Extended Abstract
Extended Abstract
Extended Abstract
Abstract
This work studied the hydrogen purification in the cold box (Aromatics-2 facility) at Dow’s plant in The Netherlands,
with the main aim of rising the benzene production in 35% at the hydrogenation reactor (Aromatics-1 facility), whenever
the hydrodealkylation reactor is stopped.
A model of the cold box was developed by using the Aspen Plus® V8.0 software and it was validated based on the
plant process data, revealing deviations inferior to 10%.
Subsequently, three distinct scenarios were analysed by using the developed Aspen model. Case Study 1 corresponded
the current plant situation, in which the hydrodealkylation reactor is working. Case Study 2 represented the main goal of
this study, i.e. it considered the hydrodealkylation reactor stopped. Case Study 3 was similar to Case Study 2 but it
considered the washing column working normally.
In the three case studies, a cold box feed of 22,7 t/h and a hydrogen purity of 90% (v/v) were set. Sensitivity analyses
were conducted for Case Study 2 whereas Case Studies 1 and 3 were analysed by single simulations.
In Case Study 1, the heavies ended up in the ethane outlet stream, unlike Case Study 3, in which the heavies ended up
in the methane outlet stream. In Case Study 2, the methane stripper was dispensable and yet the model yielded trustworthy
results.
Keywords: Aspen; benzene production; cold box; hydrodealkylation; hydrogen purification; hydrogenation reactor.
1
After leaving the dryer at 10°C and 53 bar, the feed
gas stream flows towards the cooler no.2, where it is
cooled down to -60°C in counter-current with respect to
the cold product streams rich in hydrogen, methane and
ethane, respectively. Figure 2 represents the flowsheet of
the cryogenic section.
The next cryogenic exchanger is called deep cooler
no.1 because it is located inside the cold box. This
equipment is responsible for cooling down the feed gas
to -91°C in counter-current with respect to the cold
product streams. After passing through the deep cooler
no.1, a condensate product is obtained.
Afterwards, the flow rate of condensate product is
slightly increased by using an ethylene chiller, which
uses ethylene as refrigerant. The ethylene chiller is the
Figure 1 Schematic overview of the Aromatics-2 facility. cold source of the plant responsible for the refrigeration
TX - toluene/xylenes mixture, HDA – hydrodealkylation. The needs due to heat gains throughout the process, and in
blue dashed line indicates the option to send high purity H2
the dryer cooling. Thereby, the ethylene evaporation is
from the cold box unit (Aro.2) to the hydrogenation reactor
(Aro.1), when HDA reactor is stopped. used as cold source for the cooling and liquefaction of
the process streams. In this apparatus, the ethylene
In Terneuzen site, the HDA reactions that produce refrigerant is totally vaporized by removing the heat
benzene are the following: from the feed gas stream. Then, the vaporized ethylene
is discharged to the suction vessel of the ethylene plant.
C 7H 8 + H2 à C6H6 + CH4 (1) The separator no.1 is located downstream of the
(toluene) (hydrogen) (benzene) (methane) ethylene chiller, and a flash separation takes place
C8H10 + 2 H2 à C6H6 + 2 CH4 (2) therein. The liquid product that leaves the bottom of the
(xylene) (hydrogen) (benzene) (methane) separator is rich in ethane and most hydrogen sulfide is
C8H10 + H2 à C6H6 + C2H6 (3) dissolved in this stream. This liquid stream is expanded
(ethylbenzene) (hydrogen) (benzene) (ethane) and then stripped in the methane stripper. It is
noteworthy that the hydrogen sulfide in the feed gas
Obviously, for HDA reactions to take place it is stream does not entail troubles in the cryogenic section
necessary to feed certain flow rates of TX and hydrogen since the quantity remaining after the drying process is
into Aro.2. This hydrogen comes from the ethylene plant dissolved in the liquefied ethane and does not cause
towards the methanator, located at Aro.2 facility. The cloggings during the re-evaporation.
HDA reactor is the final step of the aromatics facilities Meanwhile, the gas stream that leaves the top of the
wherein benzene production increases by means of the separator no.1 is rich in methane and hydrogen. Its
hydrodealkylation reactions. Depending on the daily temperature has to diminish in order to raise the
market situation and the benzene production costs, the concentration of H2 up to 90% (v/v). Thus, another
HDA reactor might be shut down temporarily, meaning cryogenic exchanger, named deep cooler no.2, is used to
that the TX mixture may be directly sold in the market cool down the feed gas stream to -136°C by means of
instead of reused at Aro.2 facility. the cold products streams rich in methane and hydrogen,
The outlet product from the HDA reactor is cooled respectively. At this temperature, part of the methane in
down through a chain of heat exchangers, allowing the the feed gas condenses and forms a new cooling agent.
separation of a liquid phase rich in benzene. The gas Thereby, the product stream rich in liquid methane is
phase is purified in a washing column and a cryogenic collected in the separator no.2 wherein the flash
section, and in between these units, the gas phase separation takes place yielding two different products.
undergoes a pre-cooling and a drying process. The gas stream that leaves the top of the separator is
rich in hydrogen with a purity in the range 90-95% (v/v).
2.2 Cryogenic section The purified hydrogen flows back into the system
The cryogenic section contains all the cooling through the heat exchangers battery, being heated up
equipments that maintain the process streams at negative until it leaves the cryogenic section at 2°C. In the
temperatures. Basically, it comprises the cooler no.2 and meantime, the liquid product that leaves the bottom of
the cold box. the separator, rich in methane, is expanded by a control
The cold box unit consists of a box containing a valve undergoing a pressure drop from 53 to 5,6 bar,
package of equipments, such as piping, distillation assisted by the pipe diameter increase from 3” to 10”.
columns, heat exchangers and instrumentation to Then, the methane product stream flows back to the deep
maintain liquids at extremely cold conditions. The space cooler no.2, where it is the main responsible for the
between the equipments and the inner walls of the cold cooling effect. It is noteworthy that the outlet
box is filled with perlite insulation. To prevent the temperature of the cryogenic exchanger deep cooler no.2
formation of explosive gas mixtures and humidity, the is determined by the quantity of the expanded methane.
cold box is purged with nitrogen under a low pressure. In case a heavy refrigeration is needed in the plant, a
This procedure prevents humidity deposits, such as ice, higher quantity of methane should be expanded. Further,
that would impair the insulating effect. the methane cold product flows back through the several
heat exchangers, being heated up until it leaves the
2
cryogenic section at -1,8°C. Moreover, the recovered A few tasks had to be performed to build the process
methane from the methane stripper (stream 30) and from model of Dow´s cold box unit, namely:
the separator no.3 (stream 17) are added to the methane • Selection of the property method (NRTL modifications
product stream (streams 13 and 26). of the PSRK-Predictive Redlich-Kwong-Soave
On the other hand, the liquid product that leaves the equation of state);
bottom of the separator no.1 is expanded and supplied to • Selection of the convergence tolerance (10-7) and
the head of the methane stripper. In this equipment, number of flash iterations (75);
CH4/C2H6 separation takes place as the feed stream flows • Identification of the relevant components involved in
in counter-current with vapour, via 5 trays, such that a the process [(hydrogen (H2); methane (CH4); ethane
fraction rich in ethane is formed in the sump and a (C2H6); ethylene (C2H4); hydrogen sulfide (H2S);
fraction rich in methane is obtained at the top of the toluene (C7H8) and benzene (C6H6)];
column, where a pressure control valve maintains the • Design of the flowsheet diagram.
rectification system pressure. A fraction of the main feed
gas stream is used in the reboiler to deliver heat for the 3.2 Flowsheet diagram
rectification system.
The liquid product that leaves the bottom of the Many process data of Dow’s cold box unit was
methane stripper is rich in ethane and flows towards the required to run a realistic simulation by using the Aspen
deep cooler no.1, where it plays the role of cold fluid Plus software, but there was a wide lack of available
occurring its partial vaporization in counter-current with information. To start building the process flow diagram
the main feed gas stream. After partially vaporized, this of the cold box, a few input parameters, such as the
stream undergoes the final separation in separator no.3. temperature, pressure, and composition of the cold box
As aforementioned, the final flash separation of feed stream were demanded by the software. To get the
CH4/C2H6 occurs, whereas the remaining methane maximum relevant plant process data, it was essential to
(stream 17) is added to the methane outlet stream use the tool Aspen Tag Browser, containing the majority
(stream 13). From the separator no.3, a bottom liquid of the daily plant process data, named InfoPlus.21
product rich in ethane is pumped towards the cooler no.2 (IP.21)[3]. Using the Aspen Tag Browser, the temperature
(outside the cold box) where it is vaporized, leaving the of 10°C and the pressure of 53 bar were specified for the
cryogenic section at 5°C. The ethane outlet stream heads feed gas stream. However, the current composition of
to the ethane cracking furnaces in the ethylene plant, to the cold box feed stream was not available because there
be cracked. was not a composition analyser in this zone neither
Although the cryogenic section comprises the cold inside the cold box. The only available record of this
box unit and the cooler no.2 right before it (feed of the composition was found in documents from 1980, which
cold box), for the sake of simplicity the designation of were certainly not updated, thus it was not used as
“cold box unit” stands for the whole cryogenic section current process data, though a comparison would be
throughout most text. interesting.
In this study, the main purpose of the cold box unit Therefore, some samples were taken from the cold
was to cool down the feed gas until the hydrocarbons, box outlet streams, because the inside of the box was not
CH4 and C2H6, become liquid and separate from H2 in accessible for this kind of operations. The samples were
three distinct outlet streams. Moreover, the hydrogen taken while the process was running smoothly, on
purity of the feed gas would increase to be reused in the November 19, 2014. The composition of each outlet
HDA reactor (Aro.2) or in the hydrogenation reactor stream was analysed and the data are listed the Table 1.
(Aro.1). The required composition for the simulation input was
the composition of the cold box feed stream, which was
obviously related with the data of the outlet streams,
3. Modelling of Dow’s cold box unit shown in Table 1. The total mass flow rate of each outlet
stream was selected from IP.21 data based on the
3.1 Materials sampling date and time.
The model for Dow’s cold box unit was developed As a reminder, this simulation run in steady state,
by using the Aspen Tech software. The acronym and assumed that there were no mass losses in the cold
“ASPEN” stands for “Advanced System for Process box, i.e. the mass flow rate of the fluid entering in the
Engineering”. Aspen is a process simulation software cold box was equal to the sum of the mass flow rates
widely used in industry and known as an important tool going out. Hence, the hydrogen (H2), methane (CH4) and
for chemical engineers, thus it is often used to conduct ethane (C2H6) mass flow rates (and respective
simulations of chemical processes. Given a process compositions) of the cold box feed stream were
design and an appropriate selection of thermodynamics calculated through their partial mass balances and
models, Aspen uses mathematical models to predict the presented in Table 2. Although there were many
process performance. Furthermore, this software compounds in the process, the hydrogen, methane and
performs mass and energy analyses. Among the wide ethane represented over 99% (w/w) of the total mass
variety of Aspen packages, Aspen Tech Process flow rate, meaning that the other compounds
concentrations were negligible. Thus, a major
Modelling V8.0 - Aspen Plus® V8.0 was selected to
simplification was assumed for the simulation, viz. the
conduct the current simulations since it is a steady state
cold box feed gas stream only contained the main
process simulator and the most adequate for the level of
compounds in Table 2.
complexity of this specific process [2].
3
As the main input parameters were obtained, Aspen based on the process operating conditions by the
could finally start building the cold box flowsheet with sampling date and time. The proposed overall process is
all the streams and blocks identified in Table 3. The represented in Figure 2.
inlets specifications set throughout the simulations, were
Table 1 Compositions of the cold box outlet streams (sampling on November 19, 2014).
Hydrogen outlet stream Methane outlet stream Ethane outlet stream
C (vol%) C (wt%) C (vol%) C (wt%) C (vol%) C (wt%)
H2 90,6 53,5 8,42 1,05 0,04 <0,01
CH4 9,09 42,6 82,0 81,1 7,15 3,75
C2H6 0,01 0,09 9,32 17,3 89,8 88,2
Others 0,29 3,75 0,24 0,63 3,02 8,03
Total 100 100 100 100 100 100
Table 2 Compositions and mass flow rates of the cold box feed stream.
Cold box feed stream
Qm (t/h) C (wt%) C (vol%)
H2 3,3 14,6 60,2
CH4 15,3 67,2 34,8
C2H6 4,1 18,2 5
Total 22,7 100 100
4
Table 3 Blocks description according to Aspen library and their specifications.
Name Quantity Description Specifications
ΔPmain stream = 0,3486 bar ; ΔPH2 stream = 0,3486 bar ;
Cooler no.2 1 Multi stream heat exchanger ΔPCH4 stream = 0,16319 bar ; ΔPC2H6 stream= 0,16319 bar ;
Tmain stream= -60,5°C; TCH4 stream = -1,75°C; TC2H6 stream = 4,98°C
ΔPmain stream = 0,3486 bar ; ΔPH2 stream = 0,3486 bar ;
Deep cooler
1 Multi stream heat exchanger ΔPCH4 stream = 0,16319 bar ; ΔPC2H6 stream= 0,16319 bar ;
no.1
Tmain stream= -91,05°C; TCH4 stream = -65°C ; Vf C2H6 stream = 0,628
Deep cooler ΔPmain stream = 0,3486 bar ; ΔPH2 stream = 0,3486 bar ;
1 Multi stream heat exchanger
no.2 ΔPCH4 stream = 0,16319 bar ; Tmain stream= -140°C (DesignSpec)
Ethylene
1 Thermal and phase state changer ΔPmain stream = 0,3486 bar; Tout= -95°C (DesignSpec)
chiller
Reboiler 1 Thermal and phase state changer ΔP = 0 bar ; Q = -88,21 kW
Methane 1 Rigorous 2-phase fractionation
No. trays = 5 ; PTop= 6,92 bar; PBottom= 6,72 bar
stripper for single columns
Separator 3 Two outlet flash ΔP = 0 bar; Q = 0 kW
Pump 1 Pump Pout = 7 bar
Valve 1 Valve Pout = 5,6 bar
Mixer 3 Stream mixer ΔP = 0 bar
Splitter 1 Stream splitter ΔP = 0 bar ; Limit flow rate of S9 = 1136,1 kg/h
5.1 Case Study 1 Using the new model (Table 7), a single simulation
was run, considering the heavies in the cold box feed
The Case Study 1 corresponded to the current plant stream, whereas its total feed flow rate remained
situation, in which the HDA reactor was normally constant at 22,7 t/h.
working, as well as each and every equipment in the
Aromatics-2 facility. 5.1.1 Cold box mass balances
So far, the developed Aspen model had just
considered the main components, CH4/H2/C2H6. To get a The composition of the cold box feed stream was
more realistic response from the model and predict how slightly different from the one set in section 3. Table 8
the plant worked, it was mandatory to consider small lists the feed flow rates, weight and volume percentage
flow rates of heavies, namely benzene (C6H6) and compositions assumed in Case Study 1.
toluene (C7H8), in the cold box feed stream (Table 8). The easiest way to check the heavies evolution
Actually, these heavies were known to end up in the through the cold box, is by observing the compositions
ethane outlet stream, reinforcing the motivation of this of the cold box outlet streams (Table 9). Assuming the
case study to verify if this fact was really predicted by feed conditions of Table 8, most heavies ended up in the
the model. Moreover, a small flow rate of hydrogen ethane outlet stream indeed, as expected.
sulfide (H2S) was also considered in the cold box feed
stream to prevent coking in the HDA reactor.
Table 8 Case Study 1 – feed flow rates and compositions of the cold box unit.
H2 CH4 C2 H6 C6 H6 C7 H8 H2S Total
Qm (t/h) 3,31 15,2 4,12 0,023 0,023 2,27x10-3 22,7
C (wt%) 14,6 67,1 18,1 0,10 0,10 0,01 100
C (vol%) 60,2 34,8 5,02 0,01 0,01 0,002 100
7
Table 9 Case Study 1 – Aspen results for the operating conditions, flow rates, and compositions of the cold box outlet streams.
The Case Study 2 represented the main goal of this purity in the cold box feed stream varied from 0 up to
work. Unlike Case Study 1, in this plant scenario the 100% (v/v), whereas the total feed flow rate remained
HDA reactor was stopped. This case study is sketched in constant at 22,7 t/h.
Figure 10, displaying the methanator outlet stream, The sensitivity analyses revealed that the model
which was compressed and transported towards the cold representing the plant scenario of Case Study 2, was
box unit through the existent piping and equipments only valid from 22,4 up to 90% (v/v) of H2 in the cold
without suffering any chemical reaction or physical box feed gas.
process since these equipments were also stopped. The
high purity hydrogen leaving the cold box unit was 5.1.1 Cold box mass balances
subsequently fed to the hydrogenation reactor located in
the Aromatics-1 facility. As sketched in Figure 10, the The Aspen results pointed out that up to 10% (v/v) of
green blocks were working normally, whilst the grey H2 in the cold box feed stream, the separation of H2 and
blocks were shut down. CH4 did not occur at all because the model was dealing
with supercritical fluids. Besides, the model faced
convergence errors from 10 up to 22,4% (v/v) of H2,
thus it did not yield reliable results in that range.
The results of the sensitivity analysis are plotted in
the validity range of the model, Figure 11, that displays
the composition distributions of the cold box outlet
streams.
Figure 11 shows that the ethane outlet stream was
absent (null flow rate), since C2H6 was not taken into
account in this case study, as the HDA reactor was
stopped. The methane outlet stream revealed a realistic
composition distribution, starting with a high flow rate
of CH4, and gradually reducing as the flow rate of H2 in
the cold box feed stream increased. Last, but not least,
Figure 11 presents the expected response of the total
flow rate of the hydrogen outlet stream. The latter
gradually rose as the H2 purity increased in the cold box
feed stream up to 90% (v/v). Beyond 90% (v/v) of H2 in
the cold box feed stream, the total flow rate of the feed
stream exceeded 22,7 t/h and the results were
unrealiable.
Figure 10 Schematic overview of Case Study 2 (Aro.2
The equipment mainly responsible for the increase of
facility): grey blocks stopped and green blocks working. LHC
I/II: ethylene plants; DC-MTH: methanator; GB1, 2: H2 purity was the deep cooler no.2, in which a
compressors; DA-WASH: washing column; DC-HDA: DesignSpec varied the outlet temperature from -200 to
hydrodealkylation reactor; CB: cold box unit. 0°C to condense methane till a H2 purity of 90% (v/v)
was achieved. A constant H2 purity of 90% (v/v) was
Using the new Aspen model, the plant scenario of
achieved in the hydrogen outlet stream throughout the
Case Study 2 technically meant to vary the composition
model validity range. If H2 purity in the cold box feed
of the cold box feed stream. To observe the effect of the
stream would exceed 90% (v/v), the deep cooler no.2
composition variation on the cold box performance, would presumably condense H2 instead of CH4, but as
several sensitivity analyses were performed, in which H2 8
this process would be unfeasible at the operating 5.1.4 Ethylene chiller
conditions specified, the separation of CH4/H2 would not
occur, i.e. everything that would enter into the cold box As Case Study 2 did not consider ethane neither
would end up in the hydrogen outlet stream. heavies in the cold box feed gas, the flow rate of
ethylene refrigerant fed to the cold box was lower than
24
in Case Study 1.
Total flow rates of the outlet
20
The Aspen results showed that the ethylene chiller
partially condensed the feed gas, only containing H2 and
streams (t/h)
16
Hydrogen
stream
CH4, in a small range from 22,4% up to 24,6% (v/v) of
12
Methane
H2 in the cold box feed stream (Figure 13). For hydrogen
stream
8
Ethane purities beyond 24,6% (v/v), the ethylene chiller could
stream
4
not proceed with CH4 condensation and just a cooling
process took place.
0
0
10
20
30
40
50
60
70
80
90
100
Purity of H2 (vol%) in the feed gas stream of the cold 3
0
Figure 12 displays a constant response for the 0
10
20
30
40
50
60
70
80
90
100
temperature of the methane outlet stream, since this
Purity of H2 (vol%) in the feed gas stream of the
variable was specified in the model. Conversely, the cold box
temperature of the hydrogen outlet stream showed a
steep variation of about 100°C because it was the sole Figure 13 Case Study 2 - sensitivity analysis regarding the
refrigerant flow rate in the cold box unit.
process variable.
Anyway, in the range of H2 purity in the cold box
feed stream practically applicable in the plant (60%-75% 5.1.5 Methane stripper
(v/v)), the temperature presented a rather small variation,
meaning that it would be safe working within this range The Aspen results of the methane stripper were in
without significant disturbances of the hydrogen outlet close agreement with the ethylene chiller results, i.e.
stream temperature. since ethane was absent in the cold box feed stream, the
Hydrogen Stream
Methane Stream
stripper feed was null (Figure 14). Yet, small flow rates
10
across the methane stripper were noticed in a small
Outlet streams temperature
-50
-65
Top
Bottom
Feed
-80
3
Methane stripper flow rates
-95
-110
Purity of H2 (vol%) in the feed gas stream of the 2
cold box
(t/h)
Table 10 Case Study 3 – feed flow rates and compositions of the cold box unit.
H2 CH4 C6 H6 C7 H8 Total
Qm (t/h) 4,05 18,62 0,027 0,027 22,7
C (wt%) 17,8 81,9 0,125 0,125 100
C (vol%) 63,3 36,6 0,01 9,5x10-3 100
Table 11 Case 3 – Aspen results for the operating conditions, flow rates, and compositions of the cold box outlet streams.
11