Detailed Modelling and Optimisation of An Ethylene Plant
Detailed Modelling and Optimisation of An Ethylene Plant
Detailed Modelling and Optimisation of An Ethylene Plant
Chemical Engineering
Examination Committee
Chairperson:
Supervisors:
October 2013
ii
iii
iv
Acknowledgements
First of all, I would like to express my gratitude to Professor Carla Pinheiro, Professor Henrique
Matos and especially Professor Costas Pantelides for giving me the chance to work in Process Systems
Enterprise for seven months and learn a lot about gPROMS ModelBuilder.
I would also like to thank my housemates Pedro, Vasco and Tiago for easing my homesickness
and for playing an important role during the time we spent in London.
A word of appreciation goes also to all my friends from Instituto Superior Tcnico, especially my
best friends Bernardo and Ins for all theyve done for me in the past five years.
Im also thankful to my PSE supervisor, Maarten Nauta, for being such a selfless and down to
earth guy and for helping me every time I needed help.
Also, huge thanks to my family, especially my parents and my brother for everything theyve
taught me in twenty three years of life and for investing in my education.
Last, but not least, I would like to thank Mariana, for her patience, her belief in me and for
keeping up with me in the last seven months weve been apart.
vi
Abstract
This work comprises the detailed modelling and optimisation of an ethylene plant processing
fresh propane and recycle streams of ethane and propane.
Propane is converted in a steam cracker, operating at low pressure, but high temperature. Then,
the effluent is quickly chilled in one or more transfer line exchangers and then in a quenching tower.
After that, the cracked gas goes into a compression train, with intercoolers, to remove water, and is
dried in molecular sieves. Finally, the water-free cracked gas goes into the distillation column train, in
order to separate and recover valuable products, like ethylene and propylene.
This process is energy demanding: temperatures go as high as 900 C and as low as -121 oC
and the key of it is the choice and design of the right utilities.
Due to the cryogenic temperatures, at least one refrigerant has to be used in order to chill the
cracked gas and condense some of the distillate streams. In this case, two refrigeration cycles were
designed: an ethylene refrigeration cycle and a propylene refrigeration cycle. Ethylene refrigerant is
condensed using propylene refrigerant, whereas propylene refrigerant is condensed using cooling
water.
Finally, some key aspects of the designed flowsheet were subjected to an optimisation, and the
total annualized cost was reduced from 87.67 M$/yr to 50.55 M$/yr, corresponding to a reduction of
42.34% on the total annualized cost.
Keywords
Ethylene plant, steam cracking, olefins, alkenes, gPROMS, refrigeration
vii
viii
Resumo
O presente trabalho consiste no modelo detalhado e otimizao de uma fbrica de olefinas que
processa uma alimentao fresca de propano e reciclagens de etano e propano.
O propano convertido num steam cracker, que opera a baixas presses, mas altas
temperaturas. O efluente reacional rapidamente arrefecido num ou mais transfer line exchangers e,
depois, numa torre de quenching. Posteriormente, o gs de cracking segue para uma unidade de
compresso, com arrefecimento intermdio, de forma a remover-se parte da gua presente e depois
secado em peneiros moleculares. Finalmente, o gs de cracking livre de gua segue para a unidade
de colunas de destilao, de forma a recuperar-se produtos de alto valor, como o etileno ou propileno.
Este processo requer o uso de muita energia: as temperaturas vo desde 121 C at
900 C e a chave do processo a escolha e design das utilidades adequadas.
Devido s temperaturas criognicas, pelo menos um refrigerante tem de ser usado para
arrefecer o gs de cracking e para condensar algumas correntes de destilado. No presente projeto, dois
ciclos de refrigerao foram modelados: um ciclo de refrigerao de etileno e um ciclo de refrigerao
de propileno, sendo que a condensao do etileno conseguida usando propileno, enquanto que o
propileno condensado com gua de arrefecimento.
Finalmente, alguns aspetos-chave do flowsheet elaborado foram sujeitos a uma otimizao,
onde o custo total anualizado foi reduzido de 87.67 M$/ano para 50.55 M$/ano, correspondendo a uma
reduo de 42.34% no custo total anualizado.
Palavras-chave
Fbrica de olefinas, steam cracking, olefinas, alcenos, gPROMS, refrigerao
ix
Contents
1.
2.
Introduction ......................................................................................................................................1
1.1
Motivation ................................................................................................................................1
1.2
1.3
1.4
Background......................................................................................................................................3
2.1
2.1.1
2.1.2
2.1.3
2.1.4
Compression..................................................................................................................13
2.1.5
2.1.6
Drying ............................................................................................................................14
2.1.7
2.1.8
Fractionation ..................................................................................................................15
2.1.9
Refrigeration ..................................................................................................................16
2.2
3.
3.2
3.2.1
3.2.2
3.3
4.
4.2
4.2.1
4.2.2
4.3
4.3.1
4.3.2
4.4
4.5
4.6
4.6.1
Mixers ............................................................................................................................33
4.6.2
4.6.3
JT Valves .......................................................................................................................33
xi
4.6.4
5.
7.
8.
5.1.1
5.1.2
5.1.3
5.1.4
5.1.5
5.1.6
5.1.7
5.2
6.
5.2.1
5.2.2
5.2.3
Demethaniser ................................................................................................................56
5.2.4
Deethaniser ...................................................................................................................57
5.2.5
C2 Splitter .......................................................................................................................57
5.2.6
Depropaniser .................................................................................................................58
5.2.7
C3 Splitter .......................................................................................................................58
5.2.8
5.2.9
6.2
Case 1 ...................................................................................................................................67
7.2
Case 2 ...................................................................................................................................70
7.3
Case 3 ...................................................................................................................................72
7.4
Conclusions ...........................................................................................................................77
8.2
Bibliography ...........................................................................................................................................81
Appendices .......................................................................................................................................... A-1
A-1. Final Flowsheet (A3 Size) ........................................................................................................ A-1
A-2. Shortcut Column Model ........................................................................................................... A-3
A-3. Heat Exchanger Detailed Sizing Model ................................................................................... A-5
A-4. First Optimisation Case Inputs and Outputs ............................................................................ A-8
A-5. Second Optimisation Case Inputs and Outputs ....................................................................... A-9
xii
List of Figures
Figure 1: Simplified flowsheet of an ethylene plant based on steam cracking [5]...................................3
Figure 2: Flowsheet of an ethylene plant with a gaseous feedstock [3]. .................................................4
Figure 3: Flowsheet of an ethylene plant with a liquid feedstock [3]. ......................................................5
Figure 4: Cracking example - ethane being converted to ethylene [6]. ...................................................5
Figure 5: Ethylene's molecular structure. ................................................................................................6
Figure 6: Propylene's molecular structure. ..............................................................................................7
Figure 7: Thermodynamic stability of hydrocarbons [1]. .........................................................................8
Figure 8: Main reactions involved in hydrocarbon pyrolysis [1]. ..............................................................9
Figure 9: Furnace scheme [11]. ............................................................................................................12
Figure 10: Industrial TLE (Alfa Laval) [12]. ............................................................................................12
Figure 11: Water quenching tower [13]. ................................................................................................13
Figure 12: Ethylene plant compressor train (Siemens) [14]. .................................................................13
Figure 13: Scrubber scheme [16]. .........................................................................................................14
Figure 14: Structure of the molecular sieve [18]....................................................................................15
Figure 15: Fractionation train [1]. ..........................................................................................................16
Figure 16: Recovery of ethylene/propylene process from FCC Off-gas [20]. .......................................18
Figure 17: MTO process flowsheet [21]. ...............................................................................................19
Figure 18: DCC Process simplified diagram [22]. .................................................................................19
Figure 19: 218.15K isotherm P-x for the ethylene-ethane binary pair. .................................................24
Figure 20: Model development workflow. ..............................................................................................24
Figure 21: Reactor model. .....................................................................................................................25
Figure 22: Two-stream heat exchanger model......................................................................................26
Figure 23: Water quenching tower model. ............................................................................................27
Figure 24: Water quenching tower model inside view. ..........................................................................27
Figure 25: Segment of the compression section. ..................................................................................28
Figure 26: Splitter model. ......................................................................................................................29
Figure 27: Cold box model. ...................................................................................................................30
Figure 28: Distillation column model. ....................................................................................................31
Figure 29: Mixer model. .........................................................................................................................33
Figure 30: Stream splitter model. ..........................................................................................................33
Figure 31: JT valve model. ....................................................................................................................33
Figure 32: Conversion reactor model. ...................................................................................................34
Figure 33: Ethylene plant flowsheet from a previous project [4]. ..........................................................35
Figure 34: Replication of the ethylene plant flowsheet from previous project [4]. .................................36
Figure 35: Simulation pressure profile and industrial data [56]. ............................................................37
Figure 36: Simulation gas temperature profile and previous simulation profile [56]. ............................37
Figure 37: Main process flowsheet. .......................................................................................................42
Figure 38: Ethylene refrigeration cycle flowsheet. ................................................................................43
Figure 39: Propylene refrigerant cycle flowsheet. .................................................................................46
Figure 40: Cascade refrigerant cycle flowsheet. ...................................................................................48
Figure 41: Cascade refrigerant cycle flowsheet with two-stream heat exchangers. .............................51
Figure 42: Ethylene plant flowsheet. .....................................................................................................52
Figure 43: Reactor temperature profile. ................................................................................................54
Figure 44: Reactor pressure profile. ......................................................................................................54
Figure 45: Components' mass flowrate profile. .....................................................................................55
Figure 46: Components' mass flowrate profile (continuation). ..............................................................55
Figure 47: Reaction rates' profiles. ........................................................................................................56
Figure 48: Total annualized cost distribution for the base case. ...........................................................64
Figure 49: Total annualized CAPEX distribution for the base case. .....................................................64
Figure 50: Ethylene and propylene flowrates as function of several optimisation iterations. ................67
xiii
Figure 51: Total cost as function of several optimisation iterations (case 1). .......................................68
Figure 52: Total annualized cost distribution after case 1. ....................................................................69
Figure 53: Total annualized CAPEX distribution after case 1. ..............................................................69
Figure 54: Variation of the compressors pressure ratios for several optimisation iterations. ...............70
Figure 55: Intercoolers outlet temperatures as function of several optimisation iterations. .................71
Figure 56: Total cost as function of several optimisation iterations (case 2). .......................................71
Figure 57: Total cost and deviance of optimisation cases from the base case. ....................................75
Figure 58: Final flowsheet (A3 Size). .................................................................................................. A-1
Figure 59: FUG model specifications. ................................................................................................. A-4
Figure 60: Heat exchanger detailed sizing model specifications. ....................................................... A-7
Figure 61: First optimisation case inputs and outputs. ........................................................................ A-8
Figure 62: Second optimisation case inputs and outputs.................................................................... A-9
xiv
List of Tables
Table 1: Ethylene's physical properties [3]. .............................................................................................6
Table 2: Propylene's physical properties [3]. ...........................................................................................7
Table 3: Radical reaction mechanism of ethane thermal cracking [1]. .................................................10
Table 4: Typical dilution ratios for different feedstocks [10]. .................................................................11
Table 5: Comparison between reactor outlet compositions after dehydration.. ....................................38
Table 6: Different levels of ethylene refrigeration. .................................................................................44
Table 7: Different levels of propylene refrigeration. ..............................................................................45
Table 8: Connection between ethylene refrigerant cycle and propylene refrigerant cycle. ..................47
Table 9: Connection between cascade refrigerant cycle and main process. ........................................50
Table 10: Reactor feed and reactor outlet. ............................................................................................53
Table 11: Demethaniser top and bottom results. ..................................................................................56
Table 12: Deethaniser top and bottom results. .....................................................................................57
Table 13: C2 Splitter top and bottom results. .........................................................................................57
Table 14: Depropaniser top and bottom results. ...................................................................................58
Table 15: C3 Splitter top and bottom results. .........................................................................................58
Table 16: Main results for case 1. .........................................................................................................68
Table 17: Main results for case 2. .........................................................................................................72
Table 18: Main results for case 3. .........................................................................................................73
Table 19: Main results of optimisation. ..................................................................................................74
Table 20: List of FUG model main variables and parameters. ............................................................ A-3
Table 21: List of Heat exchanger detailed sizing model main variables and parameters. .................. A-5
xv
xvi
List of Abbreviations
KO Knock-out
FCC Fluid catalytic cracking
DCC Deep catalytic cracking
VGO Vacuum gas oil
W hp Compressor horse power
NLP Non-linear programming
LP Linear programming
MILP Mixed integer linear programming
MINLP Mixed integer non-linear programming
CAPEX Capital expenditure
OPEX Operational expenditure
LMTD Logarithmic mean temperature difference
xvii
xviii
List of Units
P-1
P-2
P-3
P-4
P-5
P-6
P-7
P-8
P-9
P-10
P-11
P-12
P-13
P-14
P-15
P-16
P-17
P-18
P-19
P-20
P-21
P-22
P-23
P-24
P-25
P-26
P-27
P-28
P-29
P-30
P-31
P-32
P-33
P-34
P-35
P-36
P-37
P-38
P-39
P-40
P-41
P-42
P-43
P-44
P-45
P-46
P-47
P-48
P-49
P-50
P-51
P-52
P-53
Main process
Fresh feed
Feed mixer
Steam cracking reactor
Steam source
First transfer line exchanger
Steam sink
Steam source
Second transfer line exchanger
Steam sink
Water quenching tower
Quench water source
Quench water outlet
First compressor stage
Cooling water source
First intercooler
Cooling water sink
First KO drum
First KO liquid outlet
Second compressor stage
Cooling water source
Second intercooler
Cooling water sink
Second KO drum
Second KO liquid outlet
Third compressor stage
Cooling water source
Third intercooler
Cooling water sink
Third KO drum
Third KO liquid outlet
Fourth compressor stage
Cooling water source
Fourth intercooler
Cooling water sink
Fourth KO drum
Fourth KO liquid outlet
Fifth compressor stage
Cooling water source
First cracked gas cooler
Cooling water sink
Molecular dryer
Dried water
Cooling water source
Second cracked gas cooler
Cooling water sink
Third cracked gas cooler
Tail gas
Hydrogen rich gas
First multistream heat exchanger
First cold box interstage cooler
First cold box KO drum
Second multistream heat exchanger
Second cold box interstage cooler
P-54
P-55
P-56
P-57
P-58
P-59
P-60
P-61
P-62
P-63
P-64
P-65
P-66
P-67
P-68
P-69
P-70
P-71
P-72
P-73
P-74
P-75
P-76
P-77
P-78
P-79
P-80
P-81
P-82
P-83
P-84
P-85
P-86
P-87
P-88
P-89
P-90
P-91
P-92
P-93
P-94
P-95
P-96
P-97
P-98
P-99
P-100
P-101
xix
Refrigerant cycles
Ethylene cycle
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
55
56
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
57
58
Propylene condenser
Propylene first level expander
Demethaniser reboiler
Propylene first level KO drum
Propylene compressor
Propylene stream splitter
Propylene second level expander
Propylene second level expander
Propylene second level expander
Depropaniser condenser
Third cracked gas cooler
Second ethylene chiller
Propylene mixer
C2 Splitter reboiler
Propylene stream splitter
Propylene third level expander
Propylene third level expander
Deethaniser condenser
Third ethylene chiller
Propylene mixer
Propylene third level KO drum
Propylene compressor
Propylene stream splitter
Propylene fourth level expander
Propylene fourth level expander
C2 Splitter condenser
Ethylene condenser
Propylene mixer
Propylene fourth level KO drum
Propylene compressor
Propylene fifth level expander
First cold box interstage cooler
Last compressor inlet stream analyser
Propylene compressor
Compressed propylene mixer
Superheated propylene
Cooling water source
Cooling water sink
2/30
3/37
4/45
29/P-46
50/P-50
6/P-53
13/P-56
12/P-62
21/P-63
36/P-67
44/P-75
32/P-76
28/P-82
xx
Propylene cycle
1. Introduction
Process Systems Enterprise is a worldwide company specialized in the simulation, modelling
and optimisation of chemical processes. Its software, gPROMS ModelBuilder, is a powerful solver that
is constantly evolving, allowing a faster and more robust solving of multivariable systems.
This work has been developed under the General Processes department, using models that
already existed in General Process Engineering (GPE) library, as known as PML-SS (Process Model
Library Steady State), except for the reactor model which has been developed in this thesis from a
previous version of the model.
GPE library is a rich library that involves the majority of the most important equipment of the
chemical industry and offers the possibility to size and estimate the cost of several equipment.
1.1
Motivation
The steam cracking is a process with a history of more than 30 years, used everywhere on the
planet. Though having simple components, its the number of equipment and the existence of
refrigeration cycles, due to cryogenic operation, that make this a complex process.
The main motivation of this project is to find a combination of operational variables that are valid
for all flowsheets (main process and refrigeration cycles) and optimize it.
Having that in mind, gPROMS is a great tool to study this process, since several scenarios can
be evaluated and multiple variables can be optimized.
1.2
1.3
Original Contributions
From what has been found in the state of the art section, several studies have been made
regarding the refrigeration cycles of an ethylene plant, which is the main part of the plant with more
potential to be optimized. A lot can still be done to make this a more profitable and energy efficient
process.
1.4
Dissertation Outline
This thesis is organized in the following way: In chapter 2, a background review of the whole
process is presented. Chapter 3 describes the software used for the modelling of all the equipment, the
chosen physical properties package and the model development workow. In chapter 4, the models
functionality is described in detail. In chapter 5, the owsheet implementation is described sequentially.
In chapter 6, the optimisation formulation and cost estimation are explained and a list of all the
assumptions is presented. Chapter 7 shows all the optimisation results for all the attempted formulations.
Finally, chapter 8 presents the main conclusions of this dissertation and suggestions for future work.
2. Background
Natural gas and the petroleum fractions obtained after the primary fractionation of crude oil by
distillation consist chiefly of saturated, paraffinic and naphthenic hydrocarbons, whose chemical
reactivity is mediocre, precluding the development of diversified families of chemical compounds of
varying complexity. This can only be achieved by using unsaturated aliphatic or aromatic hydrocarbons
which, due to their many reactive potentialities, offer outstanding flexibility for organic synthesis. In this
respect, acetylene, which was for many years the most widely used basic hydrocarbon in aliphatic
industry, has gradually been superseded by ethylene, propylene and butadiene, owing to its high
production cost. Despite the fourfold increase in the price of crude oil, which occurred in 1973 and its
subsequent steady increase, ethylene still retained its economic advantage over acetylene from natural
gas or from coal. [1].
At the industrial level, this technique was first developed in the United States. As early as 1920,
Union Carbide and Carbon Co. built a pilot plant operating on ethane and propane, and this company
went on to create the first chemical complex using products derived from the pyrolysis of gas oil. This
type of facility only made significant headway in Western Europe, beginning with the United Kingdom,
and in Japan, after the end of the Second World War. In 1942, British Celanese built the first European
steam cracking unit at Spondon operating on gas oil, with a production capacity of 6 kton/year of
ethylene. In 1946, Shell Chemical built the first petrochemical complex at Stanlow, using refinery gases
as the pyrolysis feedstock. Between the 40s and the 50s, the minimum capacity of ethylene production
plants grew progressively from 10 to 50 kton/year. Giant installations subsequently appeared, routinely
producing 300 kton/year of ethylene from petrochemical naphtha [1].
Steam cracking primarily produces ethylene, but also propylene and, as secondary products,
depending on the feedstock employed, a C4 cut rich in butadiene and a C5+ cut with a high content of
aromatics, particularly benzene [1].
2.1
petroleum hydrocarbons with steam, as known as steam cracking or pyrolysis. First, the hydrocarbon
feedstock is heated, mixed with steam and further heated to incipient cracking temperature (500-650
oC).
The mixture is then fed to a fired tubular reactor, where under controlled residence time,
temperature profile and hydrocarbon partial pressure, is heated until 750-900 oC. These reactions are
highly endothermic, so a high energy input is required. After this step, the stream is severely cooled
down, in order to stop side reactions [3].
In the downstream section, the cracked gas is treated and valuable products are recovered.
Main processing of the cracked gas includes a chilling section comprising one or more transfer line
exchangers and a quenching tower, the removal of heavy compounds, compression, the removal of acid
gases, drying, cryogenic and conventional fractionation and the selective hydrogenation of minor
components, like acetylene, methylacetylene or propadiene. Downstream processing of cracked gas
from a liquid feedstock as naphtha is more complex than from a gaseous feedstock (ethane, propane
or butane) because heavier components are present [3].
With a gaseous feedstock, the downstream processing starts with a chilling section comprising
transfer line exchangers and a water quenching tower, followed by a multistage compression,
comprising 4 to 6 stages. Before the last compression stage, the cracked gas is treated in an acid gas
removal unit, involving a scrubbing process, where caustic soda (NaOH) reacts with hydrogen sulphide
(H2S) and carbon dioxide (CO2 )[3].
After the last compression stage, the cracked gas is chilled and then dried by the use of
molecular sieves. Methane and hydrogen are then removed in the first fractionator, and they can be
used as fuel or purified (in a PSA unit, for instance) and sold, while the bottom stream, with a high
content of C1+ is fed to the deethaniser. In the deethaniser, a C2 stream is produced overhead and a C2+
stream is a bottom product. The overhead stream is hydrogenated to remove acetylene and then
fractionated to recover ethylene on the top and ethane on the bottom, which generally is recycled to the
steam cracking reactor. The bottom product of the deethaniser is fed to the depropaniser, where a C3
stream is produced overhead and a C 3+ stream is a bottom product. The overhead stream is
hydrogenated to remove methylacetylene and propadiene and then fractionated to recover propylene
on the top and propane on the bottom, which generally is recycled to the reactor as well. The
fractionation process can continue after the depropaniser, depending if its economically viable to do so,
but usually a debutaniser is only used when a liquid feedstock is processed.
Downstream processing of the cracked gas from a liquid feedstock is more complex, since there
are heavier compounds to be cracked, recovered and sold. After the liquids are cracked, primary
fractionation removes the heaviest components for use as fuel oil. Gasoline and lighter compounds are
cooled either indirectly or by direct contact with water, which condenses some gasoline. Uncondensed
material is sent to the compressor, where additional gasoline is collected. The gasoline is collected and
stripped to produce raw pyrolysis gasoline. The remaining steps are functionally identical to those in gas
processing [3].
2.1.1
The term cracking refers to the conversion of high molecule hydrocarbons to low molecule
hydrocarbons, for example, the conversion of ethane to ethylene, yielding hydrogen as well.
Steam cracking (or pyrolysis) refers to the process where a hydrocarbon feedstock is cracked
in the presence of steam at a very high temperature. Dehydrogenation of hydrocarbons also occurs in
a cracker reactor.
With a complete dehydrogenation, coke is slowly but continuously produced on the internal
surfaces of the pyrolysis tubes. The heavier the feedstock, the more the coke formed. The coke
deposition may lead to a considerable increase in the pressure drop between the inlet and outlet of the
tube, since the actual usable cross section of the tubes is decreased. This phenomena influence the
yield in olefins, since the total pressure and hydrocarbon partial pressure are increased [7].
Steam acts like a coking inhibitor inside the reactor, since it reacts with coke, although in a small
rate in comparison with cracking reactions [4]. The two primary feedstocks for ethylene production are
naphtha and natural gas components like ethane, propane or a mixture of both [8].
2.1.1.1
Ethylene is the lightest olefin of the chemical world. It is a double bonded two carbon chain, with
two atoms of hydrogen attached to each one.
Its the largest-volume petrochemical produced worldwide. However, this compound has no
direct use, serving instead as an intermediate in a wide range of industrial processes like the following
[3]:
Oxidation to oxirane/acetaldehyde;
Hydration to ethanol;
End products made with ethylene include food packaging, film, toys, food containers, bottles,
pipes, antifreeze, carpets, insulation, housewares, etc. [8]. Table 1 summarizes some of its physical
properties [3].
Table 1: Ethylene's physical properties [3].
28.05
-169.15
-103.71
1.15
52.32
Propylene, as ethylene, is one of the most produced olefins in the world. It is a triple carbon
chain, with a double bond between two carbon atoms, consisting of one CH 2, one CH and one CH3
groups.
Propylene was the first petrochemical raw material to be employed on an industrial scale and
was used for more than 60 years ago in the production of isopropanol. For a long time, propylene was
to some extent overshadowed by its olefin homologue, ethylene. However, its consumption suffered an
expansion due to the success of ethylene: Being a by-product of ethylene production, many important
areas of application were opened up to propylene by the chemical industry. In Table 2, some major
physical properties are presented [3].
Table 2: Propylene's physical properties [3].
42.08
-185.25
-47.70
Standard density at 25 C
(kg/m3)
1.74
20.43
Although propylene is one of the most important feedstocks for the organic chemicals industry,
it is produced almost entirely as a by-product because it is obtained in sufficient amounts in ethylene
production by steam cracking and in some refinery processes.
Like ethylene, it isnt an end use product, but its involved in the production of several
compounds, like polypropylene, acrylonitrile, propylene oxide, 2-propanol, cumene or oligomers [3].
2.1.1.2
Thermodynamic Aspects
The desired hydrocarbons only appear to be stable in relation to the saturated structures from
which they are derived at relatively elevated temperatures. This fact is illustrated by Figure 7, which
o
shows the variation of the free enthalpy of formation Gform
as a function of temperature, related to a
carbon atom, of a number of characteristic hydrocarbon compounds. In this plot, and at a given
temperature, a substance is unstable in relation to all the compounds or elements (C + H 2), whose
representative point remains below its own, since formation from these compounds requires an input
energy: the substance is stable in the opposite case. Accordingly, hydrocarbons are unstable at all
temperatures in relation to their elements, except for methane, which is stable at low and medium
temperatures [1].
Acetylene only becomes stable in relation to the simplest paraffins at temperatures substantially
above 1000 oC. The situation is more favourable for unsaturated hydrocarbons with lower energy
content, such as ethylene, which is stable in relation to ethane at above 750 oC, and benzene, which is
favoured in relation to normal hexane above 400 oC [1].
Given the extreme simplicity of the chemical structure of a saturated hydrocarbon, thermal
activation can only cause the scission of a C-C or C-H bond. The random scission of a C-C bond of the
carbon chain the cracking reaction produces a paraffin and an olefin.
Cm+n H2(m+n)+2 Cm H2m + Cn H2n+2
(Eq. 1)
The scission of a C-H bond gives rise to the formation of an olefin by dehydrogenation with the
same number of carbons atoms as the initial saturated hydrocarbon, as well as hydrogen.
Cp H2p+2 Cp H2p + H2
(Eq. 2)
These conversions are highly endothermic and take place with an increase in the number of
molecules, which are therefore favoured, in terms of thermodynamics, at high temperature and low
pressure.
The comparison of the energies of the C-C and the C-H bonds (345 kJ/mol and 413 kJ/mol,
respectively) confirms that the primary act of pyrolysis of saturated hydrocarbons resides in the scission
of a C-C bond, because this process requires less energy than that required to split a C-H bond [1].
2.1.1.3
Kinetic Aspects
The basic reaction governing the cracking of heavy fractions consists in the cracking of a
saturated aliphatic hydrocarbon into a paraffin and an olefin.
For example, for a C7H16 feedstock, the primary cracking reaction yields propane (paraffin) and
1-butene (olefin). By secondary cracking reactions, the entities thus formed give rise, at various points
of their hydrocarbon chain, to a number of light products, rich in olefins, whose composition and yield
depend on the operation conditions selected [1].
Reactions achieving the more thorough dehydrogenation of olefins directly produced by
cracking provide highly unsaturated compounds, such as acetylene derivatives, which are undesirable
impurities in the use of C2 and C3 olefinic streams or diolefin derivatives, which display pronounced
chemical reactivity. In fact, the latter react in the reverse direction to cracking and give rise to heavy
products by the Diels and Alder reaction (cycloaddition) [1].
The compounds thus formed, if subjected to subsequent intense dehydrogenation are capable
of producing a number of aromatic hydrocarbons and particularly benzene. These constitute the natural
precursors of condensed polyaromatic substances which are designated by the general names of tar or
coke. This product can in no way be compared with graphite, because although it is rich in carbon, its
hydrogen content is still substantial and variable, depending on the feedstock and the operating
conditions [1].
Whereas the cracking reaction rate becomes significant above 700 oC, dehydrogenation
reactions only take place substantially above 800 oC to 900 oC. Moreover, the processes of the formation
of polyaromatic hydrocarbons and coke only occur rapidly at temperatures above 900 oC to 1000 oC.
The adoption of long residence times or the elevation of the reaction temperatures hence favour the
reaction yielding heavy aromatic derivatives at the expense of the production of light olefins by cracking
[1].
As for the polymerization of unsaturated aliphatic compounds, due to their high intrinsic
reactivity, their polymerization is extremely rapid, even at low temperatures. However, since these
reactions represent the reverse of cracking, they are not favoured from the thermodynamic standpoint
in the operating conditions of pyrolysis [1].
As a rule, with respect to the actual steps in cracking, the reactivity of the hydrocarbons
increases with the number of carbon atoms, in each family. For a given number of carbon atoms,
paraffins exhibit higher reactivity than alkylnapthenes but lower than olefins [1].
The fact that pyrolysis reactions proceed by a free radical and a chain mechanism was pointed
out by F.O. Rice. Initiation takes place by the homonymic scission of a C-C bond with the production of
free radicals. They give rise to the reaction chain by extracting a hydrogen atom from the hydrocarbon
and by forming a new free radical [1].
Considering the simple case of ethane, ethyl radicals are obtained by attacking ethane by
methyl radicals. Ethyl radicals are stabilized by supplying ethylene and a hydrogen radical, which in turn
attacks an ethane molecule to form an ethyl radical, and the reaction continues in this manner. Through
this mechanism, a single methyl radical can initiate the conversion of large quantities of ethane to
ethylene and hydrogen. In fact, the radicals disappear in a number of so-called termination reactions,
which culminate in the stoppage of the chain. Whenever a new chain is initiated, a methane molecule is
formed, hence the pyrolysis of ethane produces hydrogen, methane and ethylene as primary products
[1].
Initiation
2.1.1.4
(Eq. 3)
Propagation
(Eq. 4)
(Eq. 5)
(Eq. 6)
(Eq. 7)
Termination
H + H H2
H + CH3 CH4
H + C2 H5 C2 H6
CH3 + C2 H5 C3 H8 CH3 CH = CH2
(Eq. 8)
(Eq. 9)
(Eq. 10)
(Eq. 11)
Other Considerations
The thermal cracking reactions take place in non-catalytic tubular coils built into the radiant
section of the fired heaters. The temperature requirement in a cracking reaction is inversely proportional
to the carbon chain length in the feedstock used [9]. The cracking reactions are highly endothermic, so
a lot of energy is needed and is provided by side-wall or floor burners or a combination of both, which
use gaseous and/or liquid fuels [9].
Residence times are longer for heavy than for light feedstocks. Thus, in the steam cracking of
ethane, propane and, to a lesser degree, butane, the differences in product yields for residence times
ranging from 0.2 to 1.2s are slight.
10
For liquid feedstocks, on the other hand, residence times range from 0.2 to 0.3s [1]. Long
residence times will favour secondary reactions.
From the thermodynamic standpoint, pyrolysis reactions producing light olefins (by cracking or
dehydrogenation) are more advanced at low pressure, a range in which the condensation reactions are
highly disadvantaged. This is why, owing to the pressure drops inherent in the circulation of the reaction
mixture, furnace tubes operate at exit pressures close to atmospheric pressure [1].
Moreover, the condensation side-reaction rate is much more heavily influenced by the
hydrocarbon content of the reaction mixture that the rate of the primary reactions, which are substantially
of the first order with respect to the reactants. A decrease in the partial pressure of the hydrocarbons,
by dilution with steam, reduces the overall reaction rate, but also helps to enhance the selectivity of
pyrolysis substantially in favour of the desired light olefins [1].
The use of steam also involves a number of drawbacks which impose a limit value to its content
in the feedstock. Since steam must be heated to the reaction temperature, its presence increases the
required reactor volume and hence the furnace investment. Its separation from the reactor effluent
requires very large condensation areas and results in high utility consumption [1].
The amount of steam employed, which is normally expressed as the weight of steam per weight
of feedstock, depends on the molecular weight of the feedstock, as shown in Table 4.
Table 4: Typical dilution ratios for different feedstocks [10].
2.1.2
Feed
C2/C3
0.2-0.4
Naphtha
0.3-0.6
0.5-1.0
Cracking Furnace
A cracking heater consists of two main sections: the convection section and the radiant section.
The convection section is the upper offset arrangement of a traditional cracking heater and the radiant
section is the lower end. The two names describe the type of heat transfer transferred to the process
gas.
The convection section functions as a heat recovery device by accumulating heat from the
preheated hydrocarbon feeds that produce flue gas. The preheating takes place by the heated BFW
entering the steam drum and then superheating the saturated steam in the steam drum [9].
In the radiant section, the gas is cracked in radiant coils with the help of heat, which is absorbed
from burned fuel gas and then transferred to it by radiation.
Combustion reactions take place between fuel and oxygen and it is conducted by at multiple
burners, being the primary heat source that generates the heat for the cracking reactions. These
reactions produce flue gas which is passed through the convection section and exits the cracking heater
through a stack on top of the convection section [9]. In each furnace, multiple coils of tubes are used.
11
2.1.3
Chilling Section
The cracked gas is quickly cooled by passing it through one or more transfer line exchangers
and then through one or more quenching systems. The transfer line exchangers recover heat by
generating high and medium pressure steam.
If the feedstock is a light stream such as natural gas, the gases may be pumped directly to a
water quenching tower. Heavier feedstock pyrolysis products are sent to an oil quenching tower [10].
Quenching is employed to provide the rapid cooling necessary to retard the secondary
reactions, which decrease olefin yield and foul the equipment. Heavier feedstocks like naphtha are
usually cooled by direct contact with quench oil. The gases enter the bottom of a vertical quench tower
where they are contacted with a countercurrent flow of oil as they flow to the top of the tower. In addition
to cooling the gases, the quench oil removes entrained hydrocarbon liquids from the pyrolysis products.
The cracked gas then go to a fractionator for removal of fuel oil. The overhead gases from the
fractionator go to the water quench tower, while the bottoms are stripped of light oils, which are
recirculated to the quench tower. For naphtha feedstocks, the design outlet temperature of the oil
12
quenching tower is 360-520 oC, whereas for a gas oil feedstock, the outlet temperature is 500-650 oC
[10].
When a gaseous feedstock is processed, the cracked gas is directly quenched with water after
the heat recovery in the transfer line exchangers. Water quenching cools the gas very rapidly by direct
contact with water in a vertical quenching tower. Condensable hydrocarbons are also removed in this
step. The flow is countercurrent with the gases entering on the bottom of the quenching tower and
cooling water entering on the top [10].
2.1.4
Compression
After the cooling of the effluent of the quenching tower or the primary fractionator, the stream
must be compressed before being fractionated. Compression usually is performed in four to six stages,
with intercooling. However, the number of stages depends mainly on the cracked gas composition and
the highest temperature allowed for interstage discharge. Interstage cooling and temperature control
keep the cracked gas below 100 oC to prevent olefin polymerization and equipment fouling. Maximum
cooling removes as much water as possible to reduce the load on the dryers, but that requires some
attention, since the effluent temperature must remain above the hydrate formation point [3].
The compressor discharge pressure is selected according with the chosen refrigerant so that
methane will condense in the demethaniser condenser. Condensation temperature varies with the
H2/CH4 ratio. The discharge pressure, with the coldest ethylene refrigerant level typically at ca.
101oC is normally 32-38 bar.
13
2.1.5
In the production of ethylene via pyrolysis, a major ongoing challenge to producers is to reduce
coke deposition on the inside of the radiant section tubes and transfer line exchangers. The historical
method for coke reduction and controlling carbon monoxide formation is the use of an inhibitor such as
dimethylsulfide (DMS) and dimethyldisulfide (DMDS). Injection of DMDS results in the passivation of
metal tubes in gas (e.g., ethane) and liquid (e.g., naphtha) steam crackers that produce ethylene, via
adsorption of sulphur atoms produced by its pyrolysis. Hydrogen sulphide formed from DMDS inhibits
the formation of carbon monoxide [15].
Carbon dioxide and hydrogen sulphide are removed from the cracked gas by a oncethrough and regenerative solvent scrubbing with caustic soda.
H2 S + 2 NaOH Na 2 S + 2 H2 O
(Eq. 12)
(Eq. 13)
This step usually occurs before the last compression stage. Carbon dioxide must be removed
because it can freeze at low temperatures in heat-exchange and fractionation equipment. It can also be
absorbed into ethylene, compromising the quality of the product and further processing. On the other
hand, hydrogen sulphide is corrosive. The acid gases liberated from the regenerative solvent can be
incinerated or recovered [3].
2.1.6
Drying
The cracked gas is saturated with water before being compressed and after each intercooler
stage. Moisture must be removed before the fractionation zone, in order to prevent the formation of
hydrates and ice. This is usually accomplished by chilling and adsorption on molecular sieves [3].
Water is adsorbed on molecular sieves after the highest compression pressure is reached to
minimize adsorption cost, because higher pressure allows smaller dryer volume with lower adsorbent
cost and less water removal since water is removed after each compression stage [3].
Molecular sieves are a microporous material which can selectively adsorb gases and liquids.
They are synthetic zeolites (crystalline metal aluminosilicates) which upon dehydration can efficiently
and selectively remove water or other solvents. The selectivity of a molecular sieve is determined by
14
its pore size; molecules with a critical diameter which is less than the pore size will be efficiently
adsorbed while larger molecules will be excluded.
For the drying of the cracked gas, the most used in the industry is the 3A Zeolite Molecular
Sieve, where 3A stands for its pore diameter: 3 angstroms. This zeolite adsorbs water and excludes
all the molecules with a higher critical diameter. Metal cations contained in the crystalline structure of
molecular sieve adsorbents balance the negative charge of the framework. These metal cations create
an electrical field resulting in a strong attraction for polar molecules. Higher polarity generally results in
a higher attraction for a molecular sieve. Since the pore size of the 3A sieve is very close to the critical
diameter of a water molecule and because of the potassium cation incorporated into the 3A molecular
sieve crystal, the 3A sieve has a very strong affinity for the polar water molecule.
. Due to the zeolites high affinity with water, smaller molecules like hydrogen arent adsorbed
[17].
The process typically involves passing the contaminated stream through one or more beds at
ambient temperature to remove (adsorb) the contaminants, while simultaneously, heating a previously
contaminant loaded bed with hot gas at a high temperature (> 200 oC) to desorb the contaminant. The
hot bed is then cooled and is ready for another adsorption step [17].
2.1.7
Cold Box
Before the cracked gas enters the fractionation train, it is chilled to temperatures that can go up
to -150 oC, depending on the front end column and composition. The cold box comprises four
multistream heat exchangers and three intercoolers to remove additional heat. The most used front end
column is the demethaniser, with four feeds, all liquids at different temperatures, coming from
intermediate KO drums. In the multistream heat exchangers, the cracked gas sends heat to the
demethaniser distillate and to the vapour outlet of the last KO drum (the one with the lowest temperature
level) [19].
2.1.8
Fractionation
The cryogenic separation is the predominant method for cracked gas separation. In the last
years, new processes have arisen like adsorption, absorption or membrane technology, but they havent
found major applications in the ethylene industry, yet [8].
The hydrocarbon separation sequence usually begins in the demethaniser with the removal of
methane (CH4) and light components, mostly hydrogen, from high molecule compounds. Hydrogen can
15
be recovered in a PSA unit. The demethaniser normally operates at a pressure near 32 bar, where
methane is condensed at the top around -100 oC [1].
The bottom product is routed to a deethaniser column, which operates at about 26 bar. In the
top stream, ethylene (C2H4) and ethane (C2H6) are recovered and are fed to the C2 Splitter, which
operates at about 19 bar [1], where ethylene is a top product and ethane is a bottom product. The bottom
stream of the deethaniser enters the depropaniser, where in the top stream propylene (C3H6) and
propane (C3H8) are recovered and fed to the C3 Splitter, where propylene is a top product and propane
a bottom product. Depending on the feedstock and the operational parameters, there might be some
butane and butylene that are recovered in a debutaniser.
If the formation of acetylene (C2H2) and methylacetylene/propadiene (C3H4) is considered,
then, there must be a hydrogenation reactor after the deethaniser and after the depropaniser.
Hydrogenation is typically conducted in packed bed or, less frequently, in trickle bed reactors. For
example, the hydrogenation of acetylene may occur in one or more reactors, sometimes featuring
several beds with intermediate cooling, with a palladium or nickel based catalyst. The temperature rises
from 40 to 80 oC between the inlet and outlet of a bed and the operating pressure is near 26 bar [1].
2.1.9
Refrigeration
16
The C3 Splitter separates propylene from propane, which is sold as such or combined with
ethane and recycled for cracking. This separation requires 100-300 stages and a reflux ratio of 12-20
because of the close boiling points of propylene and propane. A polymer-grade fractionator operates
typically at ca. 15 bar, with cooling water in the overhead condenser. The fractionation design depends
on the concentration of propylene in the feed.
Refrigeration in ethylene plants is important and costly. Refrigeration optimisation is vital in plant
design. Typically, two different closed-loop refrigeration systems are employed, like ethylene and
propylene, each generating two to five different temperatures. Separate closed loops protect the purity
of products and refrigerants. Refrigerant temperatures are chosen to accommodate diverse plant needs
most efficiently. Ethylene and propylene are used because of their physical properties and availability
in the ethylene plant. The less expensive refrigerant, propylene, is used from +25 to -40 C and the more
expensive, ethylene, from ca. -100 to -75C. Refrigerants are usually generated by compression to 1620 bar in multistage units, followed by expansion to generate the lowest refrigeration level. Superheated
ethylene is cooled and condensed with propylene refrigerant. Propylene, after compression, is cooled
with process cooling water. With open-loop systems, which are used less frequently, products are
circulated in the refrigeration loops rather than isolated [3].
2.2
long-range feedstock availability and price, as well as in-place refining and chemical processes,
influence process design and feedstock selection because of their high economic impact on
manufacturing cost. Other drivers for new developments are the utilisation of associated gas and natural
gas as petrochemical feedstock, as the value of these fractions offers advantages with respect to costs
compared to crude oil based feedstock. In addition, as flaring of associated gas is no option anymore,
other utilization is required for those streams [3].
17
Dehydrogenation
Dehydrogenation of ethane over Cr or Pt catalysts is limited by equilibrium and leads to a very
poor yield in ethylene. This route is not competitive with conventional routes. [3]
MTO Process
Methanol can be catalytically dehydrated and partially converted to ethylene over alumina and
zeolite catalysts. The process based on novel zeolite catalysts (ZSM-5) was developed by Mobil (MTO,
Methanol to Olefins). A new development introduced by UOP/Norsk Hydro, converts methane to
methanol in a first stage and then converts methanol to olefins. Economics of this new process seem to
be competitive with conventional processes. The process is based on a fluidized-bed reactor for
conversion of methanol. 80% of the carbon content of methanol is converted into ethylene and propene.
Methanol obtained from syngas can be converted in high selectivity (but at a low rate) to ethanol with
the help of catalysts and promoters. The ratio of propylene to ethylene can be adjusted over a wide
range and propylene/ethylene ratios as required by the market (0.8 to 0.9) can be produced via this
route. Another development by Lurgi is the MTP (Methanol to propylene) process, similar to MTO,
however, with a maximized propylene output [3].
18
Syngas
FischerTropsch and modified FischerTropsch processes are used to produce olefins directly
as a byproduct of gasoline and diesel fuel production; liquefied petroleum gases and paraffin
intermediates, which are cracked to produce olefins; and intermediate oxygenated liquids, which are
dehydrated and separated to produce olefins. Syngas is produced by gasification of coal and other. [3]
19
Metathesis
Metathesis of propylene leads to ethylene and 1-butene:
2 C3 H6 C2 H4 + C4 H8
(Eq. 14)
Since this reaction is reversible, the ethylene/propylene production of a plant can be adjusted
to market conditions. So far this technology has been applied in a plant in Canada and at the U. S. Gulf
Coast. As the demand for propylene is increasing faster than the ethylene demand and due to the lack
of propylene production in new ethane crackers in the Middle East, metathesis has gained importance
for the production of propylene based on ethylene and 2-butene feedstocks. Several commercial units
applying metathesis for propylene production via this route are in operation. Contents of isobutene and
butadiene in the feed create problems, as coking rates of the applied catalysts are increased. [3]
20
gPROMS ModelBuilder
The software used for the simulation, modelling and optimisation of all flowsheets was
gPROMS, a process simulator software that is used to create or simulate models, optimize flowsheets,
solving a system of equations meanwhile.
A model created in gPROMS includes a set of equations and both variables and parameters.
The value of the parameters is defined on the SET section of the model. In contrast, variables can be
either calculated in an equation or its value can be defined on the ASSIGN section. Each variable is
related to a variable type and the bounds and default value of the variable will be the same as for the
variable type [23].
Also in the model entity, there is the TOPOLOGY section, where the connections between
objects are defined. These connections can be either written by code or by dragging and dropping
objects and connecting them in the topology window. Each connection has a certain connection type
[23].
gPROMS also features the possibility to have submodels inside a top level model. For
example, for the present case, the distillation column model instead of having all the code written in one
single model, has several submodels, like E equations, H equations, M equations or S equations
(MESH). On the top model, several units are defined, making the connection between the top level
model and all the submodels.
Another fundamental part of a gPROMS project is the PROCESS entity, where all the assigned
variables are stored. To simulate a flowsheet, the process is run, while gPROMS calculates the degrees
of freedom and solves the multivariable system.
gPROMS ModelBuilder features as well the possibility to use initialization procedures, instead
of assigning initial conditions on the models. Although it is really handy, it may be really time consuming,
especially in the column models, which are really sensitive to initial guesses.
Since there are several flowsheets in this project, all of them had to be initialized, before
connecting them. When a flowsheet converges using initialization procedures, a variable set is saved
with all the values for all the variables in the equations that describe the model and it may be simulated
from that time on, using the variable set, as long as the changes arent too severe.
Another feature of gPROMS ModelBuilder is the optimisation toolkit, where the objective
function can be minimized (e.g., total cost) or maximized (e.g., profit), through varying a series of control
variables (assigned variables) and defining some constraints (not assigned variables).
21
3.2
(Eq. 15)
where fiv is the fugacity of component i in the vapour phase, and fil is the fugacity of component i in the
liquid phase. The models used in Multiflash to represent the fugacities in terms of temperature, pressure
and composition fall into two groups: equation of state (EoS) methods and activity coefficient methods.
When using an equation of state method, all thermodynamic properties for any fluid phase are
derived from the equation of state. When using an activity coefficient method, the vapour phase
properties are derived from an equation of state, whereas the liquid properties are determined from a
combination of models which include a representation of the excess properties.
The transport properties of a phase (viscosity, thermal conductivity and surface tension) are
generally derived from semi-empirical models.
Equations of state can be used over wide ranges of temperature and pressure, including the
subcritical and supercritical regions. They are frequently used for ideal or slightly non-ideal systems
such as those related to the oil and gas industry where modelling of hydrocarbon systems, perhaps
containing light gases such as H2S, CO2 and N2, is the norm.
The simple cubic equations of state, Peng Robinson (PR) and Redlich Kwong Soave (RKS),
are widely used in engineering calculations. They require limited pure component data and are robust
and efficient. Both PR and RKS are used in gas-processing, refinery and petrochemical applications.
There is some evidence that RKS gives better fugacities and PR better volumes (densities) [24].
3.2.1
In this project, the RKS Equation of State was chosen, since it is better for fugacities calculation
[25]. The total pressure and each variable needed for its calculation, with the standard (Van der Waals
1-fluid) mixing rules, are given by equations 16 to 19.
P=
NRT
Vb
a
V(Vb)
components
N = i
ni
a = ij a i a j (1 k ij ) ni nj
components
b = i
bi ni
(Eq. 16)
(Eq. 17)
(Eq. 18)
(Eq. 19)
where ni is the number of moles of the component i and k ij is the binary interaction parameter between
component i and component j.
22
For each component, Multiflash calculates a i and bi , from its critical temperature (Tci ), its critical
pressure (Pci ) and its acentric factor (i ), according to equations 20 to 22.
2
Tci
a i = a ci [1 + k i (1
)]
3.2.2
RTci
Pci
(Eq. 20)
(Eq. 21)
(Eq. 22)
Binary interaction parameters (BIPs) are adjustable factors that are used to alter the predictions
from a model until these reproduce as closely as possible the experimental data. For the components
of the system, Multiflash lacked some binary interaction parameters, setting its default value as 0,
considering no interaction at all [24]. The best way to get realistic BIPs is to get vapour-liquid equilibrium
data for a binary or ternary system and compare it to Multiflashs predictions.
Several datasets of vapour-liquid equilibrium were collected for the following pairs: Acetyleneethane [26]; Propane-ethane [27]; Propylene-ethane [28]; Acetylene-ethylene [26], [29]; Ethyleneethane [30], [31], [32], [33], [29]; Ethylene-propane [34], [35]; Ethylene-propylene [36]; Methanebenzene [37]; Methane-ethylene [38]; Methane-propane [27], [39]; Methane-propylene [27]; Propylenebut-1-ene [40]; Propylene-propane [41], [42]; Hydrogen-ethane [43]; Hydrogen-ethylene [43]; Methaneethane [34], [44], [45]; Hydrogen-methane [46]; Hydrogen-benzene [47]; Ethylene-but-1-ene [36], [48],
[49],; Benzene-water [50], [51], [52]; But-1-ene-1,3-Butadiene [27], [53]; 1,3-Butadiene-benzene [54],
[55].
Then, the vapour-liquid equilibrium data for each pair was validated against Multiflashs
predictions. For isothermal P-x data, the pressure at the bubble point of the mixture was calculated,
whereas for isothermal P-y data, the pressure at the dew point was calculated. The residue between
both experimental and predicted pressure profiles was minimized, by varying the BIP for the pair of
components. Figure 19 shows the P-x experimental equilibrium data for a 218.15 K isotherm for the
binary pair ethylene-ethane, compared against Multiflashs predictions with the default BIP and the
optimized BIP.
23
950
P (kPa)
850
750
Experimental
BIP = 0.013
650
BIP = 0
550
450
0
0,2
0,4
xethylene
0,6
0,8
Figure 19: 218.15K isotherm P-x for the ethylene-ethane binary pair.
There is a big deviation between both pressure profiles when the BIP is set as 0. For the
optimized BIP, Multiflash predicts a pressure profile closer to the one for the experimental equilibrium
data.
Also, the binary interaction parameters can have a constant, linear and quadratic dependence
on the temperature, according to the Eq. 23.
k ij = k 0 + k T T + k T2 T 2
(Eq. 23)
According to the datasets available, when a serious temperature dependence was found, the
BIP variation with the temperature was fitted to a linear or second degree polynomial equation and the
different k constants were introduced on Multiflash GUI. Otherwise, only the constant k (k 0 ) was
introduced.
3.3
Figure 20 shows the sequence of actions that was followed to cover all the parts of this project.
Upgrading of
existing cracker
model
Preliminary
process flowsheet
simulation
Binary Interaction
Parameters
estimation
Main process
flowsheet
simulation
Ethylene
refrigerant cycle
flowsheet
simulation
Propylene
refrigerant cycle
flowsheet
simulation
24
Coupling of
process flowsheet
and refrigeration
cycle flowsheets
Optimisation of
coupled flowsheet
4.1
The 1-D model of the reactor offers the possibility to choose the geometry of the reactor, the
pressure drop calculation, whether to include dynamics, choose the kinetics based on the feedstock and
select the heat profile. The top level model includes a heat transfer coefficient calculation submodel, a
kinetic model for cracking and a kinetic model for coking.
Geometry Specifications
Length of tubes
Number of bends
Number of coils
Angle of bends
Radius of bends
Number of junctions
Coke density
25
Dynamics Specifications
Kinetics Specifications
Type of heat profile (constant heat flux, constant wall temperature, uniform heat flux, predicted
or using industrial data)
4.2
Chilling Section
The chilling section includes the transfer line exchangers, that are non-direct contact heat
exchangers and a water quenching tower, which is a direct contact heat exchanger.
4.2.1
The detailed modelling of this equipment hasnt been done, so it was represented as a simple
two-stream heat exchanger. In these models, there are main specifications, pressure drop
specifications, design specifications and costing specifications.
Main Specifications
Thermal specification (hot/cold stream specification, heat duty, hot/cold temperature difference
or none of the above)
Hot/cold stream specification (if chosen, the user may define outlet temperature, vapour fraction
or enthalpy)
Design Specifications
26
Cost Specifications
Material of construction
Tube length
4.2.2
The detailed modelling of this equipment hasnt been done, so it was represented by a mixer
(where both gas and water streams are mixed together) and by a P-T flash drum, so, both streams leave
the quenching tower at the same temperature and pressure.
In Figure 24, one can have an inside view of the water quenching tower model. Both inlets in
the mixer are the same as both inlets on the water quenching tower model. Since the gas inlet pressure
is near the atmospheric pressure and the water enters the tower at the atmospheric pressure, the knockout drum operates at the temperature of the mixture and at atmospheric pressure. Both knock-out drum
outlet streams are the same as the ones in the water quenching tower model.
27
Main Specifications
Water flowrate
4.3
4.3.1
Compression Section
The compression section involves five stages of compression, with intercooling in between, in
order to condense water, and a knock-out drum to remove it from the cracked gas stream. Since the
multistage compressor model doesnt include the option to apply intercooling, each stage was simulated
as one single compressor. The specifications of the intercoolers arent described since they were
described in the previous subchapter.
28
Material of construction
Separator specification (Duty, vapour fraction, product purity, liquid outlet flowrate or
temperature)
4.3.2
Drying Section
The detailed modelling of this equipment hasnt been done, so it was represented by a
component splitter, where a water-free cracked gas is obtained.
Basic Specifications
Outlet specification (split fractions, component flowrates or total flowrates and composition) of
each component
Advanced Specifications
Temperature specification (same temperature as inlet, different temperatures for each outlet,
different increments/drops of temperature for each outlet, adiabatic and equal outlet
temperatures)
Pressure Specifications
Outlet specification (same pressure as inlet, different pressures for each outlet, different
pressure drops for each outlet)
4.4
Cold Box
In the cold box, there is heat integration between the cracked gas, which is cooled, and the top
of the demethaniser column and the last knock-out drum, which are heated. Since the heat transferred
in these multistream heat exchangers isnt sufficient to cool down the cracked gas, there are heat
29
exchangers between the multistream heat exchangers and the knock-out drums, in order to achieve the
desired temperatures. This heat is transferred to ethylene and propylene refrigerants. Both heat
exchangers and knock-out drum models were already described.
Hot/cold streams outlet temperature (the user must assign number of streams 1 temperatures)
Outlet specification (pressure, pressure ratio, power controlled, temperature, vapour fraction,
enthalpy, saturated liquid, saturated vapour, subcooled liquid or superheated vapour)
Mechanical efficiency
30
4.5
Separation Section
The demethaniser, deethaniser, C2 Splitter, depropaniser and the C3 Splitter were all simulated
Main Specifications
Number of stages
Feed stage(s)
Condenser/reboiler Specifications
Condenser specification (reflux ratio, normalized reflux ratio, top liquid/vapour product flowrate,
bottom product flowrate, bottom product purity, top liquid/vapour product purity, condenser duty
or condenser temperature)
Pressure Specifications
Pressure profile (linear, uniform, calculated from column hydraulics or custom profile)
31
Operation Specifications
Operation specifications (Reflux ratio, normalized reflux ratio, boil up ratio, normalized boil up
ratio, top liquid/vapour product flowrate, bottom product flowrate, bottom product purity, top
liquid/vapour product purity, condenser duty, condenser temperature, reboiler duty or reboiler
temperature)
Design Specifications
Plate efficiency
Plate spacing
Costing Specifications
Column material
Tray material
Steel density
4.6
Other Equipment
Apart from the equipment that was already described, mixers, JT valves, stream splitters and a
32
4.6.1
Mixers
Mixers are used throughout the whole flowsheet. Since there isnt the possibility to calculate the
mixture pressure yet, the user must specify the outlet pressure.
Pressure Specifications
Mixture outlet pressure (minimum inlet pressure, stream to propagate pressure or the value of
the pressure itself)
4.6.2
Stream Splitters
As the mixers, there are many stream splitters used throughout the flowsheet.
Main Specifications
4.6.3
JT Valves
Main Specifications
Outlet specification (pressure drop, flow coefficient, pressure, temperature, vapour fraction,
saturated liquid, saturated vapour, subcooled liquid or superheated vapour).
33
4.6.4
Conversion Reactor
Main Specifications
Operation Specifications
Reaction Specifications
34
Name of reactions
Reference component
Flowsheet Implementation
The flowsheet implementation was divided in seven parts: the PML-SS library models validation,
the steam cracker model validation, the main process flowsheet, the ethylene refrigerant cycle
flowsheet, the propylene refrigerant cycle flowsheet, the cascade refrigerant cycle flowsheet and the
coupled final flowsheet.
5.1.1
Before starting implementing the main process flowsheet, a previous project flowsheet [4] was
replicated, as part of the background review, involving the steam cracking of a mixture of ethane,
propane and butane. In this previous project [4], all the models were made by the author.
Since this project was made essentially using PML-SS library models, it served as well to
validate them against the previous project (Figure 33).
In Figure 33, there are six feeds: propane, ethane, butane, steam, recycled propane and
recycled ethane. These feeds are mixed, at a pressure of 4 bar and a temperature of 600 oC (in the
range of conditions from the outlet of the convection zone of a fire heated reactor). This mixture enters
the reactor, where a series of reactions take place, based on the kinetics presented on the model. The
chilling section is reduced to a single cooler, instead of two transfer line exchangers and one water
quenching tower. The outlet pressure is near atmospheric and the outlet temperature is 30 oC. Then,
the cooled gas is dried in a component splitter that removes all the water present. The compression
section is neglected and the feed pressure and temperature of the columns are assigned, instead of
being propagated from compressors, expanders or heat exchangers before or between columns.
35
Along with the simplifications made on the flowsheet, the main disadvantage was the use of
IPPFO (Ideal Physical Properties Foreign Object), which leads to much less realistic values than
Multiflashs predictions.
Figure 34 shows the replication of the previous project, using PML-SS library models. In
opposite to assigning the feed conditions of each equipment, the inlet pressures and temperatures were
propagated from the equipment behind. Since the cracker model was not available at that time, a
component splitter and a feed source were used to simulate the cracking reactor outlet composition of
the previous project.
Figure 34: Replication of the ethylene plant flowsheet from previous project [4].
All the results matched, which validated PML-SS library models used on the flowsheet.
5.1.2
After the validation of PML-SS models, the cracker model had to be updated in order to be
compatible with PML-SS library models. Some equations had to be changed, as well as the variables
propagated through the port connections.
The cracker model was validated against a case from [56] involving the simulation of a stream
cracker, using the kinetics presented by [57].
In the case of the pressure profile, the paper presents the results of a simulation and
correspondent industrial data. A reproduction of the case was run and the pressure profile was
compared to the industrial data found on the paper.
Regarding the temperature profile, the paper only presents the simulation results and made no
reference to industrial data, so the simulation profile was compared against the previous simulation
profile from [56]. Both profiles can be seen Figures 35 and 36.
36
Simulated profile
Pressure (bar)
2,8
2,6
2,4
2,2
2
0
0,2
0,4
0,6
0,8
The simulation profile for the pressure along the coils and the industrial data show a perfect
match which validates the pressure calculation for the reactor model.
850
Temperature (oC)
800
750
Previous simulation [56]
700
Simulation profile
650
600
0,0
0,2
0,4
0,6
0,8
1,0
Figure 36: Simulation gas temperature profile and previous simulation profile [56].
On the other hand, there are slight differences between the simulation profile for the temperature
and the simulation profile from a previous simulation [56]. As was stated before, there are a lot of
possibilities of specifying the heat flux profile. One can assign a constant heat flux and the temperature
is calculated assuming a constant heat profile, one can assign a constant wall temperature and the
temperature and heat flux are calculated assuming a constant wall temperature, one can assign an
uniform/predicted heat flux and a flue gas temperature and the temperature is calculated by radiation
laws, assigning an emissivity factor as well or one can use industrial data for the heat flux and flue gas
37
temperature profiles, so its not that strange to get slight differences between the two plots, since there
are several possibilities for the heat flux profile. In the present case, a constant heat flux was considered.
Although there are differences, given the slight degree of them, the simulation profile is considered good.
Finally, for a conversion of propane of 90% and a steam to propane ratio of 0.3 kg/kg, the reactor
outlet composition after dehydration using the gPROMS model was compared to a reactor outlet
composition after dehydration described on [3] and the deviances are shown in Table 5.
Table 5: Comparison between reactor outlet compositions after dehydration.
Simulation
Ullmans [3]
Deviation
Methane
22.90%
23.43%
-0.53%
Ethane
2.87%
3.06%
-0.19%
Ethylene
36.28%
37.15%
-0.87%
Acetylene
0.27%
0.46%
-0.19%
Propane
9.98%
9.97%
0.01%
Propylene
14.45%
14.81%
-0.36%
But-1-ene
1.15%
1.00%
0.15%
1,3-Butadiene
3.00%
2.85%
0.15%
Benzene
7.37%1
2.15%
5.22%
Hydrogen
1.75%
1.51%
0.24%
Carbon monoxide
N/A
0.04%
Carbon dioxide
N/A
0.01%
Hydrogen sulphide
N/A
0.01%
Propadiene
N/A
0.52%
C4H4
N/A
0.08%
C4H10
N/A
0.04%
Toluene
N/A
0.43%
Xylenes
N/A
0.05%
Ethylbenzene
N/A
0.01%
Styrene
N/A
0.21%
Pyrolysis Gasoline
N/A
1.27%
N/A
0.94%
The deviations are really good except for benzene. The reason for that deviation is that in the
reactor kinetics model, none of the compounds listed as N/A were considered. In contrast, there is one
reaction in the kinetics that yields a pseudo-component named C6. That pseudo-component groups
all the hydrocarbons with five or more carbon atoms (as in benzene, toluene, xylenes, ethylbenzene,
styrene, pyrolysis gasoline and pyrolysis fuel oil). As it is stated in [56], it is common to attribute
benzenes physical properties to the C 6 component, so, basically, the mass fraction of benzene is not
only the mass fraction of benzene but also the mass fraction of all five or more chained carbons
molecules. Basically, benzene composition on the simulation outlet composition is, in reality, C 6
composition. In contrast, neither carbon monoxide (from coke residue), carbon dioxide (from coke
This value is in reality C6 composition, grouping not only benzene but all the other molecules with 5 or more
carbon atoms.
1
38
residue), hydrogen sulphide (coke inhibitor), propadiene, C4H4 and C4H10 were considered in the kinetics
presented in [56], [57], [58], [59], [60], [61] and [62], so they were neglected in this simulation as well.
5.1.3
First, an ethylene capacity of 850 kton/year was chosen from [63], keeping in mind that usually
ethylene capacities vary between 500 kton/year and 1500 kton/year. Based on the capacity value and
assuming that the plant operates 24 hours a day, 330 days a year, a flowrate of propane was chosen to
feed the reactor. Since only the radiation zone is considered in the reactor model, the temperature (650
oC)
and pressure (3 bar) of the feed are typical outlet conditions of the convection zone of the reactor
[3], [64].
For the reactor, the geometry and its properties were chosen by assigning typical values
presented on [3], [56] and [64]. The coils have a length of 70 m, an internal diameter of 0.108 m and a
wall thickness of 0.008 m.
The kinetics were obtained from [56], [57], [58], [59], [60], [61] and [62], which present the
reactions that occur for an ethane feed, a propane feed or a mixed feed and the activation energies and
Arrhenius pre-exponential factor for each reaction and for each feedstock. For a mixed feed, the
following reactions were considered:
C3 H8 C2 H4 + CH4
(Eq. 24)
C3 H8 C3 H6 + H2
(Eq. 25)
C3 H8 + C2 H4 C2 H6 + C3 H6
(Eq. 26)
2 C3 H6 3 C2 H4
(Eq. 27)
2 C3 H6 C6 H6 + 2 CH4
(Eq. 28)
C3 H6 C2 H2 + CH4
(Eq. 29)
C3 H6 + C2 H6 1 C4 H8 + CH4
(Eq. 30)
C2 H6 C2 H4 + H2
(Eq. 31)
C2 H4 + C2 H2 C4 H6
(Eq. 32)
C2 H6 + C2 H4 C3 H6 + CH4
(Eq. 33)
A constant heat profile of 90 kW/m2 [3] was assumed across the reactor, since a more detailed
heat profile would require industrial data.
The chilling section comprises two transfer line exchangers that chill the cracked gas from
around 900 oC to 250 oC [3]. In the first one, the outlet temperature of the cracked gas is 450 oC and
high pressure steam is produced, while in the second, the outlet temperature is 250 oC and medium
pressure steam is produced. The chilling section also comprises a water quenching tower that cools
down the cracked gas to a temperature near ambient (about 40 oC).
After the chilling section, the cracked gas is compressed in a five stage compressor with
intercooling in between. The inlet pressure of the first stage is 1 bar, the outlet pressure of the last one
39
is 32 bar and a constant pressure ratio [65] of 2 was chosen (1 2 4 8 16 32 bar). There is a
physical constraint in this section, since that temperature cant go up past 100 oC when pressurizing the
system, since olefins can polymerize [3]. There are intercoolers between the compression stages that
condense part of the water content and it is removed in interstage knock-out drums. The cracked gas is
cooled in the intercoolers, with cooling water that enters at 25 oC and leaves the cooler at 40 oC, in order
to be possible to cool it back again in a cooling tower.
Since the outlet of the last compression stage is at about 90 oC, which is the maximum allowable
temperature to dry the gas [15], the cracked gas is first cooled to an intermediate temperature and then
is dried. After it is dried, the water-free cracked gas is cooled again to 30oC with water and then to 15
oC
is a sequence of four multistream heat exchangers where the cracked gas is cooled while the distillate
of the demethaniser (as known as tail gas) and the vapour outlet of the last knock-out drum (as known
as hydrogen rich gas) are heated. There are four temperature levels, as four feeds on the demethaniser:
-121 oC, -96 oC, -71 oC and -43 oC [19]. Since the cracked gas cant transfer all its heat to the cold
streams (due to crossover reasons), there are intercoolers after the first, the second and the third multistream heat exchangers. The heat is transferred to propylene refrigerant in the first intercooler, while in
the second and in the third, the heat is transferred to ethylene refrigerant [19].
After the cold box, the cracked gas is submitted to the fractionation train in order to separate
the several components. The number of stages and feed stages were followed by [2] and [68] and the
reflux and boil up ratios were varied in order to achieve the desired purities in both top and bottom of
each distillation column. The pressure of the columns was chosen in order to use the refrigeration levels
presented by [2], [69] and [70].
The demethaniser, which has 65 stages and 4 feed locations (15 th, 20th, 25th and 33rd stages),
operates at 32 bar and its top product contains a high content of methane and traces of hydrogen,
whereas the bottom is a mixture of C1+. The demethaniser bottom is then expanded to 26 bar and enters
the deethaniser, which has 60 stages and 1 feed location (27th stage). The top comprises a high content
of C2, whereas the bottom comprises a mixture of C2+. The distillate of the deethaniser is then fed to the
acetylene hydrogenation reactor, which operates at 26 bar and 67 oC, where acetylene is converted to
ethane and ethylene, by reaction with added hydrogen. It was assumed a 37% yield in ethylene and a
63% yield in ethane [3].
C2 H2 + H2 C2 H4
(Eq. 34)
C2 H2 + 2 H2 C2 H6
(Eq. 35)
The effluent of the acetylene hydrogenation reactor is then expanded to 19 bar and fed to the
C2 Splitter, which has 120 stages and 1 feed location (90th stage), recovering high purity ethylene on
the top and high purity ethane on the bottom.
On the other hand, the bottom stream of the deethaniser is expanded to 8 bar and fed to the
depropaniser, which has 60 stages and 1 feed location (25th stage). The top is rich in C3, whereas the
bottom comprises a high content of C3+. Finally, the top of the depropaniser is compressed to 15 bar
and fed to the C3 Splitter, which has 230 stages and 1 feed location (120th stage). The top product is
40
high purity propylene, whereas the bottom comprises a high content of propane. Both ethane and
propane are recycled to the reactor.
Figure 37 shows the main process flowsheet of the ethylene plant.
41
42
5.1.4
Two refrigerant cycles were designed, based on [2], [69] and [70]. The ethylene refrigeration
cycle (Figure 38) includes two temperature levels, namely -73 oC (4.6 bar) and -101 oC (1.18 bar).
Although two compressors are represented in the flowsheet, they are, in reality, two stages: one that
compresses the vapour from 1.18 bar to 4.6 bar and the other which compresses the vapour from 4.6
bar to 20.2 bar. This pressure is the minimum pressure that allows the condensation of ethylene by
propylene refrigerant.
Superheated ethylene, at 20.2 bar is chilled first with water to 33 oC and then with two different
levels of propylene refrigerant to 8 oC and - 17oC, respectively. After that, ethylene refrigerant is
completely condensed by another level of propylene refrigerant, leaving the last heat exchanger at
30.6 oC. Ethylene is then expanded to 4.6 bar and fed to the only user on this level, the second cold box
interstage cooler, leaving it with an increased vapour fraction. The partially vaporized stream is fed to a
knock-out drum, where the vapour is sent to a compressor stage and the liquid is split in two streams,
since there are two users on this temperature level. Both streams are expanded to 1.18 bar. Part of the
stream is fed to the demethaniser condenser and the rest is fed to the third cold box interstage cooler.
Both streams are then mixed and fed to a compressor stage. Both compressed streams are then mixed
and the loop is closed. The stream analyser (15) reports the inlet conditions of the stream that enters
the compressor 16.
43
Table 6 summarizes the hot and cold fluid on each heat exchanger.
Table 6: Different levels of ethylene refrigeration.
Heat Exchanger
Hot fluid
Superheated ethylene
33
oC
ethylene
Cold fluid
Cooling water
5
oC
level propylene
8 oC ethylene
-17 oC ethylene
Cracked gas
Cracked gas
Demethaniser distillate
5.1.5
The propylene refrigeration cycle (Figure 39) includes five temperature levels, namely 12.8 oC
(8.5 bar), 5 oC (6.8 bar), - 20 oC (3.1 bar), -35 oC (1.7 bar) and -46 oC (1.08 bar).
Although three compressors are represented in the flowsheet, they are, in reality, three stages:
one that compresses the vapour from 1.08 bar to 1.7 bar, a second which compresses the vapour from
1.7 bar to 3.1 bar and a third one that compresses the vapour from 3.1 bar to 16.5 bar. This final pressure
is suitable to condense propylene with cooling water.
Superheated propylene, at 16.5 bar is partially condensed with cooling water, to a temperature
of 39.1 oC. This stream is then expanded to 8.5 bar and fed and condensed in the only user on this level,
which is the demethaniser reboiler. Propylene vapours are then separated in the first knock-out drum
and sent to a compressor stage. On the other hand, the liquid is split in three sub-streams, which are
expanded to 6.8 bar. The first sub-stream is fed to the first user on this temperature level, which is the
depropaniser condenser. The second sub-stream is fed to the second user on this temperature level,
which is the third cracked gas cooler. The last sub-stream is partially vaporized in the third user on this
temperature level, the second ethylene chiller. All the three partially vaporized sub-streams are then
mixed and totally condensed in the last user on this temperature level, which is the C2 Splitter reboiler.
The liquid stream is then split again in two sub-streams, which are expanded to 3.1 bar. The first substream is fed to the first user on this temperature level, the deethaniser condenser. The second substream is fed to the second user on this temperature level, the third ethylene chiller. Then, both substreams are mixed together and fed to a second knock-out drum. The vapour is sent to a compressor
stage, while the liquid is split in two sub-streams, which are expanded to 1.7 bar. The first sub-stream
is fed to the first user on this temperature level, the C2 Splitter condenser, while the second sub-stream
is fed to the second user on this temperature level, the ethylene condenser. Both sub-streams are mixed
and fed to the last knock-out drum, where the vapour is separated and sent to a compressor stage,
while the liquid is expanded to 1.08 bar and fed to the only user on this level, the first cold box interstage
cooler. Then, the vaporized stream is sent to a compressor stage. Both compressed streams are then
mixed and the loop is closed. The stream analyser (51) reports the inlet conditions of the stream that
enters the compressor 52.
44
Table 7 summarizes the hot and cold fluid on each heat exchanger.
Heat Exchanger
Hot fluid
Cold fluid
Superheated propylene
Cooling water
Demethaniser bottom
Depropaniser distillate
5 oC level propylene
5 oC level propylene
C2 Splitter bottom
30 oC cracked gas
5 oC level propylene
Deethaniser distillate
C2 Splitter distillate
Cracked gas
33 oC level ethylene
5 oC level propylene
8 oC level ethylene
45
46
5.1.6
The main challenge of the refrigeration cycles was to find a combination between both ethylene
and propylene refrigerants flowrates and the split fractions on the splitters that would respect three
constraints:
The heat transferred by ethylene refrigerant in the second and third ethylene chillers and in the
ethylene condenser must be the same as the heat received by propylene refrigerant on each
heat exchanger;
The heat transferred in the propylene condenser and in the first ethylene chiller is the same as
the heat received by cooling water on each heat exchanger;
The outlet cold stream temperature cant go higher than the inlet hot stream temperature and
the inlet cold stream temperature cant go past the outlet hot stream.
Minimum temperature difference at the inlet or outlet of the heat exchangers of 2 oC while using
refrigerants [71] or 5 oC while using water [71];
Fixed global heat coefficients (from [71], since the detailed sizing of heat exchangers is not
implemented in gPROMS yet), regardless of the flowrates or condensation/vaporization
phenomena.
Both refrigerant cycles were connected by the second ethylene chiller (2 in the ethylene
refrigerant cycle, 30 in the propylene refrigerant cycle), the third ethylene chiller (3 in the ethylene
refrigerant cycle, 37 in the propylene refrigerant cycle) and the ethylene condenser (4 in the ethylene
refrigerant cycle and 45 in the propylene refrigerant cycle). The correspondent single stream heat
exchangers were taken out from both flowsheets and replaced by two-stream heat exchangers.
Table 8 summarizes all the connections between the ethylene refrigerant cycle flowsheet and the
propylene refrigerant cycle flowsheet.
Table 8: Connection between ethylene refrigerant cycle and propylene refrigerant cycle.
refrigerant cycle
refrigerant cycle
Hot fluid
Cold fluid
33oC
5 oC level
ethylene
propylene
oC
-20 oC level
ethylene
propylene
-17 oC
-35 oC level
ethylene
propylene
47
48
5.1.7
Once the main process flowsheet (Figure 37) and the cascade refrigerant cycle flowsheet
(Figure 40) were designed, the coupled final flowsheet was obtained.
To obtain the coupled final flowsheet, the same constraints and assumptions were taken in
account when developing the cascade refrigerant cycle flowsheet.
Both heat exchangers 1 and 19 of Figure 40 dont connect with the main process flowsheet and
they were just replaced by two-stream heat exchangers with water as cold fluid.
All the other single stream heat exchangers on Figure 40 (cascade refrigerant cycle) correspond
to other single stream heat exchangers on Figure 37 (main process). Before coupling the main process
flowsheet and the cascade refrigerant cycle flowsheet, the single stream heat exchangers on the
cascade refrigerant cycle were replaced by two-stream heat exchangers and corresponding process
streams were simulated by sources, in order to verify if everything was alright (Figure 41). Then, the
sources were disconnected from the two-stream heat exchangers of the cascade refrigerant cycle
flowsheet, and the single stream heat exchangers on the main process flowsheet were removed and
both flowsheets were finally connected (Figure 42). Table 9 summarizes all the connections between
the main flowsheet and the cascade refrigerant cycle flowsheet. The final flowsheet (Figure 42) can be
seen in A3 size in the Appendices chapter (A-1).
49
Heat Exchanger
refrigerant cycle
in main process
interstage cooler
interstage cooler
(6)
(P-53)
Demethaniser condenser
(12)
Demethaniser
condenser
(P-62)
Hot fluid
Cracked gas
-73 oC level
ethylene
Demethaniser
-101 oC level
distillate
ethylene
Cold fluid
Cracked gas
(P-56)
-101 oC level
ethylene
Demethaniser reboiler
Demethaniser reboiler
12.8 oC level
Demethaniser
(21)
(P-63)
propylene
bottom
Depropaniser condenser
Depropaniser
Depropaniser
5 oC level
(28)
condenser (P-82)
distillate
propylene
(29)
cooler (P-46)
C2 Splitter reboiler
C2 Splitter reboiler
5 oC level
(32)
(P-76)
propylene
Deethaniser condenser
Deethaniser condenser
Deethaniser
-20 oC level
(36)
(P-67)
distillate
propylene
C2 Splitter condenser
C2 Splitter condenser
C2 Splitter
-35 oC level
(44)
(P-75)
distillate
propylene
30 oC cracked gas
Cracked gas
5 oC level
propylene
C2 Splitter bottom
-46 oC level
propylene
50
Figure 41: Cascade refrigerant cycle flowsheet with two-stream heat exchangers.
51
52
5.2
Simulation Results
In this subchapter, important simulation results are presented for the main equipment of the
flowsheet.
5.2.1
As indicated previously, the convection zone isnt considered in this flowsheet, since the reactor
model doesnt account for the flue gas temperature profile across the convection zone of the reactor.
Assuming that both propane and ethane recycle streams are expanded and heated to the feed
conditions, the reactor feed comprises ethane and propane. The outlet results are presented in
Table 10.
Table 10: Reactor feed and reactor outlet.
Feed
Outlet
Pressure (bar)
103
3
103
1.5
Temperature (oC)
650
892
16.7
Ethane
2.6
2.5
Ethylene
29.0
Acetylene
0.3
Propane
74.9
8.3
Propylene
11.7
But-1-ene
0.3
1,3-Butadiene
2.1
Benzene
5.2
Hydrogen
1.5
Water
22.5
22.5
Regarding the conversion, an overall conversion of 84.8% was achieved. On the other hand,
the ratio between steam and hydrocarbons is 0.29 kg/kg. Figures 43 and 44 show the gas temperature
and the pressure profile, respectively.
53
900
850
800
750
700
650
600
0
0,2
0,4
0,6
0,8
0,8
Pressure (bar)
2,5
1,5
1
0
0,2
0,4
0,6
Both pressure and temperature variations over the relative axial length are in close agreement
with the ones presented by [56]. Note that the changes in the slope of the temperature profile are due
to the fact that there are 10 reactions, with different kinetics. Figures 45 and 46 show the variation of the
mass flowrate of each component with the relative axial length of the reactor. Figure 47 shows the
reaction rate profile for the reactions involving ethane. As one can see, the mass flowrate of water
remains the same for all the relative axial length of the reactor, which makes sense since water neither
reacts nor is produced in any reaction. All the products have their mass flowrate increased with the
relative axial length, apart from the mass flowrate of propylene that starts decreasing at about z = 0.8.
Thats due to the fact that the reaction that involves propane yielding propylene and hydrogen is
dislocated to the reverse side and propylene starts being consumed.
54
0,35
0,3
Methane
0,25
Ethylene
Acetylene
0,2
Propylene
0,15
But-1-ene
1,3-Butadiene
0,1
Benzene
0,05
Hydrogen
0
0
0,2
0,4
0,6
0,8
Water
Propane follows an expected profile, being consumed throughout all the relative axial length,
whereas ethanes mass flowrate decreases until z=0.4, then starts increasing until z = 0.85, and then
decreases again. This phenomena can easily be explained with the reaction rates profiles involving
ethane.
0,029
0,8
0,0285
Propane
Ethane
0,7
0,028
0,6
0,0275
0,5
0,027
0,4
0,0265
0,3
0,026
0,2
0,0255
0,1
0,025
0,0245
0
0,2
0,4
0,6
0,8
0,9
Figure 47 shows the reaction rate profiles along the relative axial length of the reactor. r3 is
an irreversible reaction that yields ethane (C3 H8 + C2 H4 C2 H6 + C3 H6 ); r7 is an irreversible reaction
where ethane reacts (C3 H6 + C2 H6 1 C4 H8 + CH4 ); r8 is reversible reaction where ethane is a
direct reactant (C2 H6 C2 H4 + H2 ) and r11 is an irreversible reaction where ethane reacts as well
(C2 H6 + C2 H4 CH4 + C3 H6 ). In the beginning, ethane majorly is reacting, mainly in reaction r8, hence
the decrease in the mass flowrate in the beginning. On the other hand, from z = 0.4, reaction r3 starts
being dominant, along with a reverse in r8 reaction that now is yielding ethane, hence the increase in
the mass flowrate. At the end, at z = 0.8, reaction r8 falls abruptly, hence the abrupt decrease in the
mass flowrate of ethane.
55
4
2
0
-2
r3
r7
r8
r11
-4
-6
-8
-10
0
0,2
0,4
0,6
Relative axial length
0,8
5.2.2
Compressor Train
5.2.3
Demethaniser
Top
Bottom
Pressure (bar)
10.5
32
60.7
32
Temperature (oC)
-97
10.5
56
Methane
99.1
Ethane
4.1
Ethylene
0.1
48.7
Acetylene
0.5
Propane
14.1
Propylene
19.8
But-1-ene
0.4
1,3-Butadiene
3.5
Benzene
8.9
Hydrogen
0.8
Water
5.2.4
Deethaniser
Top
Bottom
Pressure (bar)
32.3
26
28.4
26
Temperature (oC)
-17.6
76.2
5.2.5
Methane
Ethane
7.7
0.1
Ethylene
91.4
Acetylene
0.9
Propane
30.1
Propylene
42.4
But-1-ene
0.9
1,3-Butadiene
7.5
Benzene
19.0
Hydrogen
Water
C2 Splitter
In the C2 Splitter condenser, 25.5 MW of heat is exchanged, whereas in the C2 Splitter reboiler,
19.9 MW is exchanged. Apart from that, Table 13 summarizes both top and bottom results.
Table 13: C2 Splitter top and bottom results.
Top
Bottom
Pressure (bar)
29.7
19
2.7
19
Temperature (oC)
-31.1
-9.7
Ethane
0.1
99.8
Ethylene
99.9
0.2
Acetylene
Propane
57
5.2.6
Propylene
But-1-ene
1,3-Butadiene
Benzene
Hydrogen
Water
Depropaniser
Top
Bottom
Pressure (bar)
20.6
8
7.8
8
Temperature (oC)
13.5
96
5.2.7
Methane
Ethane
0.1
Ethylene
Acetylene
Propane
41.5
Propylene
58.4
But-1-ene
3.4
1,3-Butadiene
27.5
Benzene
69.1
Hydrogen
Water
C3 Splitter
In the C3 Splitter condenser, 53.4 MW of heat is exchanged, whereas in the C3 Splitter reboiler,
50.5 MW is exchanged. Apart from that, Table 15 summarizes both top and bottom results.
Table 15: C3 Splitter top and bottom results.
Top
Bottom
Pressure (bar)
12.0
15
8.6
15
Temperature (oC)
35.2
43.4
58
5.2.8
Methane
Ethane
0.1
Ethylene
Acetylene
Propane
99.9
Propylene
99.9
0.1
But-1-ene
1,3-Butadiene
Benzene
Hydrogen
Water
In the ethylene refrigeration cycle, 30 kg/s of ethylene are compressed at the expense of 5.3
MW of energy.
5.2.9
In the propylene refrigeration cycle, 400 kg/s of propylene are compressed at the expense of
80.9 MW of energy.
59
60
min = 0 z dt
u
(Eq. 36)
where is the objective, u is the vector of parameterized control signals and t is the time horizon
and time intervals for parameterisation of the control variables. If there are constraints in the process,
either equality or inequality constraints, they can be defined in gPROMS ModelBuilder. In a
mathematical approach, those constrains are given by Equations 37 and 38.
f(x , x, y, u, p) = 0
(Eq. 37)
g(x , x, y, u, p) 0
(Eq. 38)
where f represents the equality constraints and g represents the inequality constraints. x is the
vector of the state variables (assigned degrees of freedom), is the derivative of x, y is the vector of the
algebraic variables (time-invariant variables) and p is the parameters vector.
By default, gPROMS treats optimisation problems as dynamic ones, optimising the behaviour
of a system over a finite non-negative time horizon. However, in some cases, it is desired to optimise a
system at a single time pointperforming a so-called "point" optimisation. From the mathematical point
of view, this is equivalent to solving a purely algebraic problem in which a generally nonlinear objective
function is maximised or minimised subject to generally nonlinear constraints by manipulating a set of
optimisation decision variables that may be either continuous or discrete.
The specification of a point optimisation problem in gPROMS is achieved simply by omitting the
horizon part of the corresponding Optimisation Entity.
In this project, the objective of the optimisation was to minimize the total annualized cost of the
coupled final flowsheet. That said, the objective function was defined as the total annualized cost.
For the first and second optimisation problems, all control variables can be a real number, so
the SRQPD solver was used. The SRQPD (Successive Reduced Quadratic Programming) solver
employs a sequential quadratic programming method for the solution of the nonlinear programming
problem.
61
For the third optimisation problem, the number of stages and the feed stages must assume
integer values, so its a MINLP problem and the OAERAP (Outer Approximation/Equality
Relaxation/Augmented Penalty) solver was used.
The OAERAP algorithm decomposes the MINLP into a NLP sub-problem and a MILP master
problem. First, the algorithm solves the NLP relaxation of the integer variables to obtain the rst
intermediate iteration to the next problem. After that, the MILP master problem nds an integer point
that features an augmented penalty function to nd the minimum over the convex linearized function.
Then, the algorithm solves a NLP, xing the integer variables, to nd the optimum value of the
continuous variables. Finally, it calculates the gradient based on the linearized functions and determines
if the optimal point was reached or rather it needs to do another iteration and calculates the respective
point. Note that, due to the linearization of the non-convex functions, there is no guarantee of nding the
global optimum.
6.1
OPEX), which was minimized after the optimisation cases. For the calculation of CAPEX, the costs of
the compressors, distillation columns, expanders and heaters were considered.
CAPEX ($) = capital cost of compressors + distillation columns + expanders + heaters
(Eq. 39)
For the calculation of OPEX, the cost of the energy spent on the compressors was considered.
OPEX ($/h) = energy spent in compressors energy cost ($/kWh)
(Eq. 40)
CAPEX
10
(Eq. 41)
Regarding OPEX, it was assumed that the plant operates 24 hours a day for 330 days a year.
$
(Eq. 42)
Lastly, the annualized total cost is the sum of the annual CAPEX and the annual OPEX.
Annualized total cost = Annualized CAPEX + Annualized OPEX
(Eq. 43)
The costs that are calculated in gPROMS have 2006 as its base year. All the costs were updated
to 2013, by applying the correspondent cost index of 569.5 [72] (2006s is 500).
62
(Eq. 44)
where Fd is a constant dependant on the drive type (electric motor, gas turbine or steam turbine) and
Fm is a constant dependant on the construction material [73]. The capital cost of the heat exchangers is
given by Eq. 45.
Heater capital cost = Fp Fm Fl exp(11.667 0.8709 ln(A) + 0.09005 ln(A)2 )
(Eq. 45)
(Eq. 46)
(Eq. 47)
(Eq. 48)
Staging cost = (number of stages 1) FNT FTT FTM 468 exp(0.1739 diameterft )
(Eq. 49)
Distillation column capital cost = Tower cost + Platforms cost + Staging cost
(Eq. 50)
where Fm is a constant dependant on the construction material, FNT is a constant dependant on the
number of trays, FTT is a constant dependant on the type of trays and FTM is a constant dependant on
the tray material [73].
The energy cost was assumed to be 0.08$/kWh (as an average price for the electric energy in
the USA).
63
A total annualized cost of 87.67 M$ per year was obtained and, as one can see in Figure 48,
the major part of it is OPEX, as it was already expectable. The estimated CAPEX is 6.80 M$/year while
the estimated OPEX is 80.87 M$/year.
8%
CAPEX
OPEX
92%
Figure 48: Total annualized cost distribution for the base case.
Regarding the total annualized CAPEX distribution, one can see in Figure 49 that the
compressors are the main expense, representing 60% of the total annualized CAPEX, as in
4.08 M$/year, while the distillation columns represent 2.56 M$/year and the heaters represent
0.16 M$/year. The cost of the expanders is negligible.
2%
38%
Columns
Expanders
Compressors
Heaters
60%
0%
Figure 49: Total annualized CAPEX distribution for the base case.
64
6.2
the above mentioned optimisation objective, but due to time constraints, only three cases were
considered. The optimisation cases were run sequentially. First, case 1 was run and the assigned
variables were updated to the optimal ones. Then, case 2 was run with the flowsheet running with the
new assigns from case 1. Finally, case 3 was run with the flowsheet running with the new assigns from
case 1 and 2.
Case 1: Mass flowrate of ethylene and propylene refrigerants and ethylene splitter 9;
Case 2: Pressure ratios and intercooling temperatures of the compressor train;
Case 3: Number of stages and feed location of the columns.
In gPROMS, the heat exchangers can operate in design or in performance mode. In design
mode, the user specifies a fixed global heat coefficient and the duty or an outlet temperature and the
model calculates the area. In contrast, in performance mode, the user specifies the heat exchanger area
and a fixed global heat coefficient and the model calculates the outlet temperatures of the cold and hot
fluids and the duty.
All the optimisation cases were run in design mode, which means that the area of the heat
exchangers varies when a duty or an outlet temperature is specified on the heat exchangers. In a
consulting project, its normal to work in performance mode, when one has industrial data for the heat
exchangers area. Other option would be to use the area that resulted from the base case, but that would
be wrong, since the refrigerant flowrates were assumed.
Even so, due to the distillation column initialisation model lack of robustness (at the numerical
standpoint), it would be quite difficult to work in performance mode. If one varied only variables related
to the refrigerant cycles and kept all the results for the columns and heat exchangers of the main process
the same, thered be only one solution for the problem. Since the duties and the areas would be fixed,
as well as the global heat coefficient (that is assigned and considered constant), thered be just one
value for the LMTD that would solve the problem. And if there was just one possible value for the LMTD,
the outlet temperature for the refrigerant would be fixed. And, if the outlet temperature of the refrigerant
was fixed, since the duty that the refrigerant receives is the duty calculated in the main process
flowsheet, the mass flowrate of refrigerant would have to be the same as well, so thered be only one
solution, the same as the one achieved when working in design mode to get the areas to run the
simulation in performance mode.
Then again, one could try to make a plant wide optimisation, varying both main process and
refrigerant cycle variables in performance mode, using as constraints the purities on the columns, for
instance, and there should be a different solution, but that would be really time consuming, due to the
distillation column initialisation model lack of robustness.
So, these optimisation cases were considered pre-implementation cases, before installing the
plant, the only situation where it makes sense to run optimisation cases in design mode.
65
For case 1, it was assumed that all the duties in the main process flowsheet remain the same;
hence all the duties of the heat exchangers that connect both main process and the cascade
refrigeration cycle remain the same as well. On the other hand, all the heat exchangers with outlet
temperatures specified instead of duty kept the same assigned variable.
For case 2, it was considered that regardless the pressure ratios of the stages on the
compressor train and intercooling temperatures, the outlet pressure of the last compressor stage must
be 32 bar. It was considered as well that the dryer will absorb all the remaining water, regardless if it is
less or more, depending on the water removed in the intercooling process (the split fraction of water in
the dryer remained assigned as 1).
For case 3, the normalized reflux and boil up ratios used as assigned variables were fixed and
the number of stages and feed location were varied, using as constraints the purities on the top and
bottom of the columns.
66
7. Optimisation Results
7.1
Case 1
In case 1, the following variables were used as control variables:
Ethylene mass flowrate (initial guess: 30 kg/s, lower bound: 20 kg/s, upper bound: 50
kg/s);
Propylene mass flowrate (initial guess: 400 kg/s, lower bound: 150 kg/s, upper bound:
500 kg/s);
Ethylene splitter split fraction (initial guess: 0.6, lower bound: 0.4, upper bound: 0.8).
Ethylene last compressor inlet vapour fraction (evaluated by the stream analyser (15)
behind it, must be 1);
Propylene last compressor inlet vapour fraction (evaluated by the stream analyser (51)
behind it, must be 1).
Figure 50 shows the variation of both flowrates for the base case for some solution iterations
that the optimizer returned as output (for the cases where all the constraints were respected).
40
400
400
Propylene mass flowrate
35
360
320
30
30
27,9627
27,883
27,8042
27,611
25
220,862
20
Base case
220,667
220,456
217,977
Iteration 2
Iteration 3
Iteration 4
280
240
200
Iteration 1
Figure 50: Ethylene and propylene flowrates as function of several optimisation iterations.
67
The total annualized cost corresponding to the same iterations can be seen in Figure 51.
90
87,67
85
80
75
70
65
60
55
51,40
50
Base case
Iteration 1
51,40
51,40
Iteration 2
Iteration 3
51,38
Iteration 4
Figure 51: Total cost as function of several optimisation iterations (case 1).
The first iteration that respects all the constraints is really close to the optimal solution, which
means that both flowrates were really over-specified, especially the propylene mass flowrate (the slopes
of both curves are quite similar). Table 16 summarizes the main results of this optimisation case.
Table 16: Main results for case 1.
Relative deviance
Variable
Base case
30
27.611
8.0
400
217.977
45.5
0.6
0.614
2.3
6.80
5.46
19.7
80.87
45.92
43.2
87.67
51.38
41.4
The total annualized CAPEX has decreased from 6.80 M$/year to 5.46 M$/year, registering a
decrease by 20%. On the other hand, the total annualized OPEX changed significantly from
80.87 M$/year to 45.92 M$/year, decreasing by 43%. Since the OPEX had a much more severe
decrease than the CAPEX, when evaluating the total annualized cost distribution after case 1, one can
see that the CAPEX percentage increased by 3%, while the OPEX percentage decreased by 3%.
68
11%
CAPEX
OPEX
89%
Since the compressor cost is calculated based on the horse power and since that the horse
power decreased by 43% (being proportional to the OPEX), the compressor share on the total
annualized CAPEX distribution decreased from 60% to 50% (or from 4.08 M$/year to 2.74 M$/year),
while the columns share increased from 38% to 47% and the heaters share from 2% to 3%. Note that
the total cost of the distillation columns almost remained the same (the only changes were in the reboiler
and condenser costs, because of the change in the areas), while the heaters cost went from
0.16 M$/year to 0.17 M$/year.
3%
Columns
47%
Expanders
Compressors
50%
Heaters
0%
69
7.2
Case 2
In case 2, the following variables were used as control variables:
First, second and third intercoolers outlet temperature (initial guess: 40 oC, lower bound:
30.35 oC, upper bound: 46.85 oC)
Fourth intercoolers outlet temperature (initial guess: 43.85 oC, lower bound: 36.85 oC,
upper bound: 46.85 oC)
Figure 54 shows the variation of all the compressor stages pressure ratios for some solution
iterations that the optimizer returned as output (for the iterations where all the constraints were
respected).
2,1
Pressure Ratio
2,05
2
1,95
1,9
1,85
First stage
Second stage
Third stage
Fourth stage
Fifth stage
1,8
Figure 54: Variation of the compressors pressure ratios for several optimisation iterations.
70
Figure 55 shows the variation of all the intercoolers outlet temperatures for some solution
iterations that the optimizer returned as output (for the iterations where all the constraints were
respected).
44
42
40
First intercoole
38
Second intercooler
36
Third intercooler
34
Fourth intercooler
32
30
The total annualized cost corresponding to the same iterations can be seen in Figure 56.
51,6
51,2
50,8
50,4
50,0
Figure 56: Total cost as function of several optimisation iterations (case 2).
In case 2, there isnt such a big difference on the total annualized cost when compared to case
1, which means that the solution for the base case (equal pressure ratio of 2 for all stages) was near the
optimal value.
On the other hand, the outlet temperature for the first, second and third intercoolers became
stuck in the lower bound, 30 oC. One could try to decrease even more the lower bound, but it wouldnt
respect the minimum temperature difference when using cooling water, which was considered as 5 oC.
71
After case 1
After case 2
Relative deviance
solution
solution
1.85935
7.0
2.04609
2.3
2.06445
3.2
2.0101
0.5
2.02695
1.4
40
30.35
24.1
40
30.35
24.1
40
30.35
24.1
43.85
41.1
6.3
5.46
5.45
0.2
45.92
45.23
1.5
51.38
50.69
1.4
Variable
The OPEX registered a decrease by 1.5% from 45.92 M$/year to 45.23 M$/year, whereas the
CAPEX almost remained the same going from 5.46 M$/year to 5.45 M$/year. Thats because, although
the compressor capital cost went from 2.74 M$/year to 2.71 M$/year, the heaters capital cost went from
0.17 M$/year to 0.19 M$/year, due to the increase in the intercoolers area (the gas outlet temperature
is closer to the cooling water inlet temperature, now). The total annualized cost distribution and the total
annualized CAPEX cost distribution are exactly the same as the ones gotten for the first optimisation
case.
7.3
Case 3
In case 3, the following variables were used as control variables:
Number of stages (initial guess: the ones assigned in the base case, lower bound: 1,
upper bound: the ones assigned in the base case);
Feed stage (initial guess: the ones assigned in the base case, lower bound: 1, upper
bound: number of stages).
Distillate mass fractions (must be the same as the ones gotten for the base case);
Bottom mass fractions (must be the same as the ones gotten for the base case).
72
Due to the complexity of this optimisation case, instead of optimising the columns in the coupled
final flowsheet, each column was simulated and optimised separately. Table 18 shows the results of this
optimisation case.
After case 2
After case 3
Relative deviance
solution
solution
65
55
15.4
15,20,25,33
15,20,25,37
10.8
60
51
15.0
27
27
120
117
2.5
90
85
5.6
60
37
38.3
25
25
230
230
120
120
5.45
5.31
2.6
45.23
45.24
50.69
50.55
0.3
Variable
Both total cost distribution and total annualized CAPEX distribution remained practically the
same.
73
7.4
Optimisation Summary
In this paragraph, the main results obtained after case 3 solution are shown. Table 19
Variable
Base case
30
27.611
400
217.977
0.6
0.614
1.85935
2.04609
2.06445
2.0101
2.02695
40
30.35
40
30.35
40
30.35
43.85
41.1
65
55
15,20,25,33
15,20,25,37
60
51
27
27
120
117
90
85
60
37
25
25
230
230
120
120
40.8
39.7
5.3
4.3
80.9
26.7
6.80
5.31
80.87
45.24
87.67
50.55
10
This variable wasnt used as a control variable but its important to notice its decrease.
This variable wasnt used as a control variable but its important to notice its decrease.
10 This variable wasnt used as a control variable but its important to notice its decrease.
8
9
74
Figure 57 shows that the main source of optimisation was the reduction of both propylene and
ethylene refrigerants, since they were over-specified, corresponding to a reduction on the total
annualized CAPEX+OPEX of 41.39%. On the other hand, after the change on the compression train
pressure ratios and intercoolers outlet temperatures, the total annualized CAPEX+OPEX was reduced
by 42.19%. Finally, case 3 led to a final reduction of 42.34% of the base case total cost.
One should remember that all the optimisation cases were run sequentially, so the reduction
percentages are cumulative.
50%
87,67
41,39%
42,19%
42,34%
80
40%
60
30%
51,38
50,69
40
20
0
Base Case
50,55
0%
20%
10%
100
0%
Case 1
Case 2
Case 3
Case
Figure 57: Total cost and deviance of optimisation cases from the base case.
75
76
Conclusions
From the literature review that was done in this project, one can see that this subject, the steam
cracking of hydrocarbons in order to get olefins, is a well-known process studied in the chemical industry,
since it has been the preferred project to produce olefins, for more than 30 years. Although there is a lot
of information about this process, its flexibility turns it into something more complex. As it was stated
before, there are several possible feedstocks for this process and the operational variables on a plant
arent the same for all of them. Apart from that, the investment in different technologies (for instance, for
the cracking furnaces) turns the steam cracking into a more and more innovative process.
To the extent of my knowledge, there isnt a public available computational model that describes
an entire plant as complete as this one, despite the assumptions that were made. Several articles and
patents were found, regarding single parts of the plant like the cold box, the cracking furnace or the
distillation column train, but none connects all the parts in just one flowsheet.
The most detailed model found [4], although making some assumptions, was the one that
served to validate all the sub-models used in this work, apart from the cracking furnace.
gPROMS can clearly be used to model an industrial ethylene plant, since all the models from
PML-SS were validated with the earlier projects replication and the cracking model was validated as
well with industrial data.
Some assumptions were made on the flowsheet, like the omission of the convection zone of the
reactor, the simplification of the molecular sieve, the simplification of the acetylene hydrogenation
reactor or even the omission of the acid gas removal unit. Regarding the omission of the convection
zone on the reactor, the only inconvenient is that the heat transfer integration is not modelled. For
example, both steam and hydrocarbons are heated with upcoming flue gas, in different parts of the
convection zone. Apart from that, the reaction rates in the convection zone arent null, as considered,
but negligible when compared with the radiation zone.
In relation to the molecular sieve, the simplification of modelling it as a component splitter wasnt
that drastic, since in several articles its stated that water is completely removed from the cracked gas
(water mass fraction on the cracked gas around ppm) and all the other components arent absorbed by
the molecular sieve.
Concerning the acetylene hydrogenation reactor, both yields in ethane and ethylene were
assumed, from [3]. No kinetics were used and the reactor was modelled as a conversion reactor. Its
considered as well a good approximation, since the flowrate of acetylene is really small in comparison
with ethane and ethylenes flowrates.
Lastly, the acid gas removal unit omission was considered reasonable as well, since the mass
balance on the downstream equipment wouldnt be affected.
Regarding the optimisation cases, case 1 was the one that led to the biggest decrease on the
total annualized cost, from 87.67 M$/year to 51.38 M$/year, since the flowrates of both refrigerants were
over-specified, specially propylenes flowrate, that decreased from 400 kg/s to 218 kg/s, leading to a
decrease in the energy spent to compress it from 80.9 to 26.7 MW. So, it makes sense that this
77
optimisation case was the most important one, since the majority of the total annualized cost is the cost
related to energy spent on the compressors.
In relation to the refrigeration cycles, although there are numerous possibilities to design them,
it was used a design similar to one of the designs available on the literature [2]. It makes no sense to
compare this projects refrigerant flowrates with industrial data if the capacity isnt the same, since that
the flowrate of the refrigerants depend directly on the cracked gas flowrate. That said, industrial flowrates
were found on [2], although no capacity was stated on the paper. If the capacity is missing, it makes
more sense to compare the ratio between the refrigerants flowrates than the flowrates themselves.
After the optimisation cases, the final flowrates of ethylene and propylene refrigerants were 27.6
and 218 kg/s, respectively, resulting in a ratio of 7.9 kg propylene/kg ethylene. On the other hand,
according to [2], Quattors plant uses 21.8 kg/s of ethylene and 151.4 kg of propylene, resulting in a ratio
of 6.9 kg propylene/kg ethylene, a value 12% inferior. This deviance is not a surprise, given the possible
differences in the heat requirements between cases. There might exist a difference in purities, perhaps
on the C2 Splitter. The reflux and boil up ratios were defined in the C2 Splitter to get 99.9% purity ethylene
(polymer grade) and maybe in Quattors plant the purity of ethylene is lower.
8.2
Future Work
Although this is a detailed flowsheet, there are some simplifications that should be corrected.
Regarding the reactor, the convection zone should be modelled, in order to represent the different heat
transfers that exist on this part of the reactor. Still, regarding the reactor, industrial data should be used
to estimate kinetic parameters, to have a more rigorous composition of the effluent.
There should exist as well a design model for the transfer line exchangers, since these
exchangers are really unique and special in terms of heat transfer.
The quenching tower should be as well modelled in a more detailed way. For the purposes of
this project, the water quenching tower was modelled as a mixer where both gas and water are mixed
together and a flash drum that separates the vapour from the liquid. This is obviously not correct, since
there are several temperature levels on a quenching tower, and both gas and liquid stream dont leave
the tower at the same temperature. Apart from that, some water is evaporated on the quenching tower
and some gases are condensed.
Also, the acid gas removal should be included, by taking to account the injecting of a sulphurous
compound to prevent coking.
Apart from that, the shortcut model of the columns, which I helped to develop meanwhile, would
be a good start for a first estimate on the number of stages and reflux ratio of the columns. See Appendix
A-2 for more details.
One thing that is missing on gPROMS ModelBuilder PML-SS library is a detailed sizing option
on the heat exchanger model. In the library model, one can specify a global heat coefficient (which is
considered fixed) and the duty or an outlet temperature and gets an area, or one can specify a global
heat coefficient and an area and gets the duty and the outlet temperatures. The actual problem is the
consideration and assumption of a fixed global heat coefficient, which is obviously wrong. I started doing
78
that model and at the moment the model is almost finished for the case of no phase change. The model
still has to account for phase change on one side or both sides. See Appendix A-3 for more details.
79
80
Bibliography
[1]
[2]
[3]
[4]
K. Y. Cheung, Site-Wide and Supply Chain Optimisation for Continuous Chemical Processes, I.
C. London, Ed., London: PhD Thesis, 2008.
[5]
[6]
G. P. S. Energy, Ethene: Production and Properties, [Online]. Available: http://www.greenplanet-solar-energy.com/ethene.html. [Accessed August 2013].
[7]
B. Meisong Yan, Simulation and Optimization of an Ethylene Plant, MSc Thesis, 2000.
[8]
[9]
[10] T. B. Parsons, C. M. Thompson and G. E. Wilkins, Industrial Process Profiles for Environmental
Use: Chapter 5. Basic Petrochemicals Industry, January: EPA, 1977.
[11] R. Karimzadeh, H. R. Godini and M. Ghashghaee, Flowsheeting of steam cracking furnaces,
Chemical Engineering Research and Design, vol. 87, pp. 36-46, 2009.
[12] A. Laval, Olmi transfer line exchanger, [Online]. Available: http://www.alfalaval.com/solutionfinder/products/olmi-transfer-lineexchanger/Pages/default.aspx?NoOfResults=13&ResultPosition1&source=http%3A//www.alfala
val.com/solutionfinder/products/pages/default.aspx%3Ftype%3DProductCategory%26firstItemID%3Ddb89f58.
[Accessed August 2013].
[13] Description of Quench Towers and Quench Separation Drums, [Online]. Available:
http://chem409group23.wikispaces.com/Description+of+Quench+Towers+and+Quench+Separation+Drums.
[Accessed August 2013].
[14] Siemens's compressor train, [Online]. Available:
http://www.energy.siemens.com/nl/pool/hq/compression/expansion%20turbines/compressortrain-large.jpg. [Accessed August 2013].
81
[15] N. C. Systems, PERP Program - New Report Alert, August 2003. [Online]. Available:
http://www.chemsystems.com/reports/search/docs/abstracts/0203-S10-abs.pdf. [Accessed 3
June 2013].
[16] Wet Scrubber, [Online]. Available: http://www.ustudy.in/node/3145. [Accessed August 2013].
[17] A. P. 13F, Use of Molecular Sieves and other Adsorbents for Contaminant Removal in Ethylene
and Propylene Production, April 2009. [Online]. [Accessed June 2013].
[18] Webest Molecular Sieves, [Online]. Available: http://www.molecularsieves.com/Molecular_Sieve_FAQ.htm. [Accessed August 2013].
[19] A. A. Al-Rabiah, K. D. Timmerhaus and R. D. Noble, Membrane Technology for Hydrogen
Separaton in Ethylene Plants.
[20] Linde, Recovery of valuable olefin products from refinery off-gas streams, [Online]. Available:
http://www.lindeindia.com/userfiles/image/File/2012_01_06_Recovery_of_Valuable_Olefin_ChemicalNews.pdf.
[Accessed August 2013].
[21] M. P. Background, Overview of the MTO Process, [Online]. Available:
http://www.cchem.berkeley.edu/molsim/teaching/fall2009/mto/background.html. [Accessed
August 2013].
[22] TPI DCC Plant, [Online]. Available:
http://images.pennwellnet.com/ogj/images/ogj2/9602jfufu03.gif. [Accessed August 2013].
[23] gPROMS ModelBuilder Guide, May 2013.
[24] I. C. S. Ltd, User Guide for Multiflash for Windows 4.1, London, 2012.
[25] I. C. S. Ltd, User Guide for Models and Physical Properties (Version 4.1), London, 2012.
[26] R. J. Hogan, W. T. Nelson, G. H. Hanson and M. R. Cines, Ethane-Ethylene-Acetylene System
Vapor Liquid Equilibrium Data At -35, 0 and 40 F., Industrial and Engineering Chemistry, vol.
47, no. 10, pp. 2210-2215, 1955.
[27] I. Wichterle, J. Linek, Z. Wagner, J. C. Fontaine, K. Sosnkowska-Kehiaian and H. V. Kehiaian,
Vapor-Liquid Equilibrium in Mixtures and Solutions. Binary Liquid Systems of Nonelectrolytes,
Landolt-Brnstein, 2007.
[28] H. Lu, D. M. Newitt and M. Ruhemann, Two-Phase Equilibrium in Binary and Ternary Systems.
IV. The System Ethane-Propylene, Proceedings of the Royal Society of London. Series A,
Mathematical and Physical Sciences, vol. 178, pp. 506-525, 1941.
[29] J. L. McCurdy and D. L. Katz, Phase Equilibria in the System Ethane-Ethylene-Acetylene,
Industrial and Engineering Chemistry, vol. 36, no. 7, pp. 674-680, 1944.
[30] J. C. Calado, E. J. S. Gomes de Azevedo, P. Clancy and K. E. Gubbins, Thermodynamic Study
of Liquid Mixtures of Ethane and Ethene, Journal of the Chemical Society, Faraday Transactions
1, vol. 79, no. 11, pp. 2657-2667, 1983.
82
[31] D. A. Barclay, J. L. Flebbe and D. B. Manley, Journal of Chemical & Engineering Data, vol. 27,
no. 2, pp. 135-142, 1982.
[32] M. S. Rozhnov, V. G. Kozya and V. I. Zhdanov, Khim Prom, pp. 674-675, 1988.
[33] A. M. Clark and F. Din, Equilibria between solid, liquid and gaseous phases at low temperatures
the system carbon dioxide + ethane + ethylene, Discussions of the Faraday Society, vol. 15,
pp. 202-207, 1953.
[34] M. Guter, D. M. Newitt and M. Ruhemann, Two-Phase Equilibrium in Binary and Ternary
Systems. II. The System Methane-Ethylene. III. The System Methane-Ethane-Ethylene,
Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, vol.
176, no. 964, pp. 140-152, 1940.
[35] X. H. Han, Y. J. Zhang, Z. J. Gao, Y. J. Xu, X. J. Zhang and G. M. Chen, Vapor-Liquid Equilibrium
for the Mixture Methane (CH4) + Ethane (C2H6) over the Temperature Range (126.01 to 140.01)
K, Journal of Chemical & Engineering Data, vol. 57, pp. 3242-3246, 2012.
[36] H. K. Bae, K. Nagahama and M. Hirata, Measurement and Correlation of High Pressure VaporLiquid Equilibria for the Systems Ethylene-1-Butene and Ethylene-Propylene, Journal of
Chemical Engineering of Japan, vol. 14, no. 1, pp. 1-6, 1981.
[37] D. Legret, D. Richon and H. Renon, Vapor-Liquid Equilibria of Methane-Benzene, MethaneMethylbenzene (Toluene), Methane-1,3-Dimethylbenzene (m-Xylene), and Methane-1,3,5Trimethylbenzene (Mesitylene) at 313.2 K up to the critical point, Journal of Chemical &
Engineering Data, vol. 27, pp. 165-169, 1982.
[38] R. C. Miller, A. J. Kidnay and M. J. Hiza, Liquid + vapor equilibria in methane + ethene and in
methane + ethane from 150.00 K to 190.00 K, The Journal of Chemical Thermodynamics, vol. 9,
pp. 167-178, 1977.
[39] G. Soave, S. Colussi and E. Stenby, Estimation of VLE phase compositions of binary systems
containing gaseous components, from the feed composition and the cell volume, Fluid Phase
Equilibria, Vols. 187-188, pp. 373-389, 2001.
[40] G. H. Goff, P. S. Farrington and B. H. Sage, Volumetric and Phase Behavior of Propene-1-Butene
System, Industrial and Engineering Chemistry, vol. 42, no. 4, pp. 735-743, 1949.
[41] H. H. Reamer and B. H. Sage, Volumetric and Phase Behavior of Propene-Propane System,
Industrial and Engineering Chemistry, vol. 43, no. 7, pp. 1628-1634, 1950.
[42] T. Hakuta, K. Nagahama and M. Hirata, Binary Vapor-Liquid Equilibrium for C3 Hydrocarbons,
Bulletin of the Japan Petroleum Institute, vol. 11, pp. 10-15, 1969.
[43] R. B. Williams and D. L. Katz, Vapor Liquid Equilibria in Binary Systems Hydrogen with
ethylene, ethane, propylene and propane, Industrial and Engineering Chemistry, vol. 46, no. 12,
pp. 2512-2520, 1954.
[44] Z. Zhang and Z. Duan, Phase Equilibria of the system methane-ethane from temperature scaling
Gibbs Ensemble Monte Carlo simulation, Geochimica et Cosmochimica Acta, vol. 66, pp. 34313439, 2002.
83
84
85
86
Caption
P-1 Fresh feed; P-2 Feed mixer; P-3 Steam cracking
reactor; P-4 Steam source; P-5 First transfer line
exchanger; P-6 Steam sink; P-7 Steam source;
P-8 Second transfer line exchanger; P-9 Steam sink;
P-10 Water quenching tower; P-11 Quench water
source; P-12 Quench water outlet; P-13 First compressor
stage;
P-14
Cooling
water
source;
P-15 First intercooler; P-16 Cooling water sink;
P-17
First KO drum; P-18 First KO liquid outlet:
P-19 Second compressor stage; P-20 Cooling water
source; P-21 Second intercooler; P-22 Cooling water
sink; P-23 Second KO drum; P-24 Second KO liquid
outlet; P-25 Third compressor stage; P-26 Cooling water
source; P-27 Third intercooler; P-28 Cooling water sink;
P-29 Third KO drum; P-30 Third KO liquid outlet; P-31
Fourth compressor stage; P-32 Cooling water source; P33 Fourth intercooler; P-34 Cooling water sink; P-35
Fourth KO drum; P-36 Fourth KO liquid outlet; P-37 Fifth
compressor stage; P-38 Cooling water source; P-39 First
cracked gas cooler; P-40 Cooling water sink; P-41
Molecular
dryer;
P-42
Dried
water;
P-43 Cooling water source; P-44 Second cracked gas
cooler; P-45 Cooling water sink; P-46/29 Third cracked
gas cooler; P-47 Tail gas; P-48 Hydrogen rich gas;
P-49 First multistream heat exchanger; P-50/50 First cold
box interstage cooler; P-51 First cold box KO drum;
P-52
Second
multistream
heat
exchanger;
P-53/6 Second cold box interstage cooler; P-54 Second
cold box KO drum; P-55 Third multistream heat
exchanger; P-56/13 Third cold box interstage cooler;
P-57 Third cold box KO drum; P-58 Fourth multistream
heat exchanger; P-59 Fourth cold box KO drum;
P-60 Hydrogen rich gas expander; P-61 Demethaniser
column;
P-62/12
Demethaniser
condenser;
P-63/21 Demethaniser reboiler; P-64 Tail gas expander;
P-65
Deethaniser
feed
expander;
P-66 Deethaniser column; P-67/36 Deethaniser
condenser; P-68
Deethaniser reboiler;
P-69 Acetylene rector feed heater; P-70 Hydrogen
source; P-71 Acetylene reactor feed mixer;
P-72
Acetylene
hydrogenation
reactor;
P-73 C2 Splitter feed expander; P-74 C2 Splitter column
P-75/44 C2 Splitter condenser; P-76/32 C2 Splitter
reboiler; P-77 Ethylene sink; P-78 Ethane recycle
expander;
P-79
Ethane
recycle
heater;
P-80 Depropaniser feed expander; P-81 Depropaniser
column;
P-82/28
Depropaniser
condenser;
P-83 Depropaniser reboiler; P-84 Heavies sink;
P-85 C3 Splitter feed compressor; P-86 C3 Splitter
column; P-87 C3 Splitter condenser; P-88 C3 Splitter
reboiler; P-89 Propylene sink; P-90 Propane recycle
expander; P-91 Propane recycle heater; P-92 Steam
source; P-93 Steam sink; P-94 Steam source;
P-95 Steam sink; P-96 Steam source; P-97 Steam sink;
P-98 Cooling water source; P-99 Cooling water sink;
P-100 Steam source; P-101 Steam sink
Appendices
A-1. Final Flowsheet (A3 Size)
A-1
Main variables/parameters
Table 20: List of FUG model main variables and parameters.
Variable
zF
zF,LK
zF,HK
xD
xD,LK
xD,HK
xB
xB,LK
xB,HK
F
D
B
,
q
Nmin
N
Rmin
R
Rfactor
(D/F)ratio
Nabove
Nbelow
NF
hF,V
hF
hF,L
gPROMS name
z_F_molar
z_F_molar(LK)
z_F_molar(HK)
x_molar_D
x_molar_D(LK)
x_molar_D(HK)
x_molar_B
x_molar_B(LK)
x_molar_B(HK)
F_molar
D_molar
B_molar
alfa_i_HK
q
teta
N_min
N_op
R_min
R_op
R_factor
D_F_fraction
N_above
N_below
N_F
h_F_V
h_F
h_F_L
Units
mol/mol
mol/mol
mol/mol
mol/mol
mol/mol
mol/mol
mol/mol
mol/mol
mol/mol
kmol/h
kkol/h
kmol/h
mol/mol
mol/mol
mol/mol
kJ/kg
kJ/kg
kJ/kg
Significance
Vector with feed molar composition
Molar fraction of the light key component on the feed
Molar fraction of the heavy key component on the feed
Vector with distillate molar composition
Molar fraction of the light key component on the distillate
Molar fraction of the heavy key component on the distillate
Vector with bottom molar composition
Molar fraction of the light key component on the bottom
Molar fraction of the heavy key component on the bottom
Feed molar flowrate
Distillate molar flowrate
Bottom molar flowrate
Relative volatility between component I and the heavy key
Column feed vapour quality
Auxiliary variable
Minimum number of stages(for total reflux)
Actual number of stages
Minimum reflux ratio (for an infinite number of stages)
Actual reflux ratio
Ratio between actual and minimum reflux ratio
Ratio between distillate and feed molar flowrates
Number of stages above feed stage
Number of stages below feed stage
Feed stage
Vapour mass specific enthalpy at P,T and composition of the feed
Actual feed mass specific enthalpy
Liquid mass specific enthalpy at P,T and composition of the feed
zFi F = xD i D + xB i B
(Eq. 51)
components
x Di = 1
i
(Eq. 52)
components
x Bi = 1
i
(Eq. 53)
xDi xBHK
xDHK xBi
= i,HK Nmin
components
i
i,HK zF i
components
i
i,HK xDi
NNmin
N+1
i,HK
i,HK
(Eq. 54)
=1q
(Eq. 55)
= 1 + R min
(Eq. 56)
= 0.75 0.75 (
R = R factor R min
RRmin 0.5668
R+1
(Eq. 57)
(Eq. 58)
A-3
(D/F)ratio =
(Eq. 59)
N = Nabove + Nbelow + 1
Nabove
Nbelow
=(
zF HK
zF LK
xBLK
xDHK
(Eq. 60)
0.206
B
) )
(Eq. 61)
NF = Nbelow + 1
q=
(Eq. 62)
hFV hF
(Eq. 63)
hFV hFL
Dialog box
A-4
Main variables/parameters
Table 21: List of Heat exchanger detailed sizing model main variables and parameters.
Variable
R
S
F
Th,0
Th,L
Tc,0
Tc,L
LMTD
U
A
Q
do
di
L
ntubes
Ftube
utube
,
Pt
Pt,factor
gPROMS name
R
S
F
T_hot_0
T_hot_L
T_cold_0
T_cold_L
LMTD
U_guess/U_real
A
Q
OD
ID
L_init/L
no_tubes_init/no_tubes
F_tube
u_tube
no_passages_tubes
p_t
p_t_factor
Re_tube
Pr_tube
Nu_tube
j_h_tube
htc_tube
A_s
D_s
l_B
F_shell
G_s
d_e
D_b
Baffle_spacing
Re_shell
Pr_shell
Nu_shell
j_h_shell
htc_shell
-
Units
Significance
Variable necessary for F calculation
Variable necessary for F calculation
LMTD correction factor
K
Hot stream temperature at z=0
K
Hot stream temperature at z=L
K
Cold stream temperature at z=0
K
Cold stream temperature at z=L
K
Log mean temperature difference
W/m2K
Global heat coefficient
m2
Heat transfer area
kJ/s
Heat duty
m
Tube external diameter
m
Tube internal diameter
m
Tube length
Number of tubes
kg/s
Tube side mass flowrate
m/s
Tube side velocity
kg/m3
Tube side density (returned by Multiflash)
Number of tube passages
m
Tube pitch
Ratio between tube pitch and external diameter
Pa.s
Tube side density (returned by Multiflash)
Tube side Reynolds number
Tube side Prandlt number
Tube side Nusselt number
J/kgK
Tube side heat capacity (returned by Multiflash)
W/mK
Tube side thermal conductivity (returned by Multiflash)
Tube side heat transfer factor
Pa.s
Viscosity at wall temperature (returned by Multiflash)
W/m2K
Tube side partial heat coefficient
m2
Shell area
m
Shell internal diameter
m
Shell baffle spacing
kg/s
Shell mass flowrate
Kg/m2s
Shell side mass flux
m/s
Shell side velocity (returned by Multiflash)
kg/m3
Shell side density (returned by Multiflash)
m
Shell equivalent diameter
m
Shell bundle diameter
Shell baffle spacing
Shell side Reynolds number
Shell side Prandlt number
Shell side Nusselt number
Pa.s
Shell side viscosity (returned by Multiflash)
J/kgK
Shell side heat capacity (returned by Multiflash)
W/mK
Shell side conductivity (returned by Multiflash)
Shell side heat transfer factor
W/m2K
Shell side partial heat coefficient
W/mK
Shell side thermal conductivity (returned by Multiflash)
(continues on next page)
A-5
,
,
R=
S=
N_b
shell_side_fouling_factor
tube_side_fouling_factor
k_tube
PD_shell
PD_tube
W/m2K
W/m2K
W/mK
Bar
bar
Number of baffles
Shell side fouling factor
Tube side fouling factor
Tube material conductivity
Shell side pressure drop
Tube side pressure drop
(Eq. 64)
ABS(Tc0 TcL )
ABS(Tc0 TcL )
(Eq. 65)
F=
R2 +1
1S
ln(1RS)
R1
2S(R+1R2 +1)
ln(
2S(R+1+R2 +1)
LMTD =
(Eq. 66)
)
(Eq. 67)
Th Tc0
0
Th TcL
L
Q = U A F LMTD
(Eq. 68)
A = do L ntubes
(Eq. 69)
utube =
Ftube
tube
(Eq. 70)
ntubes
di 2
npassages,tubes
4
Pt = Pt,factor do
Retube =
Prtube =
(Eq. 71)
tube utube di
(Eq. 72)
tube
tube cptube
(Eq. 73)
ktube
(Pt do )Ds lB
As
= ushell shell
de = 1.27(or 1.10)
Db = do (
lB = baffle spacing Ds
Reshell =
Prshell =
A-6
shell ushell de
shell
shell cpshell
kshell
htube di
ktube
(Eq. 74)
(Eq. 76)
2
(P2
t 0.785do )
ntubes n1
K1
wall
(Eq. 75)
Pt
Fshell
tube 0.14
do
(Eq. 77)
(Eq. 78)
(Eq. 79)
(Eq. 80)
(Eq. 81)
L
lB
shell 0.14
wall
hshell de
kshell
1
hshell
(Eq. 82)
(Eq. 83)
1
Ffactor,shell
Pshell = 8 jf,shell
Ptube = 8 jf,tube
Ds
de
L
di
do
2k
L
lB
ln
u2
tube
2
do
di
u2
shell
2
do
di
htube
shell (
tube (
1
Ffactor,tube
shell 0.14
wall
shell 0.14
wall
(Eq. 84)
(Eq. 85)
(Eq. 86)
A-7
A-8
A-9