems 3m
underestimated to the naked eye and hence this is
why it is important to state this point.
3.8 Total Reactor Volume
The volume of the whole reactor including all the
contents can be calculated by the equation below.

𝐷𝑠 2
𝜋 𝐷𝑆2 𝐿𝑡𝑢𝑏𝑒 4𝜋 ( 2 ) (52)
𝑉𝑟𝑒𝑎𝑐𝑡𝑜𝑟 = +
4 3 Figure 30: Complete sketch of Reactor

The pressure vessel takes up a cylindrical shape, as The sketch above details what the values calculated
does the catalyst tubes. The shell casing the tubes in the last couple of sections represent. This
can also be considered a cylinder (thick). Therefore, concludes this section which aimed to size the
the equation is based on the volume of the cylinder. reactor and its contents. As a result, the basic
Once again, taking reactor 1 as a reference, the structure of the reaction was determined in addition
reactor volume is 13 𝑚3. On the other hand, the to quantifying the amount of catalyst present in the
catalyst volume is 3 𝑚3 . From this the bed voidage reactor. This section provides a solid basis for the
can be calculated, and it is around 60%. This next section which deals with accessory
represents the empty space in the reactor. calculations for supports, heads, and thicknesses.

25
Before starting the section that details designing of accessories for the reactor, below is a table detailing the key
physical parameter findings for each reactor:

Parameter Reactor 1 Reactor 2 Reactor 3

𝒎𝒄𝒂𝒕 2572.8 𝑘𝑔 3210.1 𝑘𝑔 3315.4 𝑘𝑔

𝑹𝑵𝑯𝟐 11.6 𝑚𝑚𝑜𝑙 𝑔−1 ℎ𝑟 −1 9.3 𝑚𝑚𝑜𝑙 𝑔−1 ℎ𝑟 −1 8.99 𝑚𝑚𝑜𝑙 𝑔−1 ℎ𝑟 −1

𝑳𝒕𝒖𝒃𝒆 2.0 𝑚 2.5 𝑚 2.7 𝑚

𝒅𝒕𝒖𝒃𝒆 0.15 𝑚 0.15 𝑚 0.15 𝑚

𝒏𝒕𝒖𝒃𝒆𝒔 200 250 270

𝑫𝒔 3.0 𝑚 3.5 𝑚 3.7 𝑚

𝑳𝒓𝒆𝒂𝒄𝒕𝒐𝒓 5.0 𝑚 6.0 𝑚 6.4 𝑚

𝐃𝒓𝒆𝒂𝒄𝒕𝒐𝒓 3.2 𝑚 3.7 𝑚 3.9 𝑚

𝑽𝒄𝒂𝒕 3.0 𝑚3 3.77 𝑚3 3.9 𝑚3

𝑫𝒃𝒆𝒅 1.4 𝑚 2.1 𝑚 2.3 𝑚

𝑽𝒓𝒆𝒂𝒄𝒕𝒐𝒓 13.0 𝑚3 21.5 𝑚3 25.5 𝑚3

Table 7: Summary of Parameters Results for Reactor

26
4 Reactor accessories calculations The design procedure then for reactor wall
thickness is based on quantifying stress
4.1 Shell thickness + Wall Thickness of Vessel
circumferentially against axial stress. The larger
This will aim to quantify how thick should the wall stress is then used and quantified to then find wall
of the vessel be. To quickly recap the reactor thickness. The reasoning behind the defined design
structure, it is a pressure vessel equipped with tubes procedure is that by accounting for the larger stress
filled with catalyst bundled together and surrounded and design wall thickness based on that, the reactor
by a large shell. As a result, the wall thickness of can withstand such stresses with ease. The
the vessel represents the outer surface. Once again, equations for both stresses are shown below [16]:
the reactor sketch is used to highlight exactly what
is being done. 𝑃𝑑
𝑓𝑐 = (55)
2𝑡
Reactor
Wall 𝑃𝑑
𝑓𝑎 = (56)
4𝑡

Where:

• 𝑓𝑐 is the circumferential stress

• 𝑓𝑎 is the stress in the axial direction
• 𝑃 is the pressure inside the reactor
• 𝑑 is the shell diameter
• 𝑡 is the reactor wall thickness

Once again if reactor 1 is used a reference, and the

variable values are inputted bar wall thickness, so
the answer is in relation to wall thickness
(multiple). Circumferential stress turns out to be
larger than axial stress, and hence it is stress
selected for design. The shell inside the reactor is
produced as a result of joint in the axial direction,
and hence the efficiency of this action is termed
joint efficiency. Therefore, the thickness of the shell
Figure 31: Sketch illustrating Reactor Wall can be calculated by the below equation.

𝑃 𝐷𝑖
The design equations for the reactor wall thickness 𝑡𝑠ℎ𝑒𝑙𝑙 = (57)
𝑓
depend on the ratio of outer diameter to inner 2 𝑠 𝑗−𝑃
diameter as shown below.
Where:
𝐷0
< 1.5 (53)
𝐷𝑖 • 𝑡𝑠ℎ𝑒𝑙𝑙 is the thickness of the shell
• 𝑓 is allowed stress at 400°C for this design
• 𝑆 is factor of safety (taken as 1.5)
If the ration is less than 1.5, a particular method
follows, whereas if the ratio is larger than 1.5 then Inputting all these values for reactor 1, the outcome
another method is used. In this design’s particular is shell thickness of 94.1 mm. In order to get reactor
case, outer diameter represents diameter of the wall thickness, shell thickness is raised by
whole reactor, meanwhile inner diameter is the consideration of corrosion. In addition, since
shell diameter. If reactor 1 is once again used as a corrosion of the wall is damaging and hence
reference, then: unwanted, an extra couple of millimetres are added
to shell thickness to avoid that and this yields
𝐷0 3.5 reactor wall thickness. This figure is chosen to be
= < 1.5 (54)
𝐷𝑖 3.0 1.7 mm and final wall thickness is 95.8 mm.
27
4.2 Head Thickness Calculation 4.3 Skirt Support Thickness
In this section, the thickness of the head of the Support for a pressure vessel comes in the form of
vessel will be calculated. The head is the top a skirt. This piece of equipment is designed so that
surface of the reactor enclosing the contents. It has the reactor can be supported vertically and covered
many different geometries for different up for extra layer protection. In addition, the skirt
functionalities such as tori-spherical for example. also allows the reactor to be at the correct height.
However, for this particular vessel, a sphere-shaped The geometry of a skirt differs for different
head was chosen as the pressure profile inside the functionalities, however for this design a cylindrical
reactor is split equally with the head centre as a skirt is chosen. The skirt almost serves as skin to
point of view. reactor, however there are added design features
that aid design control. For example, the skirt is
equipped with nozzles for whatever input is
required to the reactor. The skirt is also equipped
with a manhole for during shutdown of the unit, a
process technician can assess the inside of the
reactor through this manhole. As for how the skirt
is added to the reactor set-up, the bottom head of
the vessel is welded onto the skirt which then serves
as a solid foundation for the reactor. The sketch of
the skirt is shown below [17].

Figure 32: Remodified Reactor after Head

Consideration

Therefore, the equation for a spherical head of

reactor thickness is shown below:
Figure 33: Cylindrical Skirt Sketch for the Reactor
𝑃 𝐷0 𝐵
𝑡ℎ =
𝑓 (58) As for the procedure of designing skirt thickness, it
4 𝑆+𝑃
is based on the maximum stresses it could face.
First of all, there is the stress due to the weight of
the reactor and its contents. Secondly, the skirt
Where B is a geometrical factor. Inputting all the could also face a seismic load in the case of an
variables once again using Reactor 1 values, head earthquake. Then, quantifying and using those two
thickness comes out to 51.9 mm. However, once stresses, the maximum tensile strength at the
again to be safe against corrosion, 1.7 mm is added. bottom and maximum compressive strength can be
The final value for the head thickness is 53.6 as found and then used to calculate how thick does the
shown below: skirt need to be.
51.9 + 1.7 = 53.6 𝑚𝑚 (59)

28
4.3.1 Dead Weight Stress 𝑀𝑎𝑥 𝑇𝑒𝑛𝑠𝑖𝑙𝑒 𝑆𝑡𝑟𝑒𝑛𝑔𝑡ℎ𝑏𝑜𝑡𝑡𝑜𝑚
(64)
This is the stress due to the reactor and its = 𝑓𝑠 − 𝑓𝑑𝑤
contents. It is expressed in the equation below.
It is based on the difference between total possible
𝑤 deadweight stress and seismic load stress. If the
𝑓𝑑𝑤 = (60)
𝜋 𝐷𝑠𝑘𝑖𝑟𝑡 𝑡𝑠 variables for reactor 1 is inputted, the following
calculations occurs:
Where:
• 𝑓𝑑𝑤 is the overall stress due to the reactor 𝑀𝑎𝑥 𝑇𝑒𝑛𝑠𝑖𝑙𝑒 𝑆𝑡𝑟𝑒𝑛𝑔𝑡ℎ𝑏𝑜𝑡𝑡𝑜𝑚
deadweight 66,070.9 50,093.7
= −
• 𝑊 is the weight of the reactor combined 𝑡𝑠 𝑡𝑠 (65)
with its components 15,977.2
= 𝑘𝑔 𝑚−2
• 𝐷𝑠𝑘𝑖𝑟𝑡 is skirt diameter 𝑡𝑠
• 𝑡𝑠 is the thickness of the skirt
Then, this overall value must be divided by the
The weight of the reactor was calculated by using maximum tensile strength of the material to account
the physical dimensions and relating it to a weight for how many more millimetres does the skirt
value. The mass of catalyst was already calculated thickness needs to equip. The material for the
so it was added. Finally, skirt diameter was reactor has been selected as a Molybdenum-
assumed to have a diameter of 1 m longer than the Vanadium equipped as an alloy on a steel plate.
reactor. Plugging in all the values for reactor 1, the This material is termed ASME SA542 Grade A
value of the dead-weight stress is: Class 4 and it is ideal for pressure vessels that
achieve high temperatures. As a result, the
50,093.7 maximum tensile strength for material is:
𝑓𝑑𝑤 = 𝑘𝑔 𝑚−2 (61)
𝑡𝑠
4.3.2 Seismic Load Stress 𝑀𝑎𝑥 𝑇𝑒𝑛𝑠𝑖𝑙𝑒 𝑆𝑡𝑟𝑒𝑛𝑔𝑡ℎ𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙
73,929,445 (66)
Seismic loads created a result of an earthquake = 𝑘𝑔 𝑚−2
can significantly affect the reactor and disrupt its 𝑡𝑠
structure, the main outcome would that the seismic As a result, to obtain how much should the final
load would offset the reactor off the skirt. In order skirt thickness be extended by, maximum tensile
to measure the stress that is achieved due to a strength at the bottom is divided by the material’s
possible earthquake, the below equation is used: maximum strength, this yields 0.004 m or 4 mm.

8𝐶𝑊𝐻 Now, in order to calculate what the original skirt

𝑓𝑠 = (62) thickness is before further extension, first the
3 𝜋 𝐷02 𝑡𝑠
maximum compressive strength is found as below:
Where:
𝑀𝑎𝑥 𝐶𝑜𝑚𝑝. 𝑆𝑡𝑟𝑒𝑛𝑔𝑡ℎ = 𝑓𝑠 + 𝑓𝑑𝑤 (67)
• 𝑓𝑠 is the seismic load of stress
• 𝐶 is the seismic coefficient (between 0.1 – 116,164.6
The resulting value is 𝑘𝑔 𝑚−2 . In addition,
0.3) 𝑡𝑠
this must be divided by 1/4th of the maximum yield
59,623,400
In this particular case, C is taken as 0.1. As a result, strength of the material which is 𝑘𝑔 𝑚−2.
𝑡 𝑠
if all the variables are inputted for reactor 1, then
the following value for seismic load is found: 59,623,400 149114429
÷4= 𝑘𝑔 𝑚−2 (68)
66,070.9 𝑡𝑠 𝑡𝑠
𝑓𝑠 = 𝑘𝑔 𝑚−2 (63)
𝑡𝑠
Finally, the maximum compressive strength is
4.3.3 Final Skirt Thickness Calculation divided by value to yield the skirt thickness as
Given that both possible deadweight and seismic below (skirt extension is added):
stresses are found, the total skirt thickness can be
found. Now, what must be found is maximum 116,164.6
tensile strength at the bottom of the reactor, this is 𝑡𝑠 = = 7.8 𝑚𝑚 + 4 𝑚𝑚 (69)
149114429
given by the equation in the next column. = 11.8 𝑚𝑚

29
 Figure 34: Final Sketch for Reactor

Accessory Reactor 1 Reactor 2 Reactor 3

Reactor Wall Thickness 95.8 𝑚𝑚 105.2 𝑚𝑚 107.1 𝑚𝑚

Shell Thickness 94.1 𝑚𝑚 103.5 𝑚𝑚 105.4 𝑚𝑚

Head Thickness 53.6 𝑚𝑚 57.6 𝑚𝑚 59.3 𝑚𝑚

Skirt Thickness 11.8 𝑚𝑚 18 𝑚𝑚 20 𝑚𝑚

Skirt Height 2𝑚 4𝑚 5𝑚

Table 8: Summary of Accessories Sizing for Reactor

30
5 Extra Specifications for Reactor This value in metric tons is converted to moles per
day.
5.1 Catalyst Lifetime
The iron-based catalyst’s functionality reduces
38,000 𝑘𝑔 𝑜𝑓 𝑁2 𝑎𝑛𝑑 𝐻2
when deactivation occurs. The deactivation depends
→ 1.3 × 108 𝑚𝑜𝑙 𝑑−1 (71)
on the operating conditions of the reactor, mainly
→ 4.1 × 108 𝑚𝑜𝑙 𝑦 −1
temperature and pressure. However, the most
As a result, the resultant sulphur concentration can
significant factor is the feed gas and what it could
be calculated for the theoretical situation described:
possibly contain. The magnetite catalyst could
possibly undergo activation due to sintering; this is 20
where access to the catalyst pore is blocked by 𝑆𝑢𝑙𝑝ℎ𝑢𝑟 𝐶𝑜𝑛𝑐. = (72)
potentially poisonous gases within the feed stream. 4.1 × 108
= 4.8 × 10−9 𝑉/𝑉
These poisonous species initiate sintering when
they chemisorb onto the catalyst structure, then
Consequently, the concentration even in this
these deposits block access for the reactants to enter
hypothetical situation is extremely low meaning it
catalyst pores. Another possible route to catalyst
will not affect the reactor. Also, there is no existing
deactivation is poisoning, these are undesired
sulphur in any unit within the plant, and therefore
gaseous materials in the feed or even possibly solid
the risks of sulphur entering the reactor are very
materials inserted during manufacturing. There are
low. Another potential permanent poison is
generally two types of poisons within the catalyst
chlorine. As a poison, chlorine does not affect the
field of study, permanent and temporary poisons.
iron catalyst structure but rather the potassium used
The permanent poisons, such as sulphur, are
to promote it. If another hypothetical situation is
detrimental as they induce the catalyst to
used to quantify the amount of possible chlorine
irreversibly lose catalytic activity. On the other
present, it is assumed that chlorine occupies 1 wt%
hand, temporary poisons only work to lower the
of the syngas. This accounts to 150 kg of chlorine.
catalyst activity. Examples of temporary poisons for
Again, if it assumed that 2% of this is most
this reactor would be water, carbon oxides, and
effective, then this equates to 4 kg. The 4 kg of
oxygen. Finally, the last factor that could affect the
chlorine equates to 100 moles total. Then, if this
catalyst activity could be fouling. Fouling refers to
value is divided by total syngas feed:
solid residues depositing onto the catalyst surface,
in effect blocking the entry of reactants onto the 100
catalyst pores. However, it must be noted that it is 𝐶ℎ𝑙𝑜𝑟𝑖𝑛𝑒 𝐶𝑜𝑛𝑐. = (73)
extremely rare for fouling to occur within the 4.1 × 108
= 2.4 × 10−7 𝑉/𝑉
ammonia synthesis, as it is limited to reactions
involving oils [18].
Once again this is a very low possible
concentration, and also to consider than no chlorine
The main focus of this design will be assessing the
is used in the plant it is very rare that chlorine
catalyst poisons. First, starting with permanent
poisoning occurs. Moving onto temporary poisons,
poison sulphur. If it is assumed that 3% of the
most commonly carbon oxides and oxygen. If
catalyst of the first reactor is to be poisoned by
hydrogen feed gas encounters any carbon
sulphur. Then, 5500 ppm of sulphur is considered
monoxide, the following reaction occurs:
as the maximum capacity of sulphur. Furthermore,
to calculate the active surface area the following
3𝐻2 + 𝐶𝑂 → 𝐶𝐻4 + 𝐻2 𝑂 (74)
equation is used:

𝑆𝑢𝑙𝑝ℎ𝑢𝑟 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦 Therefore, any water formed will be indicative of

𝐴𝑠𝑢𝑟𝑓𝑎𝑐𝑒 = (70)
the presence of these oxides. Water is actually used
32 in the catalyst manufacturing process where a
= 3125 𝑚𝑜𝑙
percent of water oxidizes the iron into an oxide or
magnetite. Then the magnetite is oxygen-reduced to
This value represents the total number of gas moles
form the iron-based catalyst later on in the
on that surface, there the sulphur content would
manufacturing process. Therefore, this reaction is
then be 20 mol per year. Then in order to relate this
unwanted and hence no water must be allowed into
to total ammonia produced, this value is compared
the reactor because it will convert the catalyst back
to total make-up gas for one year. The reactor
to its early structure. Hence, the possibility of this
system uses 38 m.t/d of syngas.
reaction must be worked against.
31
It is found that hydrogen sulphide in some Therefore, it is perfectly reasonable that the
proportion to water is required to achieve this ammonia catalyst lifetime will look to be replaced
equilibrium between iron catalyst and its oxide, every 10 years, and not a shorter lifetime since the
therefore this value must be found. Through poisoning hazards are very insubstantial. Below is a
literature, this value is found to be: picture of the catalyst that is going to be used.

𝐻2 𝑆
= 4.8 × 10−5 (75)
𝐻2

It was found that formation of sulphur inside the

reactor was very unlikely, and then this value above
represents that hydrogen sulphide in small
proportion is required. Hence, it is highly unlikely
that is duo will be present simultaneously inside the
reactor. In addition, it is stated this reaction only
occurs at 700 K, although inside the reactor system
this temperature is reached and exceeded, it is still
highly unlikely that it will occur. Finally, the water Figure 35: HaberVentus Catalyst Structure
can also be quantified with respect to the hydrogen
present in the feed to see how much of it is needed 5.2 Pressure Drop Across Reactors
to affect the catalyst [19]: In order to estimate the pressure, drop across the
fixed bed reactor, the Ergun equation is used.
𝐻2 𝑂 (76) ∆𝑃
= 0.15 (77)
𝐻2 = 𝑣𝑖𝑠𝑐𝑜𝑢𝑠 𝑒𝑛𝑒𝑟𝑔𝑦 𝑙𝑜𝑠𝑠
𝐿
+ 𝑘𝑖𝑛𝑒𝑡𝑖𝑐 𝑒𝑛𝑒𝑟𝑔𝑦 𝑙𝑜𝑠𝑠
This means that 15% of the hydrogen total feed 150 𝜇(1 − 𝜀)2
input must be water, this accounts to 122 kg/hr of = 𝜐𝑠
𝐷𝑃2 × 𝜀 3
water. This amount of water is highly unlikely to be
present in the reactor. The only water source in the 𝜐𝑠2 𝜌(1 − 𝜀)
+ 1.5
whole plant is the electrolysis unit which takes 𝐷𝑝 𝜀 3
desalinated water from MVC but yields 99%
hydrogen. In conclusion, it can be assumed that no Where:
water will be present inside the reactor, and hence • ∆𝑃 is pressure drop
the hazard of the catalyst returning to its initial • 𝐿 is the reactor length
shape during manufacturing is almost non-existent. • 𝜇 is viscosity
• 𝜀 is bed voidage
Lastly, the final temporary poison to consider is • 𝐷𝑝 is particle diameter
oxygen. It is found that oxygen is proportional to • 𝜌 is particle density
𝐻 𝑂
the amount of 2 yielded since the only source of • 𝜐𝑠 is superficial gas velocity
𝐻2
oxygen would be water. However, oxygen would
only be formed at 225°C – 275°C, where oxygen
Pressure Drop Across the Three
would transition into becoming permanent.
However, the initial temperature of the first reactor Beds
is 400°C and hence it is very far away from this 0.8
reaction temperature. It was already discussed why
Pressure drop

𝐻 𝑂 0.6
the 𝐻2 proportion is almost impossible to occur,
2 0.4
and since oxygen rate production is proportional to
0.2
this, the same applies. It is made sure that oxygen
does not enter the reactor as PSA, an oxygen 0
Bed 1 Bed 2 Bed 3
producing unit, yields 99.9% of nitrogen. Also, the
electrolysis unit, another oxygen producing unit
completely separates all the oxygen before Figure 36: Pressure drop for the 3 Beds
transitioning to the separators ahead.
32
The reason for why pressure drop decreases as the Next, the quench is designed for the reactor 3 inlet.
reaction progresses through the series of reactors is After the reaction occurs in reactor 2, the products
because of amount of reactant feed. The reactants formed leave at 722 K with a total mass flowrate of
steadily decrease in each due to more formation of 4844.36 kg/hr. Once again, this temperature is
ammonia occurring. Therefore, less reactants means termed too high for reactor 3 and hence must be
less contact with the walls of the reactor creating cooled. A quick energy balance is done:
less friction, and subsequently less pressure drop.
5.3 Quench System Design 𝑄𝑄2 + 𝑄1𝐵 = 𝑄1.2 (81)
Very high temperatures are achieved inside the 𝑚2 𝑇2 + 𝑚1𝐵 𝑇1𝐵 = 𝑚1.2 𝑇1.2 (82)
reactor and this potentially damage the catalyst.
Therefore, after reactors 1 and 2, a quench stream is
introduced to cool down these streams and prepare Yet again there will be two unknowns in the form
them for the next reactor. The quench comes in the of the quench streams specification. A value for
𝑚2 = 1000 𝑘𝑔/ℎ𝑟 is assumed. After that, the equation
form of unconverted fresh feed (syngas). The
is solved for the unknown parameter 𝑇2 and this is found
details of the quench design for reactor 1 outlet are to be 652 K. Therefore, a quick recap, outlet stream of
now considered. From a thermodynamics point of reactor 2 leaves at a temperature of 722 K, to cool that a
view, if a quick energy balance is done considering quench stream of 1000 kg/hr unconverted gas is
the streams as thermal reservoirs: introduced at 652 K, and this reduces the outlet stream
to 710 K. As a result, the stream then enters reactor 3 at
𝑄𝑄1 + 𝑄1𝐴 = 𝑄1.1 (78) 710 K where the final reaction occurs, then products of
that reactor leave at 748 K.
𝑚1 𝐶𝑝 𝑇1 + 𝑚1𝐴 𝐶𝑝 𝑇1𝐴 = 𝑚1.1 𝐶𝑝 𝑇1.1 (79) The implementation of the quench streams has already
been added to the mass balance shown in the beginning
of this design report. However, the freshly made
The specific heat capacity can be cancelled out as is
available feed is at 673 K, but the quench streams are
it is the same exact mixture considered (syngas). slightly cooler than that. Therefore, an extra cooler must
be implemented after the fresh feed supply. This cooler
𝑚1 𝑇1 + 𝑚1𝐴 𝑇1𝐴 = 𝑚1.1 𝑇1.1 (80) is going to be small unit since it only needs to cool the
fresh feed by a small amount, it needs to cool the feed
Considering that for reactor 1, the outlet stream from 673 K to 610 K. As for the second quench stream,
it must be cooled down to 652 K. The downside of this
leaves at 702 K and the total mass flowrate is
quench design is that it dropped the respective
4844.36 kg/hr. Then it must be cooled down to 688 conversions for all three reactors. The addition of more
K. The only unknowns will the 𝑚1𝐴 𝑎𝑛𝑑 𝑇1𝐴 , and unconverted feed makes it harder for the reactor to
both represent the quench stream. Therefore, it is convert reactants into more products. The reason for the
assumed that 𝑚1𝐴 = 1000 𝑘𝑔/ℎ𝑟. Then, if the equation conversion limitation is because the reactor has only a
is solved for the unknown variable which is temperature finite amount of catalyst to work with, and hence extra
of the quench stream, it comes out to 610 K. Therefore, feed would require extra catalyst. The reason for the
1000 kg/hr at 610 K must quench the outlet stream conversion limitation is because the reactor has only a
leaving the first reactor. This quench stream is added to finite amount of catalyst to work with, and hence extra
the pipe connecting reactors 1 and 2 where the two feed would require extra catalyst.
streams mix to form the inlet to reactor 2.

𝑚1𝐴 , 𝑇1𝐴 𝑚1.1 , 𝑇1.1 𝑚1𝐵 , 𝑇1𝐵 𝑚1.2 , 𝑇1.2

1 1 1
1

1B 1.2 2
1 1A 1.1
Fresh Feed Reactor 1 Reactor 2 Reactor 3

Q1 𝑚𝑞1 , 𝑇1 Q2 𝑚𝑞2 , 𝑇2
1

Figure 37: Reactors Quench Design

33
6 Process Control for Reactor

Figure 38: Reactors Control System

• Reactor 1 (RB-501):
transmitter FT-5041 would order control valve CV-
The fresh stream is first compressed in a
501 to stop the flow of fresh feed into the reactor.
compressor (CM-501) where the pressure of the
CV-501 is also used to manipulate temperature
syngas increases to 200 bar, as per reactor
control. If the temperature of the reactant mixture
requirement. Before the fresh feed enters the first
coming is less than the required auto-ignition
reactor, pressure controller PC-101 reads the
temperature, then temperature transmitter TT-5041
current pressure for any deviations. For instance, if
would send a signal to CV-501 to open more and
the compressor failed for some reason to prepare
stop flow. Then, pressure valve PV-503 is installed
syngas at 200 bar, and say it was 205, then this
to assess pressure of the mixture after the recycle
controller would be able to spot that deviation from
stream has been incorporated. The transmitter for
set-point. Then for example, if the pressure was
this controller is connected to the control valve that
actually 205 bar which is 5 bar higher than
controls the recycle stream, CV-504. Lastly, for the
required, transmitter PT-101 would order pressure
input side, an isolation valve IV-502 is equipped in
valve PV-501 to release some pressure and return to
case of the immediate urge needed to stop flow into
200 bar. This pressure control is prior to fresh feed
reactor 1. Once the reaction is complete, another
mixing with recycle stream. After fresh feed mixes
isolation valve is equipped to stop flow into reactor
with recycle stream, before the feed enters the
2 if necessary. The outlet of reactor is assessed for
reactor 1 inlet, a flow composition analysis is done
composition (FC-50142), temperature (TC-50142),
it through controller FC-5041. If for example there
and pressure (PC-50142) before entering reactor 2.
was a little more argon that required, the flow
All these controllers are connected to valve CV-502
34
which control the flow of quench stream into the 7 Pump and Valve Sizing
mixing point of reactor 1 outlet and this quench
stream. The quench is needed to drop reactor 1 7.1 Pump Sizing
outlet temperature of 702 K to 688 K. Therefore, Given that the node design mainly revolved
for example, if the temperature controller reads that around the reactor system, it was difficult to
the outlet for reactor 1 was 710 K, it closes the consider designing a pump inside this node. This is
valve to allow more flow of quench stream. It is because the node deals with very high pressures and
also worth noting that the quench stream is temperatures, and hence it is difficult to process a
equipped with pressure controller (PC-501) as it pump that is able to handle those conditions. As a
was compressed (CM-502) to 200 bar. result, a pump connecting the storage tank to the
separator is chosen instead (as shown in the figure
• Reactor 2 (RB-502): on the next page). This is the case if the ammonia
kept in the storage tank requires further processing.

The reactant entry to reactor 2 is equipped by an Firstly, the static head is calculated [20]:
isolation valve IV-503 in the case that inlet flow to
∆𝑃 (83)
this reactor must be stopped. After the reaction ℎ𝑠 = +∆𝑍
occurs, there is another isolation valve IV-504 𝜌𝑔
equipped if flow to reactor 3 must be stopped. Then Where:
the product stream leaving reactor 2 is assessed for
flow (FC-50143), temperature (TC-50143), and • ℎ𝑠 is static head
pressure (PC-50143). All these controllers are • ∆𝑃 is pressure difference
connected to control valve CV-503, which is also • 𝜌 is density
connected to the quench stream. Reactor 2 outlet • ∆ 𝑍 is difference in elevation
comes out at 722 K and must be cooled to 710 K
• 𝑔 is the gravitation constant
for reactor 3 inlet, therefore CV-503 is used to
manipulate process conditions towards that desired
The pressure in the storage tank is 5 bar, and the
cooling process.
distillation column is 35 bar. Density of ammonia is
already known. Static head comes out to 9.4 m.
• Reactor 3 (RB-503):
Then, the dynamic head is found:
Reactor 3 inlet and outlet is equipped with isolation
valves (505 and 506). Then, the outlet stream is to ∆𝑃𝑓 (84)
be purged. Control Valve CV-505 controls how ℎ𝑑 =
𝜌𝑔
much should be purged, and this is based on the
recycle stream going the other where temperature
(TT-504) and flow control (CT-504) is done.
Therefore, if the recycle stream reads more Where:
ammonia than expected then perhaps, we are
purging too much and so CV-505 opens more and • ℎ𝑑 is dynamic head
limits purge flow. • ∆𝑃 is pressure difference for pipe

In conclusion, the main aim of the process control The pipe was not sized for this process, and hence it
delivered was to maintain process conditions such is assumed that the pipe is 50 m long and has an
as temperature and pressure as much of the detailed internal diameter of 80 mm. Then, the Reynolds
design depends on this. By making sure process number of the pipe is found:
conditions are in-tact, the reactor will work
according to specification. As process conditions 𝑝𝑢𝑑 (85)
𝑅𝑒 =
steer away from desired values, the reactor design 𝜇
will then quickly worse than expected. Composition
was also a key factor and was controlled through Where:
flow control. It is important that nitrogen and • 𝑢 is flow velocity
hydrogen remain in the desired stoichiometric
• 𝑑 is pipe diameter
relationship of 1:3. In addition, the inerts must also
• 𝜇 is fluid viscosity
be controlled as they affect temperatures and
conversion.
35
 1.5 m
3m

--------------------------------------------------------------------------------------------------------------------------------------
Figure 39: Pump System

Flow velocity was assumed to be 1 𝑚 𝑠 −1 and so Then figure 41 is used to find the efficiency of the
the Reynolds was found to be 5.2 × 104 after pump:
calculation. The roughness of the pipe was assumed
to be 0.0005 and friction factor was found to be
0.0027. Then pressure loss in pipe is found by
Darcy-Welsbach:

𝐿 𝜌 𝑢2 (86)
∆𝑃𝑠 = 8 𝑓
𝑑 2
Finally, the result is plugged into equation 84 and
dynamic head is found to be 2.04 m, and therefore
total head is:

𝑡𝑜𝑡𝑎𝑙 ℎ𝑒𝑎𝑑 = 9.4 + 2.04 = 11.44 𝑚 (87)

Subsequently, the type of pump is found by the

figure below. Figure 41: Pump Efficiency Graph.

The pump is found to have an efficiency of 73.4%.

Given all the parameters needed to calculate pump
power are found, the following equation is used:

𝑊×𝑚 (88)
𝑃=
𝜂

Where:

• 𝑃 is pumping power
• 𝑊 is work required
• 𝜂 is pump efficiency

The pump power comes out to 4984 kW. Then also,

Figure 40: Pump Configuration Graph. a calculation was made for NPSH, and it came out
to 11.54 m total. The only downside is the
The pump is going to be a single-stage pump estimated values for the pipe dimensions and hence
operating at 35000 rpm. possible discrepancy could occur.

36
7.2 Valve Selection
The valve selected for designing is any of the
pressure relief valves. And so, the choice of valve
for pressure relief is Fisher™ GX Control Valve
and Actuator System, made by the company
Emerson [21].

Table 9: Valve Specification Sheet

The valve is designed as both a controller and

actuator. This means it can read pressure values,
transmit the information to an actuator which
opens/closes the valve per required and so the
pressure changes. The valve is simple to estimate
and choose in terms of size. There is no actuator
estimating required. Determination of support is
easily programmed. The upgraded actuator takes
into consideration a wide scope of air supply. The
actuator is designed for simple support. The
greatest part trait is conveniently shared across all
sizes. Replaceable trim Low lifetime costs. Strong,
low-profile plan. For actuator design, it is a
minimized multi-spring pneumatic actuator. It is
also accessible with incorporated, simple to-align
DVC2000 or DVC6200 computerized valve
regulator. Overall, it is the perfect choice for the
Figure 42: Valve Design ammonia reactor [23].
37
8 Costing 8.3 Method 3
8.1 Method 1
The last method is based on using the reactor
The first method entails estimating the cost by
pressure as a function of the temperature achieved
using a correlation designed for an agitated reactor,
inside the reactor. The following equation is used:
which is also jacketed for quench design. The
method is based on estimating the total volume of 𝑅𝑒𝑎𝑐. 1: 𝐶𝑒,2014 = 275𝑃 + 82,300 (92)
the reactor, the following equation is used [22]:

𝐶𝑒,2010 = 𝑎 + 𝑏𝑆 𝑛 (89) 𝑅𝑒𝑎𝑐. 2: 𝐶𝑒,2014 = 425𝑃 + 83,420 (93)

Where:
• 𝐶𝑒,2010 is cost of reactor 𝑅𝑒𝑎𝑐. 3: 𝐶𝑒,2014 = 440𝑃 + 84,000 (94)
• a = 61500
• b = 32500 And for CEPCI conversion:
• S is the volume
• N is 0.8 𝐶𝐸𝑃𝐶𝐼2021 (95)
𝐶𝑒,2021 = 𝐶𝑒,2014 ( )
𝐶𝐸𝑃𝐶𝐼2014
Then equation 90 is used to convert this to a 2020
value. Where:
𝐶𝐸𝑃𝐶𝐼2021 (90)
𝐶𝑒,2021 = 𝐶𝑒,2010 ( ) • 𝐶𝐸𝑃𝐶𝐼2021 = 614.5
𝐶𝐸𝑃𝐶𝐼2010 • 𝐶𝐸𝑃𝐶𝐼2014 = 576.1

Where: Finally, the following results were obtained:

• 𝐶𝐸𝑃𝐶𝐼2021 = 614.5
• 𝐶𝐸𝑃𝐶𝐼2010 = 532.9 • Reactor 1: £683,958.0
• Reactor 2: £1,023,410.3
The following values were obtained: • Reactor 3: £1,058,099.0

• Reactor 1: £261,697.3 8.4 Method of choice

• Reactor 2: £365,231.4 Method 3 is deemed the best choice because it is
• Reactor 3: £411,106.6 based on temperature and pressure combination.
8.2 Method 2 Method 1 is only based on volume, but volume is
The second method entails using a model not entirely representative of reactor specification,
developed for a pressure vessel, and hence it is and therefore it is underestimated. Different
dependent on the shell mass. As a result, the below reactors could operate differently in terms of
equation is used: amount of catalyst and packing material of
example. In addition, different designers choose
𝐶𝑒,2010 = 𝑎 + 𝑏𝑆 𝑛 (91) different methods of packing the bed and its shape,
it could be a tube or dispersed as a fine powder.
Where: Lastly, the arrangement inside the reactor could
• a = 12,800 differ in terms of bed voidage, space between tubes
• b = 240 or beds, and catalyst shape. Therefore, it is not
• S is the shell mass recommended to use method 1. Then for method 2
• N is 0.9 it uses shell mass as the basis for the calculation.
However, the shell mass can differ significantly in
Then once again, 2010 cost is converted to 2021 different reactors and hence there is no set rule to
cost by the CEPCI’s listed above. size it according to reactor. Also, material can play
a huge role, stronger material will yield denser
The following values were obtained: shell, however it is not representative of the
reactor’s chemical functionality. Finally, most of
• Reactor 1: £1,592,001.4 the design specifications for the reactor are based
• Reactor 2: £1,894,276.1 on temperature and pressure, and so method 3 is
• Reactor 3: £1,927,998.9 best suited to represent the cost of reactors.
38
9 References
- [1]: Rase, H., 2014. Fixed-Bed Reactor Design and Diagnostics. Saint Louis: Elsevier Science.
- [2]: Pattabathula, V., 2020. Simulation of Ammonia Production using HYSYS Software. Chemical and Process
Engineering Research,.
- [3]: Botermans, K., 2018. Modular Ammonia Plant. Undergraduate. Oregon University.
- [4]: AlKusayer, K., 2020. Ammonia Synthesis for Fertilizer Production. Worcester Polytechnic Institute.
- [5]: Kolbain, F., 2020. Renewable Ammonia. ElSevier Science.
- [6]: Lin, B., Wiesner, T. and Malmali, M., 2020. Performance of a Small-Scale Haber Process: A Techno-
Economic Analysis. ACS Sustainable Chemistry & Engineering, 8(41), pp.15517-15531.
- [7]: Afandizadeh, S. and Foumeny, E., 2001. Design of packed bed reactors: guides to catalyst shape, size, and
loading selection. Applied Thermal Engineering, 21(6), pp.669-682.
- [8]: Babu, B. and Angira, R., 2005. Optimal design of an auto-thermal ammonia synthesis reactor. Computers &
Chemical Engineering, 29(5), pp.1041-1045.
- [9]: Azarhoosh, M., Farivar, F. and Ale Ebrahim, H., 2014. Simulation and optimization of a horizontal ammonia
synthesis reactor using genetic algorithm. RSC Adv., 4(26), pp.13419-13429.
- [10]: Joshi, M. and Mahajani, V., 1996. Process Equipment Design. Delhi: Macmillan.
- [11]: Williams, D. and Seipp, T., 2007. Considerations in the Design and Analysis of an ASME Section VIII, Div.
2 Reactor Support Skirt. Journal of Pressure Vessel Technology, 129(2), pp.316-322.
- [12]: Sadeghi, M. and Kavianiboroujeni, A., 2009. The Optimization of an Ammonia Synthesis Reactor Using
Genetic Algorithm. International Journal of Chemical Reactor Engineering, 6(1).
- [13]: Zhang, C., Vasudevan, S. and Rangaiah, G., 2010. Plantwide Control System Design and Performance
Evaluation for Ammonia Synthesis Process. Industrial & Engineering Chemistry Research, 49(24), pp.12538-
12547.
- [14]: Cinar, A., 1983. Design of a Control Scheme for a Catalytic Fixed-Bed Reactor. IFAC Proceedings
Volumes, 16(19), pp.303-311.
- [15]: Khademi, M. and Sabbaghi, R., 2017. Comparison between three types of ammonia synthesis reactor
configurations in terms of cooling methods. Chemical Engineering Research and Design, 128, pp.306-317.-
- [16]: Angira, R., 2011. Simulation and Optimization of an Auto-Thermal Ammonia Synthesis
Reactor. International Journal of Chemical Reactor Engineering, 9(1).
- [17]: Panahandeh, M., Fathikaljahi, J. and Taheri, M., 2003. Steady-State Modeling and Simulation of an Axial-
Radial Ammonia Synthesis Reactor. Chemical Engineering & Technology, 26(6), pp.666-671.
- [18]: Kilpiö, T., Russo, V., Eränen, K. and Salmi, T., 2016. Design and modeling of laboratory scale three-phase
fixed bed reactors. Physical Sciences Reviews, 1(3).
- [19]: Kolios, G., Frauhammer, J. and Eigenberger, G., 2000. Autothermal fixed-bed reactor concepts. Chemical
Engineering Science, 55(24), pp.5945-5967.
- [20]: Cheema, I. and Krewer, U., 2018. Operating envelope of Haber–Bosch process design for power-to-
ammonia. RSC Advances, 8(61), pp.34926-34936.
- [21]: Cheema, I. and Krewer, U., 2019. Optimisation of the Autothermal NH3 Production Process for Power-to-
Ammonia. Processes, 8(1), p.38.
- [22]: Baltrusaitis, J., 2017. Sustainable Ammonia Production. ACS Sustainable Chemistry & Engineering, 5(11),
pp.9527-9527.
- [23]: Devarakonda, M., Holladay, J., Brooks, K., Rassat, S. and Herling, D., 2019. Dynamic Modeling and
Simulation Based Analysis of an Ammonia Borane (AB) Reactor System for Hydrogen Storage. ECS
Transactions, 33(1), pp.1959-1972.
- [24]: Sinnott, R. and Towler, G., n.d. Chemical engineering design.
- [25]: Energy Production and Consumption in the United States. Education.
- [26]: Ammonia properties. Pub Chem. Ammonia Datasheet
- [27]: The Facts About Ammonia. The Health Ny. Yorker Times
- [28]: Ammonia Uses, 2014. Health & Safety Organization
- [29]: Casale, 2014. Axial Ammonia Reactor. ASCE Publications.
- [30]: Joshi V, 2012. Ammonia Reactor Detail and Specification. ResearchGate.
- [31]: Kelly Jackson, 2013. Pumping Designs and Calculation. ASCE Publications.

39
10 Appendix

40
41
42
43
44
 45

View publication stats