UniSim Design Tutorial for

CHEE470
Queens University Department of
Chemical Engineering
00!
229462592.doc
- 2 -
Introduction
UniSim (formerly HYSYS) is a program that can be used to design chemical plants. It is
built around:
a library of the physical properties of a large number of chemical species
a set of subroutines to estimate the behaior of many types of plant e!uipment
(heat e"changers# reactors# etc.)
a graphical user interface to accept specifications for the case# and display results
$he user describes the process in terms of pieces of e!uipment interconnected by process
streams# and the program soles all the mass%energy%e!uilibrium e!uations# ta&ing into
consideration the specified design parameters for the units.
It is a ery comple" system# and there is no 'ay that this tutorial is going to demonstrate
all of the features. $he features that 'ill be sho'n are the ones that 'ill prepare you to
tac&le the plant design assignment in (H))*+,.
-i&e most programs of this type# operations can be done in different 'ays. In general# this
tutorial 'ill only describe one 'ay. You 'ill find other methods in the UniSim
documentation# but the ones sho'n here are best suited for people 'ho are ne' to the
program.
Hints for Success in Modeling
.. /uild the model one step at a time. 0eople ne' to this instinctiely 'ant to start
by adding many streams and unit operations# and then try to get the 'hole mess
'or&ing. $his is futile. 1dd the elements one at a time. 2et one 'or&ing before
you go on to the ne"t one.
2. Sae a 'hole series of bac&ups# not 3ust the latest 'or&ing ersion. 4is& space is
cheap. If you get into trouble# you may find that it is difficult to restore the model
to its preious state. 5ften you are better to retriee a preious ersion and update
it.
6. 0ut meaningful names on all streams and units. 4ebugging is difficult 'hen you
are trying to remember if stream S22 is the distillate or the bottoms in a
distillation column.
*. If a piece of e!uipment does not 'or& although the parameters all loo&
reasonable# try deleting the unit and reconstructing it.
Steps in Developing a Model of a Chemical Process
.. Select the units that you 'ant to 'or& 'ith. 4o you 'ant &ilograms and 7(# or
pounds and 7 89
2. Select the thermodynamic methods that 'ill be used for predicting physical
properties. $he decision should be based on the type of chemical species
inoled.
- 6 -
6. Specify the chemical species that 'ill be present in the process. 1t this stage you
may be gien some adice about the ability of the selected thermodynamic
method to handle these chemicals.
*. If the process inoles reactions# proide information such as stoichiometry and
&inetic constants.
:. /uild the model by adding streams and e!uipment one at a time.
;. If the process contains recycle loops# deal 'ith closing them.
+. Use the UniSim utilities to get additional information such as the mechanical
design of distillation column trays.
<. 0rint a report describing the results of the simulation.
The Methanol Process
=ethanol can be made from hydrogen plus carbon mono"ide and%or carbon dio"ide.
2 H
2
> (5 ?@ (H
6
5H
6 H
2
> (5
2
?@ (H
6
5H > H
2
5
Aecent studies suggest that the first reaction actually proceeds as
(5 > H
2
5 ?@ H
2
> (5
2
(the 'ater gas shift reaction)
follo'ed by the second reaction.
8or this e"ercise 'e 'ill 'or& 'ith the simplest ersion B the second reaction only. /y
the 'ay# running this reaction bac&'ards proides a method of operating a hydrogen fuel
cell 'ith methanol as a feed.
$he follo'ing diagram sho's the process 'e 'ill 'or& on. It is important to recogniCe
that this is not suggested as a good 'ay to ma&e methanol. $he design has been
formulated to demonstrate many &ey aspects of UniSim# 'ithout getting oer'helmed by
detail.
1 mi"ture of H
2
and (5
2
is heated to the re!uired temperature and fed to a stirred reactor.
1s noted aboe# the reaction is 6 H
2
> (5
2
?@ (H
2
5H > H
2
5.
- * -
$he product of the reaction is partially condensed. $he apour (mostly H
2
and (5
2
) is
compressed and recycled bac& to the beginning of the process. $he li!uid (mostly
(H
2
5H and H
2
5) is fed to a distillation column.
$he column produces a product stream (mostly methanol) and a 'aste stream (mostly
'ater). $he product is cooled to a temperature that is reasonable for storage. 1 pump is
re!uired proide cooling 'ater for this heat e"changer.
Setting up Data for the Model
5pen UniSim.
(lic& on the ne' file icon and get:
Units
De 'ill use the default SI units# so no action is re!uired. In future cases you may 'ant to
use different units. See section 6.2.2 of EUniSim 4esign $utorials and 1pplications.pdfF.
Thermodynamics
- : -
(lic& on the EF8luid 0&gsF tab.
E8luid 0ac&ageF is UniSimGs terminology for a collection of data that includes all the
thermodynamic# component# and reaction parameters re!uired to run the model.
It is possible to hae more than one pac&age in a model. 8or e"ample# it 'ould be
possible to use one thermodynamics model in the reactor# and another in the distillation
column. De 'ill 3ust hae one pac&age.
(lic& on E1ddF and scroll to find ESAHF. $his selects the Soae-Aedlich-H'ong method#
a popular e!uation of state model.
- ; -
Iote that the default name E/asis-.F is gien to the pac&age# and our components 'ill go
into E(omponent -ist-.F.
Components
(lose the fluid pac&age 'indo' and clic& on the E(omponentsF tab of the ESimulation
/asis =anagerF.
- + -
Io'# select E(omponent -ist B .F# clic& on EJie'F and enter E(52F into field E=atchF.
Io' clic& on E1dd 0ureF# and do the same operation 'ith EHydrogenF.
-etGs do methanol differently. Select the E8ormulaF option instead of E8ull Iame %
SynonymF and enter E(H*5F@
- < -
4o the same 'ith EH25F.
De no' hae all the components 'e need.
- K -
$his 'ould be a good time to sae the case. (lose the components 'indo' and clic& on
the E)nter Simulation )nironmentF button of the ESimulation /asis =anagerF 'indo'.
4o the usual E8ile?@SaeF operation and call the file tutor,..
Reactions
Io' 'e need to return to the Simulation /asis )nironment. (lic& on the bea&er.
(lic& here
- ., -
In the ESimulation /asis =anagerF 'indo'# select the EAeactionsF tab.
- .. -
Io'# clic& on E1dd A"nF. Dhen the EAeactionsF 'indo' appears# select EHineticF and
clic& on the E1dd AeactionF /utton.

1 EHineticF reaction is one for 'hich 'e 'ill supply the &inetic constants that define the
rate of reaction. $his allo's us to siCe the reactor. If 'e 'ere only interested in simuating
the heat and mass balances# 'e could use the simpler form# a E(onersionF reaction.
$hen 'e 'ould only hae to define the percent conersion.
(lic& on ELL1dd (ompGF and select the components as sho'n.
- .2 -
Io' fill in the follo'ing data:
$he significance of the numbers should be obious from the definition of the reaction:
6 H
2
> (5
2
?@ (H
2
5H > H
2
5. Iote that EA"n-.F has been replaced by a more
meaningful name E=ethanol AeactionF.
Io' loo& at the E/asisF tab.
- .6 -
/asis ? =olar (oncn means that the reaction rate e!uation uses concentrations of the
reactants in moles%m
6
.
/ase (omponent ? (52 means that the reaction rate e!uation describes the rate of
consumption of (5
2
# not consumption of H
2
or production of =ethanol (since (5
2

consumption ? methanol production# methanol could be specified here).
A"n 0hase ? -i!uid0hase means that the reaction ta&es place in the li!uid. Since 'e 'ill
not hae any li!uid in the reactor# this is not helpful. (hange it to EJapour0haseF.
8inally# change EAate UnitsF from E&gmole%m6-sF to E&gmole.m6-hF. Since eerything
else in the modle is in units of hours# it is best to be consistent.
$he important consideration is to ensure that the treatment here is compatible 'ith 'hat
'as used in generating the constants describing the reaction rate (probably from lab data).
- .* -
Io' go to the E0arametersF tab and enter the follo'ing aluesF
$hese are the &inetic constants for an 1rrhenius e!uation: & ? 1 e
(-)%A$)

Iote that the bar in the lo'er right has turned green and says EAeadyF. $his means that
all the necessary data hae been supplied# and are alid (that does not necessarily mean
correct).
(lose the 'indo' and return to the Simulation /asis =anager. Iote that E=ethanol
AeactionF has been added to the list of reactions.
- .: -
UniSim has put together data for a set of reactions for us called E2lobal A"n SetF. It only
contains one reaction# E=ethanol AeactionF# but 'e could add others (e.g. side reactions
that produced small !uantities of ethanol and acetone). If our model contained an
unrelated group of reactions (e.g. if 'e put a 'ater gas shift reactor on the front of the
process) 'e could create another Aeaction Set. In this 'ay 'e could model different
reactors using different reactions.
$o see 'hat is in our Aeaction Set# clic& on EJie' SetF.
- .; -
$his does not tell us anything that 'e do not already &no'# but it confirms that our
reaction is really there.
$he ne"t step is ery important# and it is easy to forget to do it. If you find yourself
unable to model a reactor because the reactions you need do not e"ist# it is probably
because you forgot this step.
1t this point UniSim has put E=ethanol AeactionF in E2lobal A"n SetF# but it has not put
E2lobal A"n SetF in E/asis-.F. Dhen building the reactor it 'ill loo& in E/asis-.F for the
reaction data.
/ac& in ESimulation /asis =anagerF clic& on E1dd to 80F.
- .+ -
Io' clic& on E1dd Set to 8luid 0ac&ageF.
Iote that E/asis-.F is no' listed in E1ssoc. 8luid 0&gsF.
De hae no' finished the 3ob of supplying the data for physical properties# components#
and reactions. Iote that 'e hae left the names E2lobal A"n SetF# E(omponent -ist-
.Fand E/asis-.F 'ith their default alues. If 'e had more than one basis# component list#
or reaction set# 'e 'ould hae been 'ise to change the names to something more
meaningful# 3ust li&e 'e changed EA"n-.F to E=ethanol AeactionF. /ut if there is only
- .< -
one of each# 'e are not going to get confused about 'hich one 'e are dealing 'ith in the
model# and this is unnecessary.
$o proceed 'ith building the process model# clic& on EAeturn to Simulation
)nironmentF.
uilding the Process !lo" Diagram
1l'ays sae the case at this point in the deelopment. In this 'ay# regardless of ho'
scre'ed up the model gets# you can al'ays go bac& to a alid case
4o a ESaeF follo'ed by ESae 1s tutor,6F.
De no' hae a blan& screen on 'hich 'e can start to build a 084 (0rocess 8lo'
4iagram) that 'ill define the process.
!eed Section
- .K -
Dhile 'e are 'or&ing on the 084 'e 'ill re!uire the use of the unit
operations palette as sho'n on the left. If it is not present# do
E8lo'sheet?@0aletteF.
$he first thing 'e 'ant to do is create the feed to the system. 4ouble
clic& on the blue material stream icon to bring up an empty stream
'indo'.
Iote that this is reminding us that the properties are being estimated
'ith the /asis-. pac&age.
)nter the follo'ing data:
Stream Iame ? 8eed
$emperature ? *, 7(
0ressure ? *,,, &0a
=ass 8lo' ? .,,, &g%h
=aterial Stream
- 2, -
$o finish the stream definition# 'e need to specify the composition. (lic& on
E(ompositionF at the left side of the 'indo'.
De 'ant to specify a stoichiometric ratio# so mole fraction is the correct units.
- 2. -
Specify:
(52 ? .2:
H2 ? .+:
EIormaliCeF
2o bac& to the E(onditionsF 'indo'.
- 22 -
Iote that the stream is no' fully defined# and UniSim has calculated the ariables that
you did not specify. 8rom no' on 'e 'ill only be able to change the blue alues# not the
blac& ones.
You may hae notices that there 'ere t'o stream icons on the palette.
$he first (blue) is a material stream going from one piece
of e!uipment to another. $he second (red) is an energy stream. $hese 'ill be handled by
UniSim. $he energy used or generated by e!uipment 'ill be displayed in these streams.
$his information 'ill be useful in costing the utilities (steam# electricity etc.) used by the
process.

Io' let us add a mi"er to combine the feed 'ith the recycle stream. In most cases# but
not all# a mi"er is not a piece of e!uipment B 3ust t'o pipes coming together.4ouble clic&
on the mi"er icon in the palette.
8or EIameF# specify EAecycle =i"erF. $hen clic& on EMMStream@@F in the EInletsF list.
You 'ill be presented 'ith a list of acceptable streams.
- 26 -
Select the only item in the list: E8eedF.
Io' 'e 'ant a second input stream# so enter EAecycleF in the line belo' E8eedF.
Ie"t# clic& on Nin the E5utletF bo"# and you 'ill see that there are no candidate streams.
So enter the name E=i"edF.

- 2* -
Dhy is it EIot SoledF9 /ecause 'e hae not described the other input stream
EAecycleF. $o do this# go to the EDor&sheetF tab because it 'ill allo' access to all the
streams connected to the mi"er.
1t this point 'e need an estimate of the recycle stream. Ho' this is arried at 'ill be
different for eery process# but usually re!uires &no'ing 'hat the conersion per pass in
the reactor 'ill be. 8or this e"ercise# use the follo'ing alues.
$emperature O pressure ? same as stream E8eedF
=olar 8lo' ? 2,, &gmole%h
Ie"t# go to the E(ompositionF 'indo' and enter the follo'ing alues.
- 2: -
Io' UniSim has sufficient data to do the necessary calculations. -et us loo& at the main
screen.
Since all of the streams are dar& blue# UniSim is happy.
- 2; -
$he ne"t step is to heat the mi"ed stream up to reactor temperature. In the palette# double
clic& on the heater icon.
Iote that the /asis-. properties pac&age 'ill be used here. If another one 'as aailable
'e could select it.
)nter the follo'ing data:
Iame ? E8eed HeaterF
Inlet ? E=i"edF (from dropdo'n list)
)nergy ? EHeater dutyF
5utlet ? E$o AeactorF (a ne' stream)
- 2+ -
Ie"t# go to the E0arametersF 'indo' and enter a alue for the pressure drop across the
heater. :, &0a is a reasonable number.
1n alternatie 'ould be to leae this empty and specify the pressure of the output stream
in the ne"t step. /ut it ma&es more sense to specify the pressure drop rather than the
output pressure. (onsider 'hat 'ould happen if the upstream pressure changed.
4o not enter a alue for duty. In a moment 'e are going to specify the output
temperature# and it is not possible to specify both. 5f course# there are situations in 'hich
'e might 'ant to define duty rather than temperature# but this is not one of them.
If you hae used other process simulators such as 0A5%II you 'ill remember that the
output temperature specification is treated as a parameter of the heat e"changer rather
than the output stream. You 'ill need to reorient your thin&ing.
- 2< -
$o specify the temperature 'e need to go to the EDor&sheetF tab. )nter a alue of 2,,7(.
Io' eerything can be calculated.
- 2K -
/efore proceeding to tac&le the reactor# it 'ould be good idea to sae the case again. 4o
ESaeF follo'ed by ESae 1s tutor,*F.
Reaction Section
- 6, -
De 'ill model the reactor as a stirred essel 'ith sufficient cooling to maintain the
output temperature at the same alue as the input. 5n the palette# double clic& on the
(S$A icon.
8ill in the data# using the same techni!ues that 'ere used in preious units. 1lthough 'e
'ill not hae any li!uid leaing the reactor# UnSim re!uires a li!uid stream 3ust in case
some is generated. $he flo' rate 'ill be set to Cero.
- 6. -
Iote the message about the need for reaction data. 2o to the EAeactionsF tab and select
the reaction set from the dropdo'n list (only one option e"ists).
Io' 'e need to specify the siCe of the reactor. 1t this stage 'e do not &no' ho' big the
reactor should be. $he best thing to do is set the dimensions unreasonably large. $his 'ill
- 62 -
drie the reaction close to the e!uilibrium point# and proide a stable enironment 'hen
'e come to deal 'ith the recycle. 2o to the EAatingF tab# and enter numbers li&e the
follo'ing.
De also hae the ability to specify a pressure drop across the reactor. $hat is done in the
E0arametersF section of the E4esignF tab.
- 66 -
Io' it is complaining that it does not hae enough data to calculate duty (the cooling
rate). De hae t'o choices:
Specify a cooling rate in &P%h (the E4utyF field in the 'indo' sho'n aboe)
Specify the outlet temperature ('e 'ill do this in the EDor&sheetF tab)
- 6* -
Io' it is happy. If you 'ant to see 'hat is happening in the reactor# loo& at the
E(ompositionF section. $he output has more methanol and less (5
2
O H
2
than the feed.
Recycle System
$he ne"t step is to condense the methanol and 'ater# and return the gases as recycle.
8irst# 'e need a cooler to condense the li!uid by cooling it to *,7(. De 'ill do this in the
same 'ay that 'e defined the feed heater# but 'ill select a cooler instead of a
heater from the palette. 1ssume a pressure drop of :, &0a.
- 6: -
- 6; -

Iote that the output# E(ondensed =i"tureF# is t'o phase. De need a separator to isolate
the t'o phases. Select a ESeparatorF from the palette and attach the streams as
sho'n.
- 6+ -
Iote that this is all that UniSim 'ants from us. /ut 'e hae made an unconscious
decision to accept a default. -oo& in the EparametersF section.
$his is telling us that both outputs are at the same pressure as the input. $here are other
parameters 'e could specify# but they are not necessary for our case.
/efore 'e recycle the apour# 'e need to split off a small purge stream to preent
buildup of noncondensible gases in the loop. In the palette# the deice 'e need is a E$eeF.
2ie it the follo'ing parameters.
- 6< -
$his may loo& li&e it is putting nothing into the purge stream. $he problem is too fe'
digits in the display. -oo& in the EDor&sheetF tab.
- 6K -
$his 'ould be a good point to see 'hat the process loo&s li&e. It should loo& something
li&e this.
If the e!uipment has got pushed around on the screen do a E084?@1uto 0osition 1llF
operation and the units 'ill be arranged in a logical order.
$he stream EAecycledF is at a lo'er pressure than the feed# so 'e need a compressor
to get it bac& to the mi"er.
- *, -
In the E0arametersF section# accept the default efficiency of +:Q.
$hen# set the output pressure in the EDor&sheetF tab.
- *. -
De are about to close the recycle loop. In doing this# it is not uncommon to end up 'ith
an un'or&able model. 1l'ays sae the case before closing the loop. 4o a E8ile SaeF
and E8ile Sae 1s tutor,:F.
Set up a EAecycleF unit from the palette and specify the t'o connections.
$he 084 should no' loo& something li&e this:
- *2 -
/efore 'e finish 'ith the reaction section there is one more 3ob to do. )arlier 'e had set
aside the !uestion of ho' big the reactor should be and 3ust set it ery big. 8or all
practical purposes the stream leaing the reactor is at the e!uilibrium concentration.
$here is a tradeoff here bet'een reactor siCe and the siCe of the rest of the e!uipment in
the recycle loop. If 'e reduce the siCe of the reactor# the conersion per pass 'ill fall.
$his 'ill re!uire a higher recycle rate# leading to larger heat e"changers compressor
separator etc. Some'here there 'ill be an minimum e!uipment cost (operating costs 'ill
also be an issue).
8or this e"ercise 'e 'ill say that preious designs hae sho'n the optimum is about K,Q
of the e!uilibrium conersion# and the height of the reactor should be t'ice the diameter.
-oo& at the EAesultsF section of the EAeactionsF tab.
De should aim for a conersion of *;.,6 L .K ? *..*
- *6 -
Jary the reactor siCe and see 'hat happens to the conersion.
4iameter Height (onersion
. 2 *2.;:
.K ..< *..:*
.<K ..+< *..*,
De 'ill end up 'ith something li&e this.
- ** -
4o a E8ile SaeF and E8ile Sae 1s tutor,;F.
- *: -
Product Separation Section
1ll of the e!uipment simulations 'e hae done so far hae been ones in 'hich any
reasonable set of parameters 'ould lead to a calculation that 'or&s. 8or e"ample# as long
as 'e did not do anything ridiculous li&e specify a temperature of -:,,(# 'e 'ould get a
alid heat e"changer design. $he ans'er might not be 'hat 'e 'anted# but 'e 'ould get
an ans'er.
Dith a distillation column 'e might not een get an ans'er. You may recall from earlier
courses in distillation that some separations are only possible 'ith a number of
e!uilibrium stages aboe a certain alue# or a reflu" ratio aboe a certain alue. Some
configurations 3ust do not 'or&.
In most cases 'e 'ant to design a column to meet certain concentration specifications. In
our case they are:
K+Q of the methanol entering the column leaes in the product
$he methanol product contains .Q (by mass) 'ater
$here are t'o specifications because a simple column 'ith t'o products and feed#
pressure# number of stages# location of feed tray specified has t'o degrees of freedom.
In some cases 'e could go directly to a model 'ith these specs. In general# it is safer to
start by creating a case that 'or&s (een though it is not 'hat 'e 'ant) and then migrate
to the case 'e 'ant. $he configuration most li&ely to 'or& is specifying:
$he flo' from either the top or the bottom of the column
$he reflu" ratio
De 'ill do it this 'ay.
Dith UniSim you must hae certain information lined up before you start putting the
column together. You 'ill be ta&en through a series of 'indo's that you must complete.
You cannot brea& out and come bac& later. De 'ill need the follo'ing information:
Iame of feed stream (E-i!uidF)
Iumber of stages (this separation is easy so try .,)
-ocation of feed ('e hae nothing to base this on so put it in the middle B stage :)
0ressure in the condenser (use .,,, &0a)
0ressure in the reboiler (use .,.: &0a)
Dill 'e ta&e the product off as a li!uid or apour (apour# do you thin& it 'ould
be a good idea to attempt to condense hydrogen9)
1 starting alue for reflu" ratio (6 is suggested)
1 starting alue for the distillate rate (.K.+2K &gmoles%hr - see the chart belo')
- *; -
4ouble clic&ing on the E4istillation (olumnF icon on the palette 'ill bring
up this:
- *+ -
)nter the follo'ing data:
De 'ant to ta&e all of the top product as a apour. 1 partial condenser allo's both li!uid
and apour. De hae specified t'o streams# but 'ill set the flo' of E4ummyF to Cero at
a later step.
Iote that the addition of this third output stream increases the degrees of freedom by ..
1t the same time a constraint (flo' ? ,) is added and the net effect is that 'e still need to
proide t'o specifications. 1nother 'ay of loo&ing at it is that a stream 'ith Cero flo'
does not really e"ist.
Ie"t# specify the pressure profile.
- *< -
In case 'ith comple" apour-li!uid e!uilibrium relationships# estimating the temperature
profile can help the program to conerge on the right ans'er. 5ur case does not re!uire
this.
- *K -
Iote that 'e hae set E4ummyF to Cero. De no' hae all the parameters specified. (lic&
on E4oneF.
- :, -
(lic& on run and get a alid case (but not the right case).
$o complete the 3ob# go to the ESpecsF section.
If you e"amine the E(olumn SpecificationsF you 'ill see that the first three are E1ctieF
(the calculations force them to be met) and the last t'o are alternaties that are EInactieF
- :. -
(ignored). De 'ill leae E4istillate AateF alone (the dummy li!uid stream) 'hile 'e
replace EAeflu" AatioF and E5hd Jap AateF by our t'o composition specs.
8irst# create the t'o ne' specs but leae them inactie. Under E(olumn SpecificationsF
clic& on the E1ddF button and start on the methanol recoery spec.
Select E(omponent AecoeryF and clic& on E1dd Spec(s)RF.
EAecoeryF means the fraction of the component in the feed that goes to the specified
stream.
- :2 -
Io' do the methanol concentration in 'ater spec.
Io' change the spec alue to 'hateer it happens to be at the present time# and ma&e the
spec actie.
- :6 -
Iote that E4egrees of 8reedomF has changed from Cero (the correct number of specs are
actie) to -. (too many are actie). 4eactiate E5hd Jap AateF and loo& at E(omp
AecoeryF.
Io' do the same 'ith the other spec.
- :* -
De no' hae t'o too many specs. 4eactiate EAeflu" AatioF and E5hd Jap AateF and
the case 'ill run.
De hae s'itched to a ne' set of specs# but the case has not really changed. 1ll of the
concentrations# temperatures etc. are the same as before. $his ma&es the transition more
or less foolproof.
$he final step is to change the alues of the specs to 'hat 'e really 'ant.
- :: -
$he 084 should no' loo& li&e this:
- :; -
Io' do a E8ile SaeF and E8ile Sae 1s tutor,+F.
/efore leaing the column simulation# let us reie' the steps 'e 'ent through to get the
case 'e 'anted.
.. )stimate the distillate rate from the feed composition and a &no'ledge of 'hich
components are to go out the top.
2. 0ic& a starting alue for reflu" ratio. 5ther programs use a default of 6# and that
'or&s most of the time.
6. /uild a model 'ith distillate rate and reflu" ratio specs.
*. 2et this model to conerge. It may be necessary to change parameters such as
number of stages# feed location# reflu" ratio etc.
:. /uild the specs that you really 'ant and set their alues to those in the 'or&ing
case.
;. 1ctiate these ne' specs and deactiate the distillate%reflu" ones. $he model
should conerge.
+. (hange the alues of the ne' specs to match 'hat you 'ant in the column. If
there is a large change# you may 'ant to do it in stages.
.
Product !inishing
$he final step in the process is to condense the methanol product and prepare it for
storage. /efore 'e start on the condenser# 'e need a source of cooling 'ater. It 'ill be
ta&en from a storage tan& and pumped to ;,, &0a. Initially# 'e 'ill set the flo' rate ery
high (.,#,,, &g%hr) to ensure that 'e hae enough for the heat e"changer. 4uring the
heat e"changer design 'e 'ill reduce this to a reasonable alue.
Set up stream EDater SourceF in the same 'ay that 'e created E8eedF.
- :+ -
- :< -
Io' add a pump.
1s 'ith the compressor# 'e set the output pressure in the output stream. Use a alue of
;,, &0a.
Io' 'e can start on the heat e"changer.
UniSim proides three leels of detail for heat e"changer design:
- :K -
.. $he simple method# using a EHeaterF or E(oolerF 3ust does the heat balance
necessary to ta&e a stream to a specified temperature.
2. $he intermediate method deals 'ith transfer of heat bet'een t'o streams. 1 alue
of U1 (oerall heat transfer coefficient L heat transfer area) is specified and
UniSim calculates the t'o output temperatures. If an estimate of U is aailable#
the area can be calculated and used for a crude cost estimate. $he recommended
procedure is E)"changer 4esign (Deighted)F.
6. $he most rigorous method is ESteady State AatingF. Dith this procedure the
mechanical design is specified (tube number%dimensions%spacing# shell number
and configuration# etc.)# UniSim estimates U and calculates the t'o output
temperatures. De 'ill defer the use of this method until 'e hae learned more
about heat e"changer design.
-et us design the condenser by method 2# using the EHeat )"changerF unit.
$here are t'o re!uirements for the design. $he process stream should e"it at *,7(# and
the cooling 'ater at *:7(.
Start by specifying the follo'ing data. Iote that the process stream is going through the
tubes and the cooling 'ater through the shell.
- ;, -
In the E0arametersF section select the E)"changer 4esign (Deighted)F method and
supply pressure drop estimates. $hen ad3ust the U1 alue so that the stream E8inal
0roductF has a temperature near the target of *,7(# as seen in the Dor&shop tab.
- ;. -
De are not finished because the cooling 'ater e"it temperature is too lo' because of the
high flo' rate. UniSim has a unit called an ad3ust. It is something li&e a process
controller in a plant# but it manipulates the model# not the process. It tells us nothing
about the dynamics of the process. -i&e a controller# it changes the alue of one
parameter in order to bring another parameter to a specified alue.
8irst# under E1d3usted JariableF clic& on ESelect JarF and specify that 'e 'ant it to
manipulate the cooling 'ater flo'.
- ;2 -
$hen# under E$arget JariableF clic& on ESelect JarF and specify that 'e 'ant to control
the outlet 'ater temperature.
$he alue 'e 'ant it to settle out at is *:7(.
- ;6 -
Ignore the 'arning about EUn&no'n =a"imumF. If there is a problem in conerging#
changing the alues in the E0arametersF tab may help.
$his message is common the first time the model is run. (lic& on EYesF.
Io' loo& at the condenser 'or&sheet tab and erify that both outlet temperatures are
correct.
- ;* -
$o finish the design there is one more piece of e!uipment to add B a tan& to
store the product.
(onnect the input stream E8inal 0roductF and supply names for li!uid and apour
outputs.
- ;: -
1lso# you can specify the pressure in the tan&.
Iote that there is some apour (mostly (5
2
and hydrogen) that 'ill be ented from the
tan&. E8inal 0roductF 'as a t'o phase mi"ture.
- ;; -
$he final model should loo& li&e this.
Io' do a E8ile SaeF and E8ile Sae 1s tutor,<F.
Modeling Tools
Haing completed the process model 'e 'ill no' ta&e a loo& at some of the facilities in
UniSim that allo' us to generate reports# do additional design tas&s# and help 'ith model
deelopment.
Reports
- ;+ -
It is possible to generate# ie'# and print reports on the 'hole model or specific pieces of
e!uipment. -et us get an oerall ie' of the streams in the model. $he E$oolsF menu has
a EAeportsF item.
(lic& on E(reateF to generate a ne' report.
- ;< -
$his is the result of E(reateF# EInsert 4atasheetF. 1 report on the 'hole model is selected.
$o ie' the report clic& on E0reie'F. Here are some samples of 'hat is in the report.
- ;K -
If you clic& on E0rintF you can print a copy.
-ater# after modifying the model# you can come bac& through the E$ools?@AeportF route
and print an updated copy. You do not hae to redefine the report content at that time.
Column Tray Design
UniSim has a utility to do a mechanical design of distillation columns# both trayed and
pac&ed. 2o to the E$oolsF menu and select EUtilitiesF.
- +, -
De 'ant the E$ray SiCingF utility.
- +. -
$o specify 'hich section of 'hich column to do# clic& on ESelect $SF# and select the
column to be siCed.
Io' clic& on E1uto SectionF and let UniSim decide 'hat needs to be done.
- +2 -
Here 'e hae selected ale trays. In this utility there are a lot of parameters that can be
specified# but 'e 'ill 3ust accept the defaults.

- +6 -
(lic& on E(omplete 1utoSectionF.
Iote that UniSim has decided to do the design in t'o parts. ESectionS.F is the trays
aboe the feed tray# ESectionS2F is belo'. /ecause the flo' rates are different in the t'o
sections# one part can be made smaller than the other# if desired.
-oo& at the EAesultsF section of the E0erformanceF tab.
- +* -
$his sho's the recommended design. Iote that the lo'er section could be made 'ith a
smaller diameter than the top. /uilding a top-heay column is not a good idea# although
the reerse is sometimes done. $he best thing to do here is to build the 'hole column at
the larger diameter. (.;,K; m ? 2 feet).
Iote that the dimensions# although displayed in metres# are actually selected from
standard siCes in feet.
Useful Techni#ues in Developing a Model
Here are a couple of techni!ues that can ma&e it easier to deelop models.
Suppose 'e are told that# oer the lifetime of the catalyst in the reactor# its actiity can
fall as much as 2,Q. 1nd 'e are as&ed the !uestion EDill this affect the concentration of
the feed to the distillation columnF. It 'ould be helpful if 'e could 'or& on the reactor
and the column feed stream at the same time so that 'e 'ould not hae to flip bac& and
forth bet'een them.
4ouble clic& on the reactor icon and moe to the ESiCingF section of the EAatingF tab.
4ouble clic& on the stream E-i!uidF and moe to the E(ompositionF section. Select
E=ole 8ractionF as the basis.
- +: -
If you minimiCe the 084 'indo'# you should get something li&e this.
-oo&ing at the dimensions of the reactor# something seems 'rong. $he height%diameter
ratio is T 'hen it should be 2. 1lthough it does not affect the reaction calculations (only
the olume is significant)# 'e should fi" this before proceeding.
De 'ant to change t'o numbers# but 'e do not 'ant UniSim to recalculate the model
until both numbers are changed. $he intermediate case 'ith one number changed 'ould
send the model off into une"plored territory and might not conerge.
1t the top of the main 'indo' you 'ill find this: $he left (green) light
indicates that calculation is actiated. (lic& on the right (red) light to deactiate
calculation.
Io' change the numbers. 1 little arithmetic sho's that 'e 'ant 4? ...2.# H ? 2.2*2 to
get the same olume. (hange both of these and then clic& on the green light.
HII$: In 'or&ing 'ith UniSim and the program seems to seiCe up# chec& to ma&e sure
that the red light is not on.
1t this point 'e should hae something li&e this:
- +; -
Io'# reduce the height 2,Q (to ..+K*) to simulate the effect of catalyst decay.
Here 'e hae our ans'er. $here is no significant change in composition.
(hange the height bac& to its original alue# close the t'o 'indo's# and restore the 084
'indo'.
De are no' finished 'ith the deelopment of the model. Sae the file (it should be
Etutor,<F) and e"it UniSim.
- ++ -
$ppendi%
UniSim Customi&ation
In $ools?@0references?@Aesources?@(olours
Set E084 /ac&groundF to 'hite
Set E084 -abel $e"tF to blac&
Set E084 1nnotationF to blac&
In $ools?@0references?@Simulation
Unchec& E(onfirm /efore 1dding if 1ctie (orrelations are 0resentF
Unchec& E)nable (ross Hairs on 084F