Part III

Dynamics and Control

1
using Aspen Dynamics
 CHAPTER

Dynamics and Control of

Flow-driven Processes

51
.
INTRODUCTION

Dynamic -imulation of a chemical process greatly helps to understand the transient

behaviour Aspen Dynamics , which is tightly integrated with Aspen Plus is widely
,

used for process design and control. This powerful simulator can automatically initialize
the dynamic simulation using the steady state results of the Aspen Plus simulation .

Interestingly, when the file containing the flowsheet is opened in Aspen Dynamics the ,

default control structures are already installed on some loops Usually, level, pressure
.

and temperature controllers are included where appropriate However these default
,

control schemes can be modified or even replaced with other suitable control loops
available in Aspen Dynamic- package Note that there is a scope to include some
additional controllers for the used process Moreover this simulation tool provides a
,

graphical environment to show the process response.

To convert a steady state simulation into a dynamic simulation there are several
,

items that should be taken care of For example the size of all equipments must be
,

specified and the control structures must be devised For steady state simulation using
Aspen Plus the size of the equipment is not needed, except for reactors. On the other
,

hand ,
for dynamic simulation using Aspen Dynamics, the inventories of material
contained in all the piece* of equipment affect the dynamic response Therefore, the
.

physical dimensions of all process units must be known.

When the steady state Aspen Plus simulation is exported into Aspen Dynamics, we
need to choose either simpler flow-driven dynamic simulation or more rigorous pressure-
driven dynamic simulation Pres ure-driven simulations include pumps and compressors
-
.

where needed to provide the required pressure drop for material flow Control valves
must be installed where needed and their pressure drops selected For flow-driven
,

simulations , however, no such arrangements are required.

229
230 PROCESS SIMULATION AND CONTROL USING ASPEN

In the present chapter, we wish to study the dynamics and control of the low-

f
driven processes. For this intention, we choose a reactor (RCSTR) as well as a distillation
column (RadFrac) example rom the model library of Aspen simulator.

f
52 .
DYNAMICS AND CONTROL OF A CONTINUOUS STIRRED TANK
REACTOR (CSTR)
Problem statement

Ethyl acetate is produced in an esteriication reaction between acetic acid and ethyl alcohol.

f
acetic acid + ethyl alcohol ethyl acetate + water
A feed mixture, consisting of 52.5 mole% acetic acid, 45 mole% ethyl alcohol and
2 5 mole% water, enters the RCSTR model with a low rate of 400 kmol/hr at 750C and
.

f
1 .
1 atm. The reactor, as shown in Figure 5.1, operates at 70oC and 1 atm.

FIGURE 5.1 A flowsheet of a CSTR

Both the reactions are first-order with respect to each of the reactants {i.e., overall
second-order). For these liquid-phase reactions, the kinetic data for the Arrhenius law
are given below:
Forward reaction: A = 2.0 x 108 m3/kmol s
S = 6.0 x 107 J/kmol
Reverse reaction: k = 5.0 x 107 m3/kmol . s
£ = 6.0 x 107 J/kmol

Composition basis = Molarity

Here, k is the pre-exponential factor and E represents the activation energy. The reactor
geometry data are reported below.
Vessel type: vertical
Head type: lat
f

Diameter: 0,45711 m
Volume: 0.15 m3

(a) Simulate the reactor model using the SYSOP0 thermodynamic model to compute
the product compositions.

Copyrighied malerial
 )YNAMICS AND CONTROL OF PI.OW-DHrVKN PROCKSSKS 231

(b) Report the default controllers tuning parameters and control actions used, and
constraints imposed on variables.
(c) Investigate the servo performance of the default liquid level and temperature
control algorithms and discuss the effect of loop interaction.
(d) Show the regulatory behaviour of both the controllers in presence of disturbance
in feed temperature.

Simulation approach
(a) To open the Aspen Plus Startup dialog box. click the desktop Start button, then
point to Programs, AspenTech, Aspen Engineering Suite, Aspen Plus Version
'
and then click the Aspen Plus User Interface. Let s select the option with
Template and then click OK (see Figure 5.2).

aWHl 1 I M*l oKlfcl MwiH -| I I I -I J

L .H ! -i I i

o-tt rt m mots; am -

bHft mm 1 ; :_ tm

FIGURE 5.2

As the next window appears (see Figure 5.3), it is appropriate to select General
with Metric Units and hit OK button.
Here we use the simulation engine at 'Local PC When the Connect to Engine dialog
pops up (see Figure 5.4), press OK. Note that this step is specific to the installation.

Creating flowsheet
The process low diagram, shown in Figure 5.5, includes a reactor, namely RCSTR,
f

with an incoming FEED stream and an outgoing PRODUCT stream.

Copyrlghiod material
232 PROCESS SIMULATION AND CONTROL USING ASPEN

Nil Zl , _
iJ_l J J
I I 1 1

I Rk Separato'-r wth Engi<h Uhb

I Air 5«paraK> ntfi Meirc 'Jnti .
PelroleuT v-rth r-lefr
.
3*3pe IPE S!fe»T Property kit Fhafnieceul cals hI Caenerai SmJalion vrth
rfBark Smutatcn L Fhamaceuticala Mettle iJnts
C ba'>gyhr tanoMr,
2Chemical3w<h English ijnts H Polymefs Mh Engis .

MMKca(/W cum/nr
i3 Chemicals w<h Heine Unrts Polymer wrO". Mibfe
j BMMMM »**> BfSlW> UnJs f. P,TT>fr«*atu / /.-eh Property Method Nooe
i S cl VM **h MeMc Lints £jj P>n)fr«(arur3y /. (h
FtewbwfoTB xi Mole
i jGas Process iMi English UrAa So!'*;-rfh EngWi I
yScWswlh fAetrcU. Strearr report
,
*n*rai wth English IJhta
G La Specia*/ QmiOk ccmpoabon Hole flow

"
£2 Specialty ChemV;a*s
d m alkjgy EntfishUnfts
dumelallurgy Mih Metnc Unfis Run Type

1 2] [Rwsheel 3

111
Fot Help prws Fl
,

Start I r- j j]Ch lei-5-Migiiio(IW | aw 2 - MCTtcB W . | - Aijob«/tarotei Pro(ett . | Aepoi Pka

FIGURE 5.3

Connect to Engine

Server type: Local PC

User Info

Node name:

User name:

PassiAiord:

Working directory:

n Save as Default Connection

OK Exit Help

FIGURE 5.4
 PVNAMICS AND CONTROL OF FLO\V DRl\-EN PROCESSES 233

M rn,-» 1
.
« I I ---

-
- -;

e i i i ysa
'

.*» MB Mtet acsT» ie«o-

-
-
.
i

FIGURE 5.5

Configuring settings
Hitting Next button, we get Global sheet of the Specifications form under Setup folder
in the left pane of the Data Browser window. Enter the Title of the present problem-
Dynamic Simulation of a CSTR'. change the Input mode from 'Steady-State' to "Dynamic"
"

and leave the remaining items at their defaults. The window looks like Figure 5.6.
? ).!
,

., 81 V
* a

jLr
~

i
3 I 1 !-! -'in: q >!i.

4
-

ia

Km "ttm "CT on

js- b

FIGURE 5.6
234 PROCESS SIMULATION AND CONTROL USING ASPEN

In the next window, as shown in Figure 5.7, the Aspen Plus accounting information
required at some installations are provided.

Hi _l 3 J J
_
j r _
I I I [»" -
I M I M

/GbUI] /OWbMoP -/AtcountIngj Di mn'ci |

O r-.i* - - 1. PVil i . tr. I . " .

Accouxlrurbar

© Mm >-
O R»MOttor# PrcrKtraw fr-OUH WISH
i -

lnoMtCflinpi<»
[V Hm>iA(«m j $etur*» j --v . I C(*ftm RMEton | PreiMeChange | Marv ai j Sokb | UwMoiw |

'

STREAMS RSttK nvdd REtwl RGbb. RCSTR RPKi Rewch

For **> cr-u Ft C \ a FoMan'Aoen »us 11 NOM RwpM b ot ntr --

FIGURE 5.7

In the subsequent step (see Figure 5.8), select Stream sheet with opening the Report
Options form under Setup folder and include Mole fraction item.

.
tfe E* V1e» 0*4 '<xk Rn ftM Ufmy WMoh -.

1 hi .JSlal _J
i r-1 i i nr i -m \ |a! «|».|
-

3 »| j ,| H.|

P Gtf«<a(e»ii»xlsrJ*Mn-feKirt
tam'o br rcUM r lDcmti itpvt
iWhua Fi«cMntt*m
P Malt P Mole TIT fGEfTM
0 RVMlOIMm
r Urn* r Mm.

.
_
l -r. - - -.

.
Jj

-
CH

fffaM ftt PQbt. RCSTR HPhg ftfiad.

- s -

FIGURE 5.8
 DVN VMICS AND CONTROL OF FLOW-DRIVEN PROCESSES 235

Specifying components
In the Data Browser window, choose Components /Specifications to obtain the component
input form. Filling out the table with the components (acetic acid, ethanol, ethyl acetate and
wateri involved in the present reaction system, the screen looks like Figure 5.9.

MM - . -

E-sC

aun
i 1

I
I*

cmt : ffstnc ff/wa

8 § O w&mh ai:s Tft
O
Re3''>
, ,,.

.
* - -« MUM

FIGURE 5.9

Specifying property method

Choosing Properties /Specifications in the column at the left side, one obtains the
property input form. As shown in Figure 5.10, we use the SYSOPO base property method.
i.|jrrxl

'
-

KW -U MSI *) nKI*M-aH 1 3)

9 «-.*«

. 2

HflUwj Wfate ffft*

I 8 y Ki)
t COM, hc IP

FIGURE 5.10
236 PROCESS SIMULATION AND CONTROL USING ASPEN

Specifying stream information

Use the Data Browser menu tree to navigate to the Streams/FEED I Input I Specifications
sheet. Specifying the FEED stream by its temperature, pressure, flow rate and
composition, we have this window, shown in Figure 5.11.

oj -
u l-w
-

-
3 ****
* 1
-

- 3
3 (.-. j| 3
T 3
('- * z
h -

i= 3 -
.
3
pr (n -%\
I J

I "in

> > Mat l«

FIGURE 5.11

Specifying block information

In the list on the left, choose Blocks IRCSTRI Specifications to obtain the block input
form. It is illed with the given data as shown in Figure 5.12.
f

-
o -

-
# l-s-g-Q-m-g-
. j
'
T*4m ! . » Iff ..-*-. |.;a .l»-. |<j w> r

FIGURE 5.12
236 PROCESS SIMULATION AND CONTROL USING ASPEN

Specifying stream information

Use the Data Browser menu tree to navigate to the Streams/FEED I Input I Specifications
sheet. Specifying the FEED stream by its temperature, pressure, flow rate and
composition, we have this window, shown in Figure 5.11.

oj -
u l-w
-

-
3 ****
* 1
-

- 3
3 (.-. j| 3
T 3
('- * z
h -

i= 3 -
.
3
pr (n -%\
I J

I "in

> > Mat l«

FIGURE 5.11

Specifying block information

In the list on the left, choose Blocks IRCSTRI Specifications to obtain the block input
form. It is illed with the given data as shown in Figure 5.12.
f

-
o -

-
# l-s-g-Q-m-g-
. j
'
T*4m ! . » Iff ..-*-. |.;a .l»-. |<j w> r

FIGURE 5.12
 DYNAMICS AND CONTROL OF FLOW-DRIVEN PROCESSES 237

In the next step (Figure 5.13), select RCSTR I Dynamic! Vessel sheet under Blocks
folder and enter the reactor geometry data.
LMf ***** -**

JUIBI U _
51 nMi>H<H 3 I H .-J JJ -1 J
i r i i i fv i lei I Pi 3>M

. i

STftfAMS WSw ffr-ipd Wgu* HGfcbi RCSIfl RPfc i BBwt*

C gUdms iw Pk* 111 MJU %

FIGURE 5.13

The forward reaction as well as the backward reaction is represented with their
stoichiometric coefficients and exponents in two sheets shown in Figures 5.14(a) and (b). ,

J l-i I IT .1 .IBI I IBI .

J
"
ti »i nl i « !

EAC

_
l ***

gff Vftrtc:

ll

-
CH-
1-0 LIZ U
.

FIGURE 5.14(a)
238 PROCESS SIMULATION AND CONTROL USING ASPEN

Mai _
1J Mffli MJ ahlahNN 1 Jl .! I"i _J El M M
i r i i i nr i m I issi
71 f. .1 I 1
. I .<lr:-
-

7i »i nl «i uf\

UMFAC Grows;
Reaction No fyT
Readmit Pto<k«<t

Compgnert EKpononi Cofr wrKf*

'
.
'
v- i- -
.
rv

1 V
-

J 0"

Prop Sett

3 -"EEC
j PRODUCT
_

3 Becks
.
J/j 3CSTR

_
j Ow stry
-
Resctcr:
w R-l
J t eti e

ll -

STREAMS flSloe RYidd .

REqul RGbtu FiCSTR March

FIGURE 5.14(b)

The power law data for both the reactions provided in the problem statement are
entered in the two Kinetic sheets shown in Figures 5.15(a) and (b)

, nt Edt '.' en Can Toot Pir. fU Lfeery Wrxto-. He*

ar -i MM <<||AI 3 »l o\s
~

/Sto-Jwnwy *Kinebc| : - |
'

; | He«yG)mt>s
UfUF Grauo*
}l) AA- EAL-) EAC . VATER d
fiwc Cte, | . -J
-
J/i Zcrvij&

| PIBpctfN r«*c(actor.)jr/To/'e-(E »'/I'l/ro)

;
"

;
i "
k. 200000000
[ W
.
.
. a P«* en i E::
_
l 0** To-
I
(qi«w.

> ) FEED
£ PRODUCT
.
1 SCSTR
y Reacscoi
_
J Ow-wy-
-

nn miiii

SIRtAHS

J- 90

FIGURE 5.15(a)
 DYNAMICS AND CONTROL OF FLOW-DRIVEN PROCESSES 239

-.- 1 3

I I

r
.
«

f " If Till I | MME->w. I C3k-s RMcMn | Ck . | I

HD->

st e
-'

py.>- pew «ac cstr

FIGURE 5.15(b)

The status indicator in the above window reveals by the message Required Input
Complete that no more input specifications are required to run the simulation .

Running steady state simulation

As we click on Next button to continue the simulation the Required Input Complete ,

dialog box appears Hitting OK on the message, we are displayed the Control Panel
.

where the simulation messages during the run are recorded (see Figure 5.16).
- ?!

- MM

iT*0»r. Kmm . Pif

i i Q
W ji
D
| -ii?»i u -J- I .ic- .i-a.-i. || .n, am, .. .. -_ |. «ij

FIGURE 5.16
240 PROCESS SIMULATION AND CONTROL USING ASPEN

Viewing steady state results

In the next, select Solver Settings, choose Results Summary /Streams in the list on the
left and finally get the steady state results as shown in Figure 5 17. .

.md.m.x. 1 - |Ftei.*« Sumn»r Sreano - DUit Browser)

'
.

f Fte &» »ew D*> Todj Run Lbraiy WMow Htjp

Setup
1 I I 1
Omponeras

Streams
3 foumi r
" ~

as Bocks
Displw 3 Stie«mTatile|
Reacts

3fl Ccr.v«gefK«
'

1 Conv Options
41 d1 d
£0 Conv Options Volume Flow cumyhi 24.497 24 01S
O S«up Enthalpy MMkcal/hi .
35035 35 951
o 0M0 Bsslc
o OMOAdv Mole Flow km*hi
o lSSQP Base M 2ia0CB 87110
0 LSSQP Ad*
EAl 180.000 57110
Tear
EAC 122,890
1 Convei ence
Conv Ortef WATER mooo 132.890
Q| Sequence id* fan
Rowsheetng Ctouons
AA 0 525 0 218
Model Analvsis Too .

EO CcHguration EAL 0 450

.
0 143
.

P«Rit5 Summaiy EAC 0 307

.

Q Run Status WATER 0 025 0 332

Q Streams . .

Convetgence

Jdll
Results Avaisse

Mstets/SpittM! | Seoatatois | Heat Enchangeis \ Cokmra Hoactoti | Pressuie Chatlseis ] Minipulatois ) Sofds | Usei Models |

STREAMS ' RSIok: RYieM

.
1 . S -U-E-U-
REqui RGibbs RCSTR RPlug RBalch i
RofHeb.ptessFl C;\ .gFolde<s\A9penPlu3l1 1 : NUM

| j£]Oaiaer5-McmsollW..| 4]Cha(»er2-Wt!roseilW...|| Aspen flu. - SIimSI.. Adohe tolal Pntesai |« iS?t}); 1545

FIGURE 5.17

(b) Exporting dynamic simulation: In the subsequent stage (see Figure 5.18),
we wish to carry out the simulation dynamically. Accordingly, at this moment,
we have to follow the sequential steps noted below:
Click on Export from the pulldown File menu or simply press Ctrl+E on the
keyboard.
Open the Drive and then Folder where we want to save the work as a file.
Type 'ChS S .
RCSTR' in the File name field.
Choose 'Flow Driven Dyn Simulation (*.dynf & *dyn.appdf)' from the options
available in the Save as type box.
Finally, hit Save button.
Also, save the work done as a backup file (e.g., Ch5_5.2_RCSTR.bkp). We may use
the same folder within which the exported dynamic simulation file is saved. Originally
many files are saved along with the backup or dynamic file. Anyway, we are now ready
to run Aspen Dynamics and we may quit Aspen Plus.
 DYNAMICS AND CONTROL OF FLOW-DRIVEN PROCESSES 241

3 ±ljili<JF; »l jlalH
. I I I

.
1 1
, -

9 Ljii=v.
Tt*-

CJ Omwrngnt **, MW
Can***

I Pc««-««r«; Ctt '-* M

| M«W .V*** Tn»» -

EAL

WATER
SrniatTi r irTi i -o/r «Caned |

Jjj

8 . S -O'll-O-
cvi
| 0>»l»5-rfco...| *]am2-«*n,. \.*]Mll<m,-lic..\\ <»«,PIU.-C , '; teJu ft |'« igsj

FIGURE 5.18

Starting Aspen Dynamics

As we click the Start button
point to Programs, then AspenTech, then Aspen Engineering
,

Suite, then Aspen Dynamics Version and then click on Aspen Dynamics a blank dynamic ,

simulation window appears as shown in Figure 5.19 .

f*e Eat Tood ftowtfiMl Rr. VMoh Hcb

D H SQ W |Sl»dySl.lc j-J
3
r «-Jt c ioos j a, k it

It*- Lfcr- f,:*.»

.
Jaw

FIGURE 5.19
242 PROCESS SIMULATION AND CONTROL USING ASPEN

Opening existing simulation

To open the low-driven dynamic file, select Open from the File dropdown menu or
f
press Ctrl+O on the keyboard. In the Open dialog box, locate the drive, then folder and
finally the file 'Ch5 _
52 .
_
RCSTR' (see Figure 5.20).

ij is H *»Q IS

UIMl
r »-
-

SSmJatn
> Ft OinMITKt

l-llff. |.|9. HT» H

.
Chi 5i
_ _
CSTR !; Ch5 EJ BCSTB
i _ ,

£l Hi

5 *«» ICM .
U f
.
. ITS | CWn |

Opin
B i -r

FIGURE 5.20

As we press Open button, the process flowsheet consisting of the automatically

inserted level (LCI) and temperature (TC2) controllers appears (see Figure 5.21).

fte &* Urn Twh HswhMt «jn Wr4» rtft

-
Li

Tt MM [MM :"
u
'
i
ft- - - - .
-Ill

. .

. ..
..

Li I
m
lad *>.

in

FIGURE 5.21
 DYNAMICS AND CONTROL OF FLOW DRIven PROCESSES
-

243

Details of the two control loops, to be used finally, are given below .

Loop 1
Controller: LCI
Type of controller: proportional (P) only
Controlled variable: reactor liquid level
Manipulated variable: product flow rate
Controller action: direct

Loop 2
Controller: TC2
Type of Controller: proportional integral (PI)
Controlled variable: reactor temperature
Manipulated variable: heat duty (cooling operation)
Controller action: reverse

Note that the direct acting control system increases the output signal as the input
signal to the controller increases. On the other hand, as the input signal to the control
structure increases, the output signal from the controller must decrease for the case of
reverse acting control strategy. The direct acting control law has negative gain and
increase/increase (or decrease/decrease) term is commonly used to represent it For the .

reverse action, increase/decrease (or decrease/increase) term is used and controller gain
has positive sign.
The reactor flowsheet includes two (LCI and TC2) single-input/single-output (SISO)
control loops. Therefore we can say that this is a multi-input/multi-output (MIMO) or
,

simply a multivariable closed-loop system .

In Aspen terminology the process variable or controlled variable is denoted by PV,

,

the set point is represented by SP and the controller output or control variable or
manipulated variable is abbreviated by OP .

For the example CSTR system level and temperature controllers are automatically
,

implemented when the Aspen Dynamics simulation is created The default values for
.

SP , PV and OP are computed from the steady state simulation. To achieve better closed-
loop process response the Aspen-generated control structures can be modified or even
,

replaced by the suitable control schemes available in the control library of Aspen
software .
In addition, the default values for controller tuning parameters, such as gain,
integral time derivative time and so on, can also be changed.
,
,

Most of the control strategies are easily tuned by simply using heuristics. As
suggested by Luyben (2004) all liquid levels should use P-only controllers with a gain
,

of 2. All flow controllers should use a gain of 0.5 and an integral time of 0.3 minute
'
also enable filtering with a filter time of 0.1 minute). The author also mentioned that
the default values in Aspen Dynamics for most pressure controllers seem to work
reasonably well But temperature controllers often need some adjustments.
.

Viewing default values of variables

In Aspen Dynamics, the steady state values of process variable and controller output
ar displayed in a table At this stage the set point value, displayed in table, shown in
.

,
244 PROCKSS SIMULATION AND CONTROL USING ASPEN

Figure 5.22, is same with the value of process variable. To show the results table of
loop I, highlight the controller block LCI, press the ight mouse button, go to Forms

r
and then select Results.

'
-
i .mi m i -- w*

ffes

El &
m £

- (M M Has M ! ! i lull aw M-M

11 ..1 in

FIGURE 5.22

We can have the same information in a faceplate, shown in Figure 5.23, simply by
double-clicking on the block LCI. But as a difference, the units are not mentioned here
with the values of SP. PV and OP.

.
. ;i J
-
.

W ,1V- 3
>

--

IP
mar

urn

FIGURE 5.23

Gopyngt-
 DYNAMICS AND CONTROL OF FLOW DRIVEN PROCESSES -

245

Similarly, we have the results table, shown in Figure 5 24 for the temperature loop 2. .
,

E3B i«ra
-

r g <: |.v» J Ji i K >ft l|

-
frol-

r-O-

HS

SiBulaticn resd? tor soJuiion

8 equadons were not elininated because they had i
*
"
(jtal ol 109 equations (29 5'/,) verc sUaiMted
ilation has 36J variablea 2(,0 equations and 91

FIGURE 5.24

Modifying controller tuning properties

First we need to open the sheet that contains the controller tuning information. To do
so for the level controller, highlight the controller block LCI, press the right mouse
button, point to Forms and then select Configure (see Figure 5.25).

|iirjf;j>,l'l. fi S I PPQ UCI l-Q

r,.-:n .

rwutlolis'm* bo .tiHinatad be aus* they had residuals over J»-00?

'

oi IH equ»ii2rii (5 J weie elmir.sted

'

LMien has HI *ui»tUi ito equiitjdni ar.d 'ifie

O j ct
-

Ij JOgggj j j Jctf Voty P-Vet || l Aspen gh f* J - * ,stM

FIGURE 5.25
246 PROCESS SIMULATION AND CONTROL USING ASPEN"1

Alternatively, to obtain the Configure dialog box, first double-click on the controller
block LCI and then click on Configure symbol (yellow colour) in the faceplate as shown
in Figure 5.26.

ll(i*H«a IS {Dyn.mlc 3 mvo t: tin a 13

r tt .t ghF* J «i a 'a i' 3

C SsD-
U 1
.

3
«liBtn»ted becAutA t . hsd residuals ouer le-OUS

Istion has J63 vortabies HO equations and 969 non-zeros

Ml

FIGURE 5.26

By the similar way, we obtain the tuning data sheet, shown in Figure 5.27, for the
temperature controller TC2.

Bf H » Q <& If {Dynamic 3 »»« «« CB'JetiHDaia

o i

u
a» F f.A

(MMM |a FT-3
£3
r Dmo>

d5.

FIGURE 5.27
 DYNAMICS AND CONTROL OF FLOW-DRIVEN PROCESSES 247

Note that the default Operator set point value is the steady state value of the process
variable (PV). The reactor liquid level is the PV for loop 1 and reactor temperature for
loop 2. Bias signal is the output from the controller when the error (= SP-PV) is zero.
From the results tables shown earlier, it is obvious that the error is zero for both loops.
Therefore, Aspen Dynamics has set the value of OP as the bias value.
The proportional integral (PI) control methodology is automatically installed with
default values for the controller gain (= 10 %/%), integral time (= 60000 minutes) and
derivative time (= 0 minute) to monitor the reactor level. However, as mentioned
previously, the proportional-only controller with a gain of 2 is sufficient to effectively
control the liquid level. Remember that to make the integral action inactive, we can
use a very large value, for example 105 minutes (even the default value of 6 x 104
minutes may also be accepted), for the integral term. For loop 1, the controller action
should be Direct' as set by default (see Figure 5.28).
'

Baa
Ffc MM Took Af) Wltttm H*

|DynBfnic ]
"

U 7f 3 ii ft i*

BLOCKSfTCn Co mv

Op**to» ie* part. jO 314029 K

-

r : i

Bw |21083 6171
Gat

jiooooo d
L

m m ConOofc* «aon -
-
-

- -
-

ft fcj
Integrftting tro. 62 43 to 62 44
Integrating tr » 62 44 to 6? 45
Mograt ing fro* 62 45 to 62 46
Step 3074 Tiw 6 2504«+001 step size- 5 OOOOe-002 step facto 1 5000e+000 accepted

j
wm -.

1* Wggg [ ' Adotoc Vrote P.o<mc

FIGURE 5.28

In loop 2 , we prefer to employ the proportional integral controller to monitoriven.

the
reactor temperature. In data sheet, shown in Figure 5.29, the default values are g
The TC2 is truly a reverse acting controller. However, we may adjust the values of
controller tuning parameters (gain and integral time) during the closed-loop study if
the control performance is not satisfactory.
Modifying ranges for process variables and controller outputs
In the Configure dialog box, hit the Ranges tab and get Figure 5.30 for level control loop.
248 PROCESS SIMULATION AND CONTROL USING ASPENIM

iDynamlc ii m * 93 O fci rj a

r ft -jt G,.t| 3 iL Si fi 'i

5

luwng 1 R«v" 1 FHwng | 0H>« LCI

Ope>«ioi ;a>pan p 1

Mi
G«n h
tmegral dm --

w-lPRoSucrt-H .
,
.

['' . v- tme I

C.>**i Chwi
r Died

Revwte

IfWiataeVabei | Htto

0|>e"i Lbra So ve*

OpBoni

m
alftulation ready for solution 3
equations were not eliMinated because they had residuals ever le-OOS
k total ol 109 equations (29 S'-i) were eliminated
Simulation has 363 variables 260 equations and 968 non-zeros

Rea*- Rea V a Iocs

Oiaplm UChajtef S -MoMsflW j Mobe Pootm PrrfrnJ [ .
-

lAjpenOrn-c.

FIGURE 5.29

ftfts -n Dyn-wwca Ch5_5 2_HCSlH.dyrf

Be Wew Tao* Run Wndow Help

D B # Gi ©i ? jDynamic I) M «

Q SniMoh

a Bi.ocKsnrncorfKi

Piocott voiabie

"

Ranjsmawwm (l 8280M (m

o
TC3

Rang* m»«num |l2167 2343 |

P Cfetorang«

Ct>0>n btT»y
.us S- Ktr
Ootore

Simulation ready lor xoluiion 3

equations were not eliminated because the? had residuals o*t
i total of 109 equations (29 S'O were elminated
"

mulaiion has 36} rarimbles 260 equations and 989 nae~:ero«

M ty«m>ce«a
I UCh e'

FIGURE 5.30
 DYNAMICS AND CONTROL OF FLOW DRIVEN PROCESSES -

249

As shown in Figure 5.30, the default ranges for both the process and output variables
are too large (± 100% of the steady state values) It may be practical to consider the .

following constraints.
Process variable
Range minimum: 0.6855 m (25% subtracted from steady state value of PV)
Range maximum: 1.1425 m (25% added with steady state value of PV)
Output
Range minimum: 15812.7 kg/hr (25% subtracted from steady state value of OP)
Range maximum: 26354.5 kg/hr (25% added with steady state value of OP)
Entering these upper and lower bounds, we have the window, shown in Figure 5.31,
for the level controller.

Ul«l
ftv Ut» Wnd?* He«

di*h - a <e [d s 3 » .i h « oaga feslaiimm :

r tt 7t teiloos J ,w ii 5V s-: ;

[M

h
jl M253G h PBODUCT)-C|>

Oulf/

S firty. BUMMM
|1SBI27

91
l-wtions <"-ie nol elikinst«d Ivrrsuse they had residuals over le-OOS
h total ol 109 equations '29 S;| vete .Uaiaat«d
StTOlation has H'i v«ri*&lM 260 aqtrnttoM end 989 iior-:eros

Vital iV«nw:*00(

FIGURE 5.31

Again the typical ranges for the temperature control loop are noted here.
Process variable
Range minimum: 52 50C (25% subtracted)
.

Range maximum: 87 50C (25% added).

Output

Range minimum: 1 1447 MMkcal/hr (25% subtracted)

.

Range maximum: 0 6868 MMkcal/hr (25% added)

-

The corresponding Aspen Dynamics window is shown in Figure 5 32. It is worthy to .

mention that the negative value of heat duty reveals the cooling operation (heat removal).
250 PROCESS SIMULATION AND CONTROL USING ASPEN

o *y .a m w |n,..mic

, .

,r ..- j 1 1«t5) |MMlS3R

'
9

Editing Siaulacion

Validation coBplete

urrent snapihois havo been saved to 11 le (r-pAflOOO anp

i - I
i

FIGURE 5.32

Both the control algorithms are completely specified above. In the next, the controller
performance will be examined in terms of set point tracking (servo) and disturbance
rejection (regulatory).
(c) Starting the Run: Before running the program, we must be accustomed with
some frequently used items of the toolbar as described in Figure 5.33.

Step Re-start Simulation

*c 601 lacJt R» Wnto*

. IMJSI

Run Pause Rewind to a saved Snapshot

FIGURE 5.33
 DYNAMICS AND CONTROL OF FLOW DRIVEN PROCESSES -

251

We wish to carry out the simulation for a certain time, say 5 hours To fix up this .

time period, select Pause At from the Run pulldown menu or simply press Ctrl+F5 on
the keyboard. Then select Pause at time, type 5 in the field or whatever we want and
click on OK (see Figure 5.34).

yrr&ntmsmfmsssi :
.
1.1
,,.

a uai
,

, s - MA

c; |W)-

1 « 1 : O -

jj j ' (fnci cil C-

,

ftC5tR

S i . 1

i ;;;.-,;"',:::/;;:,:'m;.;:'::;f iv,i:
,
J

FIGURE 5.34

Viewing servo performance of LC1

As we double-click on LCI block in the flowsheet, first the faceplate appears. In the
next, press on Configure and Plot symbols in the faceplate. Alternatively, to open the
faceplate Configure dialog box and ResultsPlot dialog box, first select LCI block, then
.

choose Forms and subsequently press one-by-one on faceplate Configure and ResultsPlot, ,

respectively Judiciously arrange all three items within the Aspen

.
Dynamics window
(see Figure 5 35) so that we can properly observe them together.
.

Eiaagi ' 'i Baaaa wpw-

! .

k* h s a © w [n n 3 B9t> esaspa

r nit a«(o» 1] i * "t

< {MB

n mm MUUI

ma : -
.
ja_!J
nMm I fmg, | <
! r.j J
-

>.,,
pi KJS» ,

|r~ 1"
l
~

1
-

,
1 «, 1

I togji I fat.

FIGURE 5.35
252 4- PROCESS SIMULATION AND CONTROL USING ASPEN

First make sure that all the items in the Configure dialog box and faceplate are
correct. In order to execute the dynamic closed-loop simulation click on Run button in ,

the toolbar. During the simulation run, give a step change in the set point value of
reactor liquid level from 0.914029 to 1.1 metre at time = 1 5 hours. Typing the new set .

point value in the faceplate, press Enter button on the keyboard so that the Operator
set point value in the Configure dialog box also changes automatically to 1 1 . meter
Note that the new set point must be within the specified ranges of PV In Figure 5 36 ,

the servo performance of the level controller is depicted for 5 hours as selected earlier
Obviously, the plot also includes the manipulated input profile .

EC

DlSySa IE -V | Dynamic 3
* ifl)_l
T tt Tl' ©id 10 05 it
! Rowsheet

Coopoient Lots
*
} Dagncsl'cs SS

OPgHi 1210356171 "

t'POC I
IT
to JDJ*]
Tuning ] flanges | Ftong | Olhw

Owralw set pewit p 1

Twmg paramtf er,

|21083.61 RCSIR

1 C; Plot 8LOCKSCLCl-l Reia»«Plol

: lrJe.;tdlme: limtjo | m

I" l d
-

De<iv alive time: "»

E

If5 11
'

Cpndolei adkm
: Direct
-
( Reveise
||
:o» 4 97
-
obi 4

0 0.5 1 15 2 25 3 35 4 45 5 1 5000e*ODO. MDCepted

Run coaplete Time Hours

Paused

FIGURE 5.36

Figure 5.36 represents an excellent set point tracking performance of the level
controller (P-only). Obviously, the LCI provides process responses with almost no
deviation from the desired set point value and with very fast approach to reach the
target liquid level.
Notice that the above plot can be edited by right clicking on that plot and selecting
Properties option or by clicking on that plot and pressing Alt+Enter on the keyboard.
In the properties window, user can modify the title, axis scale, font and colour of the
plot. Alternatively, double-click on the different elements of the plot and modify them
as we like to improve the clarity and overall presentation.
Now, we will discuss the interaction of two control loops. When we introduce a set
point step change in the reactor liquid level, the LCI scheme attempts to compensatell
for the changes through the manipulation of the efluent flow rate. This, in turn, wi t
disturb the reactor temperature and loop 2 will compensate by manipulating the hea
 DYNAMICS AND CONTROL OF FLOW DRIVEN PROCESSES-

253

removal of the CSTR appropriately. Thus we can say that loop 1 affects loop 2. In Figure 5.37,
Aspen Dynamics window demonstrates the loop interaction under the same set point
step change (0.914029 to 1.1 metre at time = 1.5 hours) as considered previously .

PS
~

rKjTj F= E],. ;- %

c;

.
o

BEDS
1- 3

J _JSS_J

FIGURE 5.37

Viewing servo performance of TC2

As described in Figure 5.38, open the faceplate along with Configure dialog box and a
blank plot sheet. Before starting the simulation run, carefully check all entries in the
faceplate as well as Configure dialog box. In the next, choose Initialization run mode in
the toolbar and then run the program once. After completion, go back to Dynamic
mode from Initialization mode (see Figure 5.38).

'

ehj
f

FIGURE 5.38

Now we wish to conduct the servo performance study for the TC2 controller with two
consecutive set point step (pulse input) changes in reactor temperature (70 -) 750C at
time = 1 2 hours and then 75 -> 70oC at time = 3 hours).
.
254 PROCESS SIMULATION AND CONTROL USING AS PEN

Clearly, the proportional integral controller with default tuning parameters values
shows a high-quality temperature tracking performance. As stated if the performance ,

of any controller is not satisfactory, we have the option to tune the parameters simply
by trial-and-error method.
If we introduce a set point change in the reactor temperature the TC2 controller ,

takes necessary action with adjusting the heat duty to compensate for the changes .

But interestingly, the liquid level remains undisturbed. Figure 5.38 confirms this fact .

At this point we can conclude that loop 1 affects loop 2, but loop 2 does not affect loop 1 .

Actually here the interaction is in a single direction.

(d) Viewing regulatory performance of LCI and TC2: To perform the
regulatory study, we need to introduce at least a single change in the input
disturbance. However, here we consider two subsequent step changes in the
feed temperature. Initially, the feed temperature changes from 75 to 80oC at
time = 2 hours and then the temperature (80oC) returns to 750C after 1.2 hours .

To change the feed temperature twice as prescribed above, first we need to open
the feed data sheet by double-clicking on the FEED block in the process flowsheet
(see Figure 5.39).

-
Ch55.2 _
BC5rRd»rf
Bt TooB Rui Van** Hefc

*? iDyn.mk 3 > « Hi « «»© 3 [3 H

Alllms

jDynamcs o -

; SyiieriLbrjry
,

Q Table STHEAMSfFEED") Re«*« KH

Description Value linrt*

F Tolnl noSe ItoW

fm T JalfTiaaa Bow 210836
T 75 0 .

P Press U'e bai

vf Milar *api>f tricion oo
n Mrls! >(ihftlp> -87 5«7I
Rho 16 3267
MM Mohii msight 52708
mn
uebnictan 0t25
ZrfEAC Ucte rrocton

Lkie Iradton

Solver Componert
ftXlOTrl LMs

OUgnogucs GMmIc
ntegrating fra» 4 96 to 4 97 1
Integrating fro« 4 97 to 4 98
Integrating fro« 4 99 to 4 99
Integrating fro* 4 99 to S
Step 361 Time- 5 0422e+0O0. step size- S 0000e-002. step tactor* 1 S000»*000 accepted
Run complete at 16 45 29
il I

| Jaae S Xerox* Wort || 'U Met* K bm PxAatr j

FIGURE 5.39

In the subsequent step, run the program with Initialization run mode. As it is
finished, go back to Dynamic mode. Then, open the plot sheets for both the controllers.
The regulatory behaviour is illustrated in Figure 5.40 giving changes in feed temperature
 DYNAMICS AND CONTROL OF FLOW-DRIVEN PROCESSES 255

are not
in the feed data sheet. For brevity, the faceplate and configure dialog box
included in the Aspen Dynamics window, shown in Figure 5.40.

2 a ts 7 I Dynamic

3
-

r u loo? s

scripts ] VUue | Unrrt

r RM -WOO - Mf
-

Ot* :iC£3=
-
-

trx*'*-** C fi

U
| _

p Pre»vc j111*58 t*r

j nr.urT
-

Ml
in

'1 r fidcn 0 *2f ijnctVnc-l

Mofen-tdian |0.B imMV-nc*
a

53 a

T mpeiatuie loop

-
i
; So

1
3
0 05 1 1= 2 25 3 35 4 45 5
Time Hems 000e+000. accepted

FIGURE 5.40

It is obvious that the reactor liquid level remains unchanged with a change in feed
temperature since there is no interaction involved On the other hand, the reactor .

temperature is disturbed However the TC2 controller provides satisfactory disturbance

.
,

rejection performance under this situation.

So far we have studied mainly the closed-loop behaviour of a reactor system coupled
with Aspen-generated control schemes. We did not include any additional controller with
the CSTR model In Section 5.3 we consider a distillation example to elaborate this point.
.

5 3 DYNAMICS AND CONTROL OF A BINARY DISTILLATION COLUMN

Problem statement
A partially vaporized binary mixture of benzene and toluene enters a RadFrac
distillation model as displayed in Figure 5 41. .

he column has total 25 theoretical stages (including condenser and reboiler) and
operates at a pressure in the reflux drum of 18 psia and reboiler of 21 psia. The
ow rate is 285 Ibmol/hr and reflux ratio is 2 2 (mole basis) .
.
256 PROCESS SIMULATION AND CONTROL USING ASPEN

Feed Specifications

Flow rate = 600 Ibmol/hr TOP o

Temperature = 225° F
Pressure = 21 psia FEED

Feed stage = 13 (above stage)

Component Mole %
BOTTOM
benzene 45
toluene 55

FIGURE 5.41 A flowsheet of a distillation column

.

In Table 5.1, the reflux drum and the base of the column (the 'sump' in Aspen
terminology) are specified. It is fair to use an aspect ratio (length to diameter ratio) of
2 (Luyben, 2004).
TABLE 5.1

Item Vessel type Head type Height / Length (ft) Diameter (ft)

Reflux drum horizontal elliptical 5 25.

Sump -

elliptical 5 25.

The column diameter is 5 ft. Use default values for other tray hydraulic parameters
(e.g., tray spacing, weir height and weir length to column diameter ratio). Consider log-
mean temperature difference (LMTD) assumptions for the total condenser. Actually the ,

LMTD is calculated using the temperatures of process fluid and coolant In the simulation.
.

assume constant reboiler heat duty and apply the UNIFAC base property method.
(a)Simulate the column model to obtain the products mole fractions.
(b)Keeping the default level and pressure control algorithms unaltered, inspect
the servo as well as regulatory performance of a proportional integral (PI1
controller that is required to insert to control the benzene composition in the
distillate by manipulating the reflux flow rate.
(0 Devising an another PI control scheme to maintain the benzene composition in
the bottom product with the adjustment of heat input to the reboiler, observe
the interaction effect between the top and bottom composition loops.

Simulation approach
(a) Select Aspen Plus User Interface and when the Aspen Plus window pops up.
choose Template and press OK. In the subsequent step, select General with
English Units and hit OK button. To open the process lowsheet window, click
f

OK when the Aspen Plus engine window appears.

Creating flowsheet
From the Model Library toolbar, select the Columns tab. Place the RadFrac model on
the flowsheet window and add the feed as well as two product streams. Renaming all
the streams along with distillation block, we have Figure 5.42.
 DYNAMICH AND rnNTKOI, OP KI-OW DIUVKN I'lfOCKHHKH 257

.
imaimAJmm lAdftiauugj*) ji.ojjiIzJII

o: 6 -i

'
twin TmT

FIGURE 5,42

Configuring settings
Ah we hit Nt'xf followed by OK button, the ROtUp input Corni appcarH (h«m* Kitfun! .43).
The diHtillation problem Is tilled oh: 'Cl0S6d-l00p Performanct; of a DihI illation ('olumn'
Iniportuntly, une tin- 'Dynanne' input noxle

mltiMI

lit I:: | n

NiMH
I
(
(

FIGURE 6 43
258 PROCESS SIMULATION AND CONTKOI, IISINC ASPEN

Figure 5.44 includes the Aspen I'lus (iccon/ilin infornuition We can fill up the
.

Accounting sheet with any name, number and ID.

'
-
i '1
t ft vW o**

mm jj ai LiinsMibd ad j ..ijj-j -j ai i i

jJgjJT; j Mil illR"ZI uL..]

.

O Ma
o - J-.

f> 1 " .
!(. [Ilflll,

I - MCM
fToirTiTTT
I

_
J M.>i*t 4» M 1><»

1 DSTWU DHH nwJFp £.ii*;i MJ/tac SCr-oc

-

Pei(iJ t»c P*iefi« BvcWi*

fa -1* prt-M 11 C - 0 :s*Kt5-/Jeei". (Vfl !1 I HUM (leoulcillnM

Uvain . I -l|a. n/ ' Hi'W.O | ' } rhjiHg fi tVio«imVcrt [[ Aapcn fits »?t>8 111'

FIGURE 5.44

We like to see the composition of all incoming and outgoing streams in mole fraction
basis in the final results table. Accordingly, we use Stream sheet under the Report
Options of Setup folder (see Figure 5.45).

rif F* Mix Otfa fir F*<3 l±rtv Wtwjo- Hrfc

i MHl iJM«l _
BMSLSfcalgl "'I "'I I I "I "I ~l g
""

JJ I I I FV . _
LLM_Lia|
"

FlipdrtOplta 351
bnnil | rbwihWI | BM ./SIiiimi| Piorniy [ ADA |
0 S*-*. ",
'
, w. «..
.

Itw* lo N> rrlaM »i i>tMn> ixoa)

O -

I! . P W .Hum.
. J 1

FIGURE 5.45
 DYNAMICS \NI) CONTUOI, OK I'l.OW DRIVKN ('UOCICSSKS 259

Specifying components
'

From (In- Data Brow.scr, hoIoc! ( oniponcnta/Speeipcctions to open the componont input
'

lorm In ill'' lahlc. shown in l- i mc S 1 (> Ihc Ivvo species are dclincd .

UM M» fVA I** Unfti

h d .Igil I lal -iW

0 H""

Qp
him nl

ni.urm rrniioii mo
Itarimi Ophor*

I Ml/ tC Qrm)«

ml
Mi

Wiwii d b rit»i».; li. i Muifi« ..ti feMtoi 0«oN>i«>

FIGURE 5.46

Specifying property method

,, >(. li'i on Hm- Nil
choose Properties/Specifications and gel the property Inpul form.
.

in Aspen simulation a property method originall} Includes several models for calculating
,

the physical properties For the distillation example, set the UNIPAC base method b>
scrolling down (see Figure 5.47).
Specifying stream Information
,"
next, (.pen Streama IFEED IInput ISpecifications sheel Entering the given
Values lor all State variables and teed eompo ion Ihe slream mpnl lorm looks like
.

Figure 6 'IH

Specifying block information

,"
lefl pane ol the Data Browser window select Blocks IRADFRACI Setup to open ,

Configuration sheet and then insorl the required datn (see Figure 5.49)
260 . PROCESS SIMULATION AND CONTROL USING ASPEN

> He E* Wew Oati Tooli ft FW Ut*»ry Window H*>

JjiJ J J
_
LJ~- I I I Jje
HU I J±hJill -ZliiJuLJgij
Relerenced |
0 Slream Oae* Piowxtyrwthocbt moded
B | S<i)*ww
PtocMt IVP*
Picwtymethod juNiFftr Tj
Q Cutlom OMi B«e meitwd jufJIFAC r Modlv P<ope<iy rmdett
O Repwt Opicna Heniy corrvcuwitt [
'

PelTOteum cflloHston option*

~
3
Fiw-watm method
~
3
b t-end Pfop«i«i W«le< »otut*y
~
3
P»e<rfoc«npOfi«rt9
3
Etecddyte c ciisbon option:
; O MrCorpi Chamtiiy ID r
17 . : .

_
Potjma

B fl Property Method!
.
_
j EHi Un
9 CD Wotecd» Slftjcti«

a
LJNIFAC wlh AedKrh-Kwof.g elation d tlate and Henty't law
Qt | Prap-S«ts
£ r j Adveoced
hpot Complete

[it* MbMi/Splilleri j Sepstfon | HhI Exchangees Cohaint { Readou | Piettue Changeft | Manptiatori | Sokh | UteiModrt |

STREAMS DSTVAJ Di'.H FladFiac E act Mijfrac SCf<ac FelioFisc fWeFrac BaichFrac
C \ d Fitfcn'Asien dd IM HUM

| .'- JOMptef -Mgoicft W j 0 «S-Hcto W j ) tjMftJ Twig-Mcn:«< |j tmpmn Pka

FIGURE 5.47

-
fi- ?d( [Ula T04I1 Run not tibrary Wride)v< HH'

d|b;|h| l l Nel ¥?! akiahl-gl i n.|

la]
3 J*J <<J( i 3 >>J Q|aJ n J
"

O Report Ot*or< 3 ClI.C FlsshOplioBt I I EOOpiw: I

-
ConpaTerts
Speofcebont Siij!li«mf»m» )yMIXED
Asw'Bend
"

$ 'JJ Peiro Owaclefuabon

Lghl-End Prcpofto
3 |M*f 3
Pteudocorvowta I 225 d CotrpGrem Vofcje

SEN2ENE 015
HonyCcrrps d TOLUENE 055
UNIFAC Gtohm
I p.. d
) Comp-Groupi
> Compel
d
_

B _
) Potywi lanlltow [Mote
-

.J/) Prtceft ee |600 d

*
Property M«hod9
Cj E«mal»3n 1 J
.
_
j Wcc iaf Sbucture
W p»wwet«
'

,
i Data
_
J MM*
-
Ji Pr -Sett
$ MvarvceC

.
_
| BOTTOM !oUI F
-

Jl) FEED
o

EOVMMu

STREAMS 1 DSTVA)
FoH* PMF1
Book

FIGURE 5.48
 DYNAMICS AND CONTROL OK KLOW-DKiVEN PROCESSES 261

cz:

'-
a
Mr

a
3
9

-
ap>
1->- d J

inc«a Si-vi M «. tan <Mf<B Klw Inrfv

' i iH* 11 6.H i iinl i ill

FIGURE 5.49

In the subsequent step (see Figure 5.50), fill up Streams sheet with informing t\
location 113th tray (above stage)!.

; M Mm Cm la* K» i»w> VM- M

1 r - i - t - : -
nr

j Kim
-

-
I .» -
, »«

fTTTTB

FIGURE 5.50
 DYNAMICS AND CONTROL OK KLOW-DKiVEN PROCESSES 261

cz:

'-
a
Mr

a
3
9

-
ap>
1->- d J

inc«a Si-vi M «. tan <Mf<B Klw Inrfv

' i iH* 11 6.H i iinl i ill

FIGURE 5.49

In the subsequent step (see Figure 5.50), fill up Streams sheet with informing t\
location 113th tray (above stage)!.

; M Mm Cm la* K» i»w> VM- M

1 r - i - t - : -
nr

j Kim
-

-
I .» -
, »«

fTTTTB

FIGURE 5.50
262 PROCESS SIMULATION AND CONTROL USINKi ASPEN

In Figure 5.51, the column pressure profile is defined.

SmWIon 1 [Bock IWDFRAC (nadFrac) Sell* - Dan Bnnrawl
Re Ed< Dala Tool* Run Plol Ubraty VtMum HHp JiJxJ
_
Li!iJ J JZJ J d
l ill itei
O Setup
/Condguistion j -/SUevni /Pioiiure|-/Conden:e( | I I

J/j Prc<iSdS Vnw: ]P(e;iNfepic.l(li¥

38
t j eorroM Pistture
_

tttt
O hpu 1 e
fiesuts
25 21
*
_
j tOP *
(j Bocks

8 S«tup
Q Design Specs
D Vaiy
S Heaierj Co era

lJ Putnoarxjxis
D*c»Tlefs
o BfoencJes
® Reeciavs

D Condenser Hc n.'
Reboter Hcijr\-es

_
J Tray Sung
.
D Tray ReU-g
CJ Peck Sa ig
Peck Relng
0 Praoerties
«3 -
1L 1 w
:6 Stages lf*Jtrcr*s*j-.
MKm/SpUleis Sep«alots j Heat Exchanges Cduran* } Reactors Piessute ChangH! j Man atojs j SofcU ) Uiei Models j
HZH1

STREAMS
(C-I-C-S-M-jfr-C-cT-
' DSTWU SCftac PeltoFrac R eFiac
"
'
SatchFfdc

C V a Fo V apenPlus 11 1 NUM 'Reajuedin{>i In .-vxwH-

| Chapter 2 - Wcfos . | 4] Chapter 5 - Mcti; | t]} Useful Temg - Mk Aapen Rib S -

Wo6e AcroM Pr I « S i Jff 2C

FIGURE 5.51

Entering heat transfer data for condenser and reboiler

Next select Dynamic under RADFRAC of Blocks folder. There are three heat transfer
options: constant duty, constant medium temperature and LMTD. As mentioned in the
problem statement, the condenser heat duty depends on the log-mean temperature
differential between the process fluid and the coolant. The coolant inlet temperature is
set constant. Here the temperature approach represents the difference between the
process temperature and the coolant outlet temperature at the initial steady state
Note that among the heat transfer specifications the coolant inlet temperature and ,

temperature approach may vary during a dynamic simulation whereas the specific ,

heat capacity of the coolant is fixed during a dynamic run (see Figure 5.52).
For the reboiler , it is fair to use constant heat duty computed in the Aspen Plus
simulation. However , the reboiler duty may be changed at dynamic state either by
manually or automatically with employing a controller (see Figure 5.53).
Entering geometry data for reflux drum and sump
The reflux drum and sump are specified in Figures 5.54(a) and (b) with their given
geometry data. The information on vessel orientation, head type, length (or height)
and diameter are used to compute the vessel holdup .
 DYNAMICS AND CONTROL OF FLOW DmVKN PROCKSSP S -

263

BOL \W» Zm» roote Rf> PW U vy Wndmr H«f.

j|a;|B| I I v\ r;|Tlal».|<|w| n i
i r -i i i rv i .isi J2J
] ±LI iijri 3 uu Nj
Sim

HmI i<«i;tai option [lmTD 1

o MM
D iVi j- Se*ci 1]
*y
i
1
"

p
"

I «np«amB appiMch
9 *«mCoolen

CJ (
_
]
« : M

_
j
a Reboier Moves

ii
_
j
Pat* Siing
i faw Rang
i
_

o
o
o Reoort
o ijm- SutraArm

0 O iarrjc Eaueme
Bock Opuon*

J 1
Mwr»/SpfcJlM« j Sep«arori | HMTExchangen | C(*in>> j RMctai | FtenusChangen | M. ScWi I UwHcmMi J
-
CM
'
MUM
5TRt»S FSfil SSpB

Book

FIGURE 5.52

i a ie] <PJ ntn M l i ~l I l"l *l s1! 1 ill

F- 1 -
i .

1 nr 1 leal i ibi
"
-

'

I ftmwwj-di
S Cleans

'

_
) - x<JerMr Mar*

O 'JMr SoOTOJUTW

0 a«*C«bor»

1*L

FIGURE 5.53
264 4- PROCESS SIMULATION AND CONTROL USING ASPEN

He Eck V*ew Data Tods Run Ubr*v Window H

In jJIZJ
l 1 1 mm
3 ±l_JiiJR 321] nil
"

3 Efb |m
2 PrccMei
, /Condense, j/Reboiier /Reltmi Dmml Sump j
.
Jfl Steams ~

Vessel lype |Hotizonial 3

: Vessd geomstiy
!1J 0«isn Speci Headlype
l Vaiy
i Lengtlk 5 fl z
9
_

Heater CocJeo
?ii«>anxx>ds : Diwnetei 25 ft z
CJ .

i Oacartefi
5
_

Iniialspdcificabon
s Reactone
Cordansaf H v

jj RebalerHcu-wss

_
i Tra)' Stw>3
a TrajRing
EJ Pack Sons
_
j PactRalrtg
9 ?ro(j«rt:«
0 Eawnatei
9 Conveyance
9 Repor.
9 Uiw SiijfDU;nes

w Oynanac

'

iTangen' lc- tange'-tl ertgUi J Heod Type-EBclic oi Hemisphenc-al Acfuattenglh it Head Tvp*=FU
t

25 5t»5£S t; Rmpifotnd f

Mwen/SpliHeft | Sepaisios j H»at£xchangeit j Cokimne | Rescloit { Piessue CKsngef: | ManipUators I Solicft j UserModeis {

Mattel
'
Mom
STREAMS J Mbw FSpil
FSpB SSpW
For Help. pre«i FT

1 $ Aspen Oyna j
_
. psty Tta . j j-j spxxK hg . j Calajbto; <m

FIGURE 5.54(a)

S3ESSESS
s % EdR Mew Oa i Too* Pin flat Ubwy

3 -I I'M lEi J Si
i r jjj_fv
_ .

"

V1
_
oc-rt es 3
J/j areams
fitoda
-
RMFRK : Headiippe

0»gr. Specs Hei t

Dimtat

Totatiqi*dvoU?ehac!wn [

Morot/SpMlaft | Sopar tcf. | Hea* Exthonffm | , | (Mom | P».«.CI»W" | Maroiaw. | S<*k | U«

0 AMk OtWMn NT*

I c* .
n- | - | <...
-

~.

FIGURE 5.54(b)
 DYNAMICS AND CONTROL OF FLOW-DRIVEN PROCESSES 265

Entering tray geometry

The example column has total 25 stages-Stage 1 being the condenser and Stage 25
the reboiler. We already have inserted the necessary information for stages 1 and
25. Now, we need to inform the simulator the tray geometry specifications for stages
2 through 24. Note that the tray holdups are computed using these geometry data
(see Figure 5.55).

n» E* Vb« Mi Tooi» ft* Lfcrao VWrdpw Hsfc

"

f : fV -
I - imi I lai 1 1
3 iilFi 3 ill d l \
~

3
.
jj Stan
-

Mi

Stage' Spaong L>WO

_
j «y .
d
« :4 0 726££ 90

Oecanfm »

_
j Retokr Haxves

lj Tufting
:
_
l i

0 BtockOptofa
- . - - V

|rt>J
SeG«*tn | Hea EKChaqot | Cckm | ReKUfl | PieumChanpn | M««i>lois { SotcH | day Mod* |
-
D->
Mam*
1 .1 pr-u j.p _
t.

. . . - NUM Seoul*! rxxl CMP-e

Caiaiacr

FIGURE 5.55

Running steady state simulation and viewing results

Hit Afecf button and press OK to run the steady state simulation. Finally, the result."
table , shown in Figure 5 56, is obtained. At this time, we should save the work.
.

(b) Exporting dynamic simulation: For process dynamics study, we wish to

export the steady state Aspen Plus simulation into flow-driven Aspen Dynamics
simulation giving a file name of'Ch5 _
53
.
_
RadFrac'. Then close the Aspen Plus
window .
266 PROCESS SIMULATION AND CONTROL USING ASPEN

E* Vte*v Data To* R« Ubr*v Window Help

1 ±Lil <iJF 3«1 ,

Setup
1 1
_
P pel s
31 9mamt -

3 F-m*;
Bseki
Sue-am Table I
Rnctane

z\ d
-

Com Optons
Zl
-

3 EO Cenv Cfilioht IHttUIB psi 21,00 21 00 18 00

9 S«t*> /apoiFiac 0 000 0«5 0 000
Q 01-10 Ejk
Mole Fte Ito.oll, 315000 600 000 285 000
Q L'HO Adv
O LSSOF Base Malt Flow b i 29009731 51497.054 22487 324
Q LSSQP idv Volume Flow ctitAm 607353 14147.249 446 618
Tea.
;nthi*(. MMBHi/hi 3 99! 11.329 6 675
Convergence .

ConvOder 4cfcFlowlhcnoi/t.
-
I Sequence BENZENE 1 036 270.000
.
268,964
j RoA-shee!r>; Options
TOLUENE
_

313.964 330000 16036

:
_
j Mc- ei ttafytii T< s
Jj EO Cc guratan oleFtac

[jj ResJls Sixmrafy BENZENE 0 003

.
0450 0 944
,

9 Pijn Status
TOLUENE 0 S97
.
0 550
.
0 056

Convef ence

Roula Avoleble
"

plj Uureit/Spllten ] Separatort ) Heal Ewhangett ] CoUww ) Reacton | Piewuie OwrgBrs | MampUatois | Srtdt ) U«e( Models |

Man
STREAMS '1 Mi»«. M « FSpa
r rJJ SSl*
rA

:FocHe|p.|il»«liFi 0 VBo*-Cheplere NUM Re>jl( AnAUe

|4)Ch8i»5 W..,|| Aipen Phm... fapm.D) .,- | fapen.IVia-.. } apocfeheat:. [ CefcuMof | « Q P'tVi 11 53

FIGURE 5.56

Starting Aspen Dynamics

Open a blank dynamic simulation window for the example column, following a similar
procedure as previously shown for the CSTR problem. In the next, simply open the
flow-driven dynamic file 'Ch5 5 3 RadFrac.dynf. As a result, the Aspen Dynamics _
.
_

window appears (see Figure 5.57) accompanying with the closed-loop process flow
diagram. The flowsheet actually includes the three default control schemes LCI, PC2
and LC3 to monitor the reflux drum liquid level, top stage pressure and column base
liquid level, respectively.
In the present discussion, we do not want to change anything of the three
automatically inserted control strategies. All data, including timing parameters, ranges,
bias values and controller actions, remain untouched. A little detail of these control
structures is given below.
Loop 1
Controller: LCI
Type of controUer: P-only (since integral time is very large (60000 minutes))
Controlled variable: liquid level in the reflux drum
Manipulated variable: distillate flow rate
Controller action: direct
 DYNAMICS AND CONTROL OF FLOW-DRIVEN PROCESSES 267

»g [Dynnmlt J h« .IS'ByiUEiafl
inM
-

r tt a«i|oo5 J ii. i» -ji f

LCI
r>0 .

C<rtenta d Snmiaaon

£5 9 »o--
BOTTOM fO

LlJ

Siaulation ready for solution

7 equations were not eliminated because they had residuals over le -
005
3
A total of S32 equations (38 b'/.) were eiinmated
Sinulation has 1991 variables. 1484 equations and 6219 non-zeros

IBjnsilcal 0.(10
'
Aspen fVjs Mty8 [>i_
I A »n IVvrtc-

FIGURE 5.57

Loop 2
Controller: PC2
Type of Controller: PI
Controlled variable: top stage pressure
Manipulated variable: coolant inflow rate
Controller action: direct
Loop 3{
Controller: LC3
Type of controller: P-only
Controlled variable: liquid level in the column base
Manipulated variable: bottoms flow rate
Controller action: direct

Adding a new PI controller for top composition loop

Now we wish to include a proportional integral (PI) law to control the benzene
composition in the top distillation product by manipulating the reflux rate. In the top
left of the window the Dynamics library is included within Simulation folder of Ml
Items pane Click on expand (+) button of Dynamics subfolder. Consequently, the expand
,

button changes to collapse (-) button as shown in Figure 5.58.

268 PROCESS SIMULATION AND CONTROL USING ASPEN

* I"-- 3 » ' i< . w t a y u n ca r-

-

r tt o p ,
ar s _

Bp

tot»l ol 91? Kiu tion* lie IK] v.c. .LktAated

ixuiation ha* I'm variabl** l«a< aquat.on. and til'

J
FIGURE 5.58

Again hit expand button next to the ControlModels icon Then select PID controller .

drag it to the flow diagram and drop the control block near to the top product stream .

Renaming the top composition controller as CCT we have Figure 5.59. ,

Uae l*.
lO i£i
Si*?0-
no
-
TwT-o
1
.

Id V
§1 - t
-Q: h tgnm K>

9
iai»i
-

PS Rnuu 3

> und to npJtQOlO

FIGURE 5.59

Connecting controller with process variable (Controlled variable)

Ixpand Stream Types under Dynamics subfolder and hold down the mouse button -
on
ow
he ControlSignal icon. As we drag it onto the flowsheet window, many blue an
appear around the process diagram. Interestingly, when we f l I Z S
wiih holding the ControlSignal icon over a port, the name of that P
.

*ame *
on the
Anyway, move the pointer and release the mouse buttonselect 0fe g wo
the dastillate compos.tio
OutputSignal originated from TOP (stream) block. To
 CONTROL OF FLOW
DRTV rPPn cc 269
of benzene as controUed variable, choose 'STREAMSCTOP") ZnCBENyFTWv
fraction' by description in the Select the Control Variable 4 bo Fi 5 6of
.

W lOyti.Mc

SIR£ AMS| TOPlfcrf BEHZE

SlfiEAMSC ICiP lfcnf I0LUE

siREMnrniPi fmanaa
SIBEAMSI'TOP'l fiTtrtTOi U
si a
.
.

SIRDMSTTOPtP

STREaMSriOPlZmrtBEftZE Moslhacten
STSfAMSlTDP
"
'

IQUJE

Ucli total

it snapshots houe been saved to lile snp*00l0 I

FIGURE 5.60

As we press 0/!l button, the cursor becomes a solid black arrow representing the
input signal to the controller. To transmit this signal to the CCT block, connect the
black arrow with a port marked InputSignal. Since this signal conveys the process
variable (PV) information to the CCT controller select 'CCT.PV by name with 'Process ,

variable
'
by description in the Select the Control Variable dialog box (see Figure 5.61).
-
Ulxl

0«B'<BQi 1% X? |Dyn.mlc

.
ri
.

FIGURE 5.61
270 PROCESS SIMULATION AND CONTROL USING ASPEN

Hit OK button and obtain Figure 5.62. Obviously the CCT controller is partially ,

configured. To complete the top composition loop, the controller output should be
connected with the manipulated variable to pass on the signal .

B pen Oynmics wiya_Oia I dyrf

Re E-*l TocJs Rowsheei Run Window Met

D fii; H # El CS iDynamic > [.

s FtwrilMl Window

r Utt G..d |0 05 t!
) Modd)
( Parametef Types
>C2
:
| Pol TflJW
i Pre5»ureCnsr»geRi 0 LCI
J PMH . rc- Bl
| Prcceduw
) F actora

| Seoaratora
I Slreani Type* CH FEED [
J Cwviedion LC3 >o
.

CortrolStgnal >o- o r

ConteniiQl ContiptSignal | B0TT0M]</

-

WMlH Controlig SySer

lation has 1931 variables 1484 equations and 6219 non-zeros 3

ing Simulation

ent snapshots have been saved to file snpAOOlO snp

I
Ready Jl focal Dynamic at 0 00
aart| p , m " Bopk | -jjChap-er 5 Mi | Jrj UaeMTgms | Adobe Acroba | Goo e Wnd | A*>en Ptu: C Dm

FIGURE 5.62

Connecting controller with control variable (Manipulated variable)

Again hold the ControlSignal icon, drag it onto the process flowsheet and drop it on
the blue outgoing arrow marked OutputSignal from the CCT block. As Select the
Control Variable dialog box appears (see Figure 5.63), choose 'CCT.OP' by name and
press OK.
Immediately, a solid black arrow representing the controller output signal is
automatically generated. Move the mouse pointer to reflux stream and make a
connection to InputSignal2 port. To use the reflux flow rate as control variable, select
'
BLOCKSC'RADFRAC). Reflux.FmR' in the dialog box and click OK (see Figure 5.64).
Now the binary distillation column is coupled with four control schemes, LCI,
PC2, LC3 and CCT, and the closed-loop process looks like Figure 5.65. The subsequent
discussion includes the modification of different tuning properties of the CCT
controller.
 DYNAMICS AND CONTROL OK FLOW DRIVKn PROCESSES -

271

HKBB

OwtB W | Dynamic J
M l
r UTtlMln *J Li \i r %

BIBB
Piocmi rmbto

3
Edit tnq SimuIsIton

Current snapcbom timvm bMn «*v«d tc 1)1* onpAOPll «np

L Uj . i

FIGURE 5.63

0 * H .[! (B V |Ovn.mlr
IS, , . 'i . 1.. - V..- , * iDl
-

3 r it 7f o-d|uir, i; wit » s

U««4 »w* l>vim cofln'

tQAiM
£1

tILUlf SI-HAI'FtUC' lMaoril.'l

nLOc»irfuiiFHw: im miii

eioc» srwM «*ri vi h ' i

eiotr srnADrrut'isi mim

yi||

i
M toed tW*

FIGURE 5.64
272 PROCESS SIMULATION AND CONTROL USING ASPEN

6* a »Q fl W |Uyn.mlt J >

r « Orfloffi ,

,
>o
J Stmam Tipca

OH FEED h
>o
J >0-- v m

Editing 5i»ulstior.

Current sn&pshcta havo bee 'ad to tile snpAOOU snp

U I

FIGURE 5.65

Modifying controller tuning properties

First we wish to see the default tuning properties. So double-click on the CCT block and ,

then hit Configure symbol in the faceplate to open the Configure dialog box (see Figure 5.66).

Fie y*w Too. Pui Wrx w Kefe

j Dynamic ] 5B»B fe»iaOESD9

3
* |r n -it c..i|o"-, jj « n

r LCI

CwT-o
Tuwnp | Raryii jf ffamg J Ottw
g>-ffnoT-
OteianMnMlai
y
TtmepaianeieK : H BOTTOM K)
Bm |o
.* V
irOjaJim |» 1- d
1- d

i 2031 varittbles ISOi equsiions and 63 3 i

jrwnt snapshots have been aawd

-
4 .- »

FIGURE 5.66
 DYNAMICS AND CONTROL OF FLOW-DRIVEN PROCESSES 273

Obviously, some of the default values set by Aspen Dynamics are not acceptable.
For example, the operator set point value of process variable (benzene composition in
distillate) should not be greater than L Secondly, the CCT controller action must be
Reverse'. In addition, the value of control variable (reflux flow rate) at steady state is
'

usually used as bias value.

We have two options in our hand to correct the default values. Either manually
we can do it or Aspen Dynamics can automatically initialize the values of set point ,

process variable, control variable, bias and ranges. Note that the controller action is
changed only manually. It is wise to initialize the values by the help of Aspen Dynamics .

For this, press Initialize Values button in the Configure dialog box and use 'Reverse'
controller action. It is obvious in the window, shown in Figure 5.67, that the values of
SP, PV and OP in the faceplate change automatically to their steady state values. If
this approach fails to initialize the simulation of controller model with the steady state
data, check and replace, if necessary, the values of PV and OP with their steady state
values by double clicking on signal transmission lines (input to the controller and output
from the controller).

fie Vx* Tods Wndow

j Dynamic ].] flS-e LJL?nH!

Allttro 3
r Gndi-JCB zl 2* H

acci Mill
.
O LCI
10 3437
10 9437
|49472 na
< -{TfLCL.j
LCI
if
Turong ] Ranges ] nam T >o-
1
1 BOTTOM K>
0|wdlo«»t|)«* 10 343 33 IfcmoMbmol
Tuwigpawieleit

Bu; |43472 1105 \bAt

a«. 4|l
1 zi
Oenvative me .

|o
Conbofct Kljon
r Dieo

nables 1494

been saved Co file snpAOOOO snp

» loot DynmcaOOO
J
It | JlJ Ouplg 5 >4ct»<I W j AJOOe toif friM \ tm>on ft» *»l»-0« jj '' «

FIGURE 5.67

Modifying ranges for process variable and controller output

we hit the Ranges tab, the Configure dialog box (see Figure 5.68) shows the default
ranges imposed on process variable and controller output.
274 PROCESS SIMULATION AND CONTROL USING ASPEN

Db Tat* Kr XVhoa* rt*

- W -; . (ft |lnlllBll?nll<in

jo g*-'.' >C
BBSiwi
Q Hc

>o-
| H bohom f<'

P Opts range

0><pki

Rw trnwun |3CS4i rill |t/N

[7

r > :-.: ¥.. rz :!.: i. v-i

_
::n FlIF t'.-r.i:. THE-: t;-:.
"

equation B10CK5CLC1 ) CalcDillTer* moved tro« ELSE branch to THEN branch

equation BLOCKS( LCI') CaicPropT«r» HMd (ro. ELSE branch to THEH branch
'

has 2031 variables 150S equations and t>32i non-zeros

Ready

5.11 r
'

FIGURE 5.68

However, here we use the typical variable ranges, shown in Figure 5.69 .

He «e» Toe* fin Wnjcm hds

»ff |rnitializMion « fi S1 © 0 [3
3
3
3 SrUMon
% iaizi

>o -

L
-Q H 60MOM [<>
nwpwjifiiiii [T

.
__g groups 664 lo 670 of type Explicit Linear
Solwing group fc'l sire S type Linear
Group 671 already converged
living group* 672 to 601 ol type Explicit Linear

fun coaplete at 11 16 14
J
..
/is

FIGURE 5.69
 DYNAMICS AND CONTROL OF FLOW DRIVEN PROCESSES
-
275

Process variable
Range minimum: 0.85
Range maximum: 10
Output
Range minimum: 10000 Ib/hr
Range maximum: 120000 Ib/hr
It is important to mention that it is a good idea to carry out Initialization as well
as Dynamic run after each new change in the control scheme so that any error in
controller installation can be detected individually.
We have now completed all required control specifications for the top composition
loop In the ongoing study, we prefer to conduct the simulation experiment to observe
the designed controller performance continuously for 5 hours. As done for the previous
CSTR problem, similarly either simply press Ctrl+F5 on the keyboard or select Pause
At rom the Run menu and put 5 hours as Pause at time.
f

In the next, we will inspect the CCT controller performance first dealing with the
servo problem followed by the regulatory problem.

Viewing servo performance of CCT

As we double-click on the CCT controller block in the flowsheet window the faceplate
,

appears. Then open the Configure as well as ResultsPlot dialog box The second one is
.

basically a blank graph sheet that presents the variations of process variable set point
,

and controller output with respect to time.

Before running the program make sure that all the items in the Configure dialog box
,

and faceplate are correct. In the next hit Run button to start the dynamic simulation. The
,

plots, shown in Figure 5 70, illustrate the servo behaviour of the PI control algorithm with
a step increase (0 9437 - 0 97 at time = 1.51 hours) followed by a step decrease
.

(0 97 - 0 9 at time = 3 hours) in the set point value of the distillate composition of benzene.
.

To achieve an improved closed-loop performance we have used the values of proportional

,

gain of 10 %/% and integral time of 10 minutes. These values have been chosen based
on a pulse input test in the distillate composition of benzene and using the trial-and-
error approach It should be kept in mind that the objective at this point is not to come
up with the best control structure or the optimum controller tuning. We only need a
control scheme and tunings that provide a reasonably good tracking performance to
drive the simulation to a new steady state.
Remember that to edit the plots, shown in Figure 5 70, double-click on different
elements of the plots and modify them as we like.
Viewing regulatory performance of CCT
In order to investigate the regulatory performance of the CCT controller, we give a step
input change in the feed pressure (21 -» 23 psia) at time = 1.48 hours and that in the feed
temperature (225 -» 230oF) at time = 3 hours. The PI controller tuning set provides good
disturbance rejection performance (see Figure 5.71) although the tuning parameter values
.
gain and integral time) have been chosen based on a pulse set point input change.
276 PROCESS SIMULATION AND CONTROL USfNG AST'EN

DgB «a IB <<l |Dyn.,nk 3

r a it Mi-nr 2} £ i s

»o

jtOTTOMfO

lor* 1 SOOOd OOO acceptttd

FIGURE 5.70

He ,4> Teds Rn ..r.x. Heb

3 S H « (i V? lOyn.mk

t Eyewntcrari

Tyl-o

>o
--

HQ- -. ccr
'

33 31

'

il
3
Z«-B£KZE«-J
5 05 1 If 2 2£ 3 3S < «5 5
Time Hoiks

1
j

FIGURE 5.71

(c) Adding a new PI controller for bottom composition loop: We have to

devise another PI control scheme to monitor the bottoms composition of benzene
by adjusting the heat input to the reboiler. As developed, the CCT controller
for the top loop, similarly we can configure the CCB controller for the bottom
loop as shown in Figure 5.72.
 DYNAMICS AND CONTROL OF FLOW-DRIVEN PROCESSES 277

I.:IM

.
:

1
s
.

41 . .
! m
.

lm< uuuk.'* w-> Ml X-M i.. I.im mmttfll M*

FIGURE 5.72

We have chosen the following tuning properties (see Figure 5.73):

Gain = 10 %/%
Integral time = 10 minutes
Controller action: Direct

"

dub .a % » |M*«.. r 3 . . nvm awocigi

_

r :: f - >. »- »

1 55 i; 7T i '».»« i
:
'w. .
-

f v..

c MM
-

.TIM mmm lia»

EkMIMi tl A I.. li.c...

| "1 .MtMlM lftj« WMtlOM m4 M3» mm-two*

FIGURE 5.73
278 . PROCESS SIMULATION AND CONTROL USING ASPEN

In addition, the used constraints are reported below:

Process variable

Range minimum: 0.0

Range maximum: 0.1
Output
Range minimum: 6000000 Btu/hr
Range maximum: 18000000 Btu/hr

Viewing interaction effect between two composition loops

To observe the effect of interaction between two composition loops, the set point value
of bottoms composition of benzene has been changed twice. The simulation result is
depicted in Figure 5.74 for a step increase (0.0033 -> 0.0045 at time = 1.5 hours) followed
by a step decrease (0.0045 -> 0.0025 at time = 3 hours).

EC MS*}
-

He Mew Toots fen Window net

| Dynamic h « CBS'S aiaocsH

9 -
=inj l
"
_

r Tt (0 05 3 i£ i « W 1* 3

LCI
>o-

%j am {33-0
SP| 100025
PV; I00O25 LCJ
--

>a-j
0PHH |12117835,; >o- CCT

BOTTOM K>
Tunuvj | Ranges | Ffenr ] Olha

Opetatot set poi* (0 0025 jbmol/bfrol

Turwtg paameten

S«t. [l1613W7r jBtu/W 63

Gart [To I .
s.
Iniegrafww. flO f i
Owwabvetimei [o
-

11
Cortiolw action -
Drad > £
i
r Reverte S?
1
-

If r
0 0 5 1 1 5 2 2 5 3 3 = 4 45
,
5 i 2 2.5 3 13
Time Hours Time Hours

TEE
teed

FIGURE 5.74

Clearly, the CCB controller shows satisfactory set point tracking performance
against a pulse input change It is observed from Figure 5.74 that owing to strong
.

interaction between the two composition loops of the distillation column the set point ,

changes in bottom loop affect the top product composition Similarly, when any set .

point change is introduced in the top composition loop, the bottom product composition
will also be affected.
 DYNAMICS IND CONTROL OF FLOW-DRIN KN I'HOrKSSKS 279

SUMMARY AND CONCLUSIONS |

This chapter has investigated the closed-loop process dynamic characteristics using
Aspen Dynamic- package. To observ e the controller performance in terms of set point
tracking and disturbance rejection, a CSTR in addition to a distillation column have
been illustrated The default control strategies have been tested for the reactor example,
whereas the two additional composition control loops have been included along with
the default control laws for the distillation example. Several simulation experiments
have been executed for both the processes under flow-driven dynamic simulation. Note
that Chapter 6 presents the dynamic simulation and control of more rigorous pressure-
driven dynamic process.

PROBLEMS|
5 1 A feed mixture of benzene and toluene is fed to a flash drum (Flash2). The
.

separator operates at 1.2 atm and 100oC For dynamic simulation, required feed
.

specifications are provided in Figure 5.75.

Feed

Temperature = 25°C
Pressure = 3 bar
FLASH
Flow rate = 100 kmol/hr
>o
PI
Component Mole fraction

benzene 06.

toluene 04
.

FIGURE 5.75 A flowsheet of a flash drum.

'
a) Use the SYSOP0 property method to compute the amounts of liquid and
vapour products and their compositions .

.
b) As shown in Figure 5 75, employ a PI control scheme to monitor the
.

temperature in the flash drum by manipulating the heat duty .

(c) Show the closed-loop servo performance with +10% and then -10% step
changes in the flash temperature .

(d) Report the tuning parameters obtained by trial-and-error method, controller

action and ranges imposed .

5 .
2 A vapour mixture of toluene, methane and hydrogen is heated using a shell and
tube heat exchanger (HeatX) The superheated steam is used as a heating medium.
.

Complete specifications required for closed-loop dynamic simulation are shown in

Figure 5 76.
.
280 PROCRSS SIMULATION AND CONTROL USING ASPEN

Hot Stream Out

Pressure = 14 psia
Cold Stream In

Temperature = 2780F I
HOT-OUT
Pressure = 500 psia
Cold Stream Out
Component Flow rate
(kmol/hr) ICOLD-INf j cold-out hoi Temperature = HOOT
Pressure = 498 psia
toluene 200
Dead time pi
methane 2300
HOT-INK >AT
hydrogen 1000 >o

ii
Hot Stream In

Temperature = 1160oF
Pressure = 14.7 psia
Flow rate = 5110 kmol/hr

FIGURE 5.76 A flowsheet of a heat exchanger.

(a) Simulate the heat exchanger model using the shortcut method, counter-
current flow direction and NRTL-RK property method.
(b) Include a PI control structure to observe the set point (cold stream outlet
temperature) tracking performance and the manipulated input (steam inflow
rate) profile. In the closed-loop simulation experiment, assume that the
temperature sensor takes 1 minute time (dead time) to measure the controlled
variable. Report the used tuning properties.
(c) Examine the regulatory performance by introducing + 10% and subsequently
-

10% step changes in the inlet temperature of the cold stream.

5 3 Device a cascade control scheme for the above heat exchanger and investigate
.

the controller performance.

5 4 A liquid mixer model with a typical ratio controller (Seborg et al. 2003) is shown
.

in Figure 5.77. The flow rates for both the disturbance or wild stream (Fw) and
the manipulated stream (FE) are measured, and the measured ratio, R = FE/Fw m ,

is calculated. The output of the ratio element is sent to a ratio controller (PI) that
compares the calculated ratio Rm to the desired ratio Rd (set point) and adjusts
the manipulated flow rate accordingly.

Input 2

Ratio
>o
PI
> | POT >

Input 1

FIGURE 5.77 A flowsheet of a mixer

 DYNAMICS AND CONTROL OF KLOW-DRIVKN PROCESSES 281

The input data are shown in Table 5.2 for simulation.

TABLE 5.2

Stream Temperature CO Pressure (aim) Flow rate (kmol/hr) Composition

E 50 1 Pe = 100 pure ethanol
W 60 1 = 150 pure water

Process variable at steady state = 0.667 (FE/FW = 100/150)

Controller output at steady state = 100 kmol/hr
Proportional gain = 4 %/%
Integral time = 20 minutes
Controller action = reverse
(a) Appljang the SYSOPO base property method, simulate the mixer model
operated at 1 atm.
(b) Using the given controller properties and default ranges, report the
ratio controller performance with two consecutive set point step changes
(0.667 -> 0.72 0.65) in the ratio.

Hint: Double-click on Input 1 transmission line and ill up Tables 5.3(a) and (b).
f
TABLE 5.3(a)

Value Spec
*,
>STREAMS("E ) Fcn("ETHANOL")
.
100.0 Free

<Ratio.Input! 100.0 Free

Similar table for Input 2 is obtained as:

TABLE 5.3(b)

Value Spec
w
>STREAMS("W ) Fcn("WATER").
150.0 Free

<Ratio.Input2 150.0 Free

In the next, double-click on Ratio element and get Table 5.4.

TABLE 5.4

Description Value Units

Inputl Input signal 1 100.0 kmol/hr

Input2 Input signal 2 150.0 kmol/hr

Output Output signal, Inputl/lnput2 0 667

.

Use Initialize Values button and incorporate the given tuning properties before
running the program.
282 PROCESS SIMULATION AND CONTROL USING ASI'RN""

5, 5 A reboiled stripper is used to remove mainly propane and lighter species from a
feed stream, shown in Figure 5.78. It has total 6 stages (including condenser
and reboiler) and no condenser.
The bottoms rate is 100 Ibmol/hr and the column top stage pressure is 150 psia
with a column pressure drop of 8 psi. The diameter of the stripper (Stages 1 to
5) is 6.5 ft. The reboiler heat duty is assumed constant, although it changes at
dynamic state. The sump has elliptical head with a height of 5 ft and diameter
of 2.5 ft.
For the closed-loop simulation, use the following data:
Dead time = 2 minutes
Magnitude of noise (standard deviation) = 0.01 Ibmol/lbmol
Proportional gain of PI = 1 %/%
Integral time of PI = 20 minutes
Controller action = Reverse

PCI
Feed

Temperature = 40oF
Pressure = 160 psia
Feed stage = 1 (above stage)
Dead time Noise Pi
Component Flow rate
(Ibmol/hr) >o
c, 60

c2 75

C3 150
n-C4 175
n-C 60
5

n-C 35
8

FIGURE 5.78 A flowsheet of a stripping column .

(a) Using the Peng-Robinson thermodynamic method simulate the RadFrac ,

(STRIP2) model and compute the product compositions.

(b) Keeping the default controllers (PCI and LC2) unaltered configure a ,
'

composition control scheme (PI) coupling with a 'Dead time' and 'Noise _

elements to maintain the propane mole fraction in the distillate by

manipulating the reboiler heat duty as shown in Figure 5 79. Use the given .

closed-loop data and execute the dynamic simulations to test the developed
composition controller performance .

5 6 Ethylene is produced by cracking of ethane in a stoichiometric reactor. The

.

irreversible elementary vapour-phase reaction is given as .

C2H6 -i C2H4 + H2
ethane ethylene hydrogen
Pure ethane feed enters the reactor model shown in Figure 5.79, with a flow
,

rate of 750 kmol/hr at 800oC and 5 .

5 atm. The reactor operates at inlet
temperature and pressure with 80% conversion of ethane .
 DYNAMICS AND CONTROL OF FTOW-DRIVKN PROCESSES 283

pi

> M >o

FIGURE 5.79 A flowsheet of a reactor

U) Using the SYSOPO thermodynamic method, simulate the reactor model.

(b) Develop a control loop as configured in the low diagram to maintain the

f
desired reactor temperature by the adjustment of heat duty. Considering the
measurement lag of 1 minute, inspect the servo as well as regulatory control
performance. Report the tuning properties used to achieve a satisfactory
closed-loop performance.
5
. 7 A binary feed mixture consisting of methylcyclohexane fMCH) and toluene is
introduced above tray number 14 of a RadFrac distillation model, shown in
Figure 5.80.

O 1 phenol [

O 1 FEED h

FIGURE 5.80 A flowsheet of a distillation column

It is dificult to separate this close-boiling system (MCH-toluene) by simple binary

distillation Therefore, phenol is used as an extractant and introduced above
.

tray number 7 of the column The two input streams have the following
.

specifications shown in Table 5.5.

,

TABLE 5.5

Stream Temperature (*C) Pressure 'bar) Flow rate Mole fraction

PHENOL 105 14
.
100 nrVhr 10
.

FEED 105 14
.
181.44 kmol/hr 0 5/0.5
.

(MCH/toluene)

The column has 22 theoretical stages (including condenser and reboiler) with a
total condenser The distillate rate and reflux ratio are given as 90.72 kmol/hr
.

and 8 (mole basisrespectively. The pressure profile is defined with Stage 1

pressure of 1 10316 bar and Stage 22 pressure of 1 39274 .
bar. Use LMTD
assumptions for the condenser The reboiler heat duty is assumed constant. The
reflux drum and sump are specified in Table 5.6.
284 PROCESS SIMULATION AND CONTROL USING ASPEN

TABLE 5.6

Item Vessel type Head type Height /Length Diameter

(m) (m)
Reflux drum horizontal elliptical 15
. 0 75
.

Sump -
elliptical 15
. 0 75
.

The column diameter and tray spacing are given as 2 m and 0.6 m respectively,
,

(a) Simulate the distillation column using the UNIFAC property method to
compute the composition of MCH in the distillate and that of phenol in the
bottom product.
(b) In addition to the default level and pressure controllers, insert a PID structure
to control the MCH composition in the top product by manipulating the flow
rate of PHENOL stream.
(c) Produce the plots to show the closed-loop control responses, and report the
tuning parameters, control actions and operating ranges for controlled as
well as manipulated variables used.

REFERENCES|
Luyben, W.L., (2004), "Use of Dynamic Simulation to Converge Complex Process
Flowsheets", Chemical Engineering Education, pp. 142-149.
Seborg, D.E., T.F. Edgar and D.A. Mellichamp, (2003), Process Dynamics and Control,
2nd ed., John Wiley & Sons, Inc.