H82CYS - Computer System Intro to Process Simulation 2

ODegree of freedom (n
df
) of a system:
n
df
= n
v
n
e
where, n
v
= variables; n
e
= independent eq
OIf n
df
= 0 (e.g. 3 unknowns & 3 independent
eq), the unknown variables can be
calculated.
OIf n
df
> 0 (e.g. 5 unknowns & 3 independent
eq n
df
= 2), specify the design variables &
calculate the state variables.
OIf n
df
< 0 (independent eq > unknowns)
process is over-specified.
(Felder & Rousseau, 2000)
H82CYS - Computer System Intro to Process Simulation 3
OUnknown variables for a single unit:
-Unknown component amounts / flowrate for all inlet &
outlet streams
-Unknown stream T & P
-Unknown rate of energy transfer (as heat & power)
OEquation to determine these unknowns:
-Material balances for each independent species
-Energy balance
-Phase & chemical equilibrium relations
-Additional specified relationship among process
variables
H82CYS - Computer System Intro to Process Simulation 4
Heated mixer
n
2
(kg O
2
)
n
3
(kg N
2
)
25C
n
4
(kg O
2
)
n
5
(kg N
2
)
50C
n
1
(kg O
2
)
40C
Q (kJ)
On
df
analysis:
6 variables (n
1
, , n
5
, Q)
3 eq (2 material balances & 1 energy balances)
= 3 degrees of freedom
OSpecify 3 design variables & solve the rest.
H82CYS - Computer System Intro to Process Simulation 5
O Given the following equations:
x
1
+ 2x
2
x
3
2
= 0
5x
1
x
2
3
+ 4 = 0
i. What is the n
df
for this system?
ii. Which design variable to be chosen for an easier
solution?
H82CYS - Computer System Intro to Process Simulation 6
ODetermine the n
df
for the system
OWhat are the design variables & state variables?
Mixing
n
A3
(mol A/s)
n
B3
(mol B/s)
n
C3
(mol C/s)
n
D3
(mol D/s)
n
E3
(mol E/s)
T
3
(C)
n
A2
(mol A/s)
n
B2
(mol B/s)
n
C2
(mol C/s)
n
D2
(mol D/s)
n
E2
(mol E/s)
T
2
(C)
n
A1
(mol A/s)
n
B1
(mol B/s)
n
C1
(mol C/s)
n
D1
(mol D/s)
n
E1
(mol E/s)
T
1
(C)
S
1
S
2
S
3
(Felder & Rousseau, 2000)
H82CYS - Computer System Intro to Process Simulation 7
OMass balance equations:
n
A3
= n
A1
+ n
A2
n
B3
= n
B1
+ n
B2
n
C3
= n
C1
+ n
C2
n
D3
= n
D1
+ n
D2
n
E3
= n
E1
+ n
E2
OEnergy balance equation:
DH = Sn
out
H
out
Sn
in
H
in

(assumption: P = 1 atm; temp = T
1
H
1
= 0; no
phase change; constant C
p
)
On
df
= 18 variables (6 on each streams) 6 equations
= 12 degrees of freedom
H82CYS - Computer System Intro to Process Simulation 8
O Specify the design variables:
OStream 1:
-n
A1
= 23.5 mol A/s
-n
B1
= 16.2 mol B/s
-n
C1
= 8.5 mol C/s
-n
D1
= 5.6 mol D/s
-n
E1
= 2.2 mol E/s
-T
1
= 135.0C
OStream 2:
-n
A2
= 0.0 mol A/s

-n
B2
= 57.0 mol B/s

-n
C2
= 29.0 mol C/s

-n
D2
= 15.6 mol D/s

-n
E2
= 0.0 mol E/s
-T
2
= 23.0

O Other info [constant heat capacity in J/(mol.C)]:
C
pA
= 77.3; C
pB
= 135.0; C
pC
= 159.1; C
pD
= 173.2; C
pE
=
188.7
O Determine the component flowrate & T for stream 3.
H82CYS - Computer System Intro to Process Simulation 9
OEnergy balance equation (cont.):
DH = Sn
out
H
out
Sn
in
H
in
= 0
0 = [ n
A3
C
pA
+ n
B3
C
pB
+ + n
E3
C
pE
] (T
3
T
1
)
[ n
A2
C
pA
+ n
B2
C
pB
+ + n
E2
C
pE
] (T
2
T
1
)
[ n
A1
C
pA
+ n
B1
C
pB
+ + n
E1
C
pE
] (T
1
T
1
)
(reference temperature taken as T
1
)
ORearrange the equation, solving for T
3
:


1 2
3 3 3 3 3
2 2 2 2 2
1 3
T T
C n C n C n C n C n
C n C n C n C n C n
T T
pE E pD D pC C pB B pA A
pE E pD D pC C pB B pA A

= 0
(Felder & Rousseau, 2000)
H82CYS - Computer System Intro to Process Simulation 10
Sequential-modular
Equation-oriented
H82CYS - Computer System Intro to Process Simulation 13
Tutorial 2 (continue from Tutorial 1)
Mixing
n
A3
(mol A/s)
n
B3
(mol B/s)
n
C3
(mol C/s)
n
D3
(mol D/s)
n
E3
(mol E/s)
T
3
(C)
n
A2
(mol A/s)
n
B2
(mol B/s)
n
C2
(mol C/s)
n
D2
(mol D/s)
n
E2
(mol E/s)
T
2
(C)
n
A1
(mol A/s)
n
B1
(mol B/s)
n
C1
(mol C/s)
n
D1
(mol D/s)
n
E1
(mol E/s)
T
1
(C)
n
A4
(mol A/s)
n
B4
(mol B/s)
n
C4
(mol C/s)
n
D4
(mol D/s)
n
E4
(mol E/s)
T
4
= ?
S
1
S
2
S
3
S
4
Heater, Q =
100,000 J

H82CYS - Computer System Intro to Process Simulation 14
OSolution is obtained by solving simultaneously all
the modelling equations.
OAdvantages:
-Flexible environment for specifications, which may be
inputs, outputs, or internal unit (block) variables.
-Better treatment of recycles, and no need for tear streams.
-Note that an object oriented modelling approach is well
suited for the EO architecture.
ODisadvantages:
-More programming effort.
-Need of substantial computing resources (but this is less
and less a problem with new PCs).
-Difficulties in handling large differential algebraic
equations systems.
-Difficult convergence follow-up and debugging.
H82CYS - Computer System Intro to Process Simulation 15
OSolve matrix equation: A X = B
where,
A = a known (i x i) coefficient matrix;
B = a know solution vector (i x 1);
X = an unknown vector (i x 1)
OExample matrix with i = 3:
A (3x3) X (3x1) = B (3x1) X = A
-1
B
x
y
z
a
1
b
1
c
1
a
2
b
2
c
2
a
3
b
3
c
3
=
d
1
d
2
d
3
x
y
z
=
d
1
d
2
d
3
-1
a
1
b
1
c
1
a
2
b
2
c
2
a
3
b
3
c
3
H82CYS - Computer System Intro to Process Simulation 16
OSolve for the following simultaneous
equations:
x + y + z = 1
2x - 2y + 5z = 1
2.5 y + z = 1
OSet up matrix equation:
A X = B
x
y
z
1 1 1

2 -2 5

0 2.5 1
=
1

1

1

H82CYS - Computer System Intro to Process Simulation 17
H82CYS - Computer System Intro to Process Simulation 18
C1 C2
35 kg B
50 kg T
15 kg X
n
1
(kg)
0.673 kg B/kg
0.306 kg T/kg
0.021 kg X/kg
n
2
(kg B)
n
3
(kg T)
n
4
(kg X)
n
5
(kg)
0.059 kg B/kg
0.926 kg T/kg
0.015 kg X/kg
n
6
(kg B)
n
7
(kg T): 10% of T in feed to C1
n
8
(kg X): 90% of X in feed to C1
C1: 4 variables (n
1
, , n
5
)
3 material balances
= 1 local n
df
C2: 7 variables (n
1
, , n
5
)
3 material balances
= 4 local n
df
Process: 5 local n
df

3 ties (n
2
, n
3
, n
4
)
2 relations (recovery of T & X in C2 bottoms)
= 0 degrees of freedom
H82CYS - Computer System Intro to Process Simulation 19
C1 C2
35 kg B
50 kg T
15 kg X
n
1
(kg)
0.673 kg B/kg
0.306 kg T/kg
0.021 kg X/kg
n
2
(kg B)
n
3
(kg T)
n
4
(kg X)
n
5
(kg)
0.059 kg B/kg
0.926 kg T/kg
0.015 kg X/kg
n
6
(kg B)
n
7
(kg T): 10% of T in feed to C1
n
8
(kg X): 90% of X in feed to C1
C1 balances:
B: 35 = 0.673n
1
+ n
2
T: 50 = 0.306n
1
+ n
3
X: 15 = 0.021n
1
+ n
4
C2 balances:

B: n
2
= 0.059n
5
+ n
6
T: n
3
= 0.926n
5
+ n
7
X: n
4
= 0.015n
5
+ n
8

10% T recovery: n
7
= 0.1 (50) = 5.0

90% X recovery: n
8
= 0.9 (15) = 13.5

H82CYS - Computer System Intro to Process Simulation 20
OSolve the mass balance equation using MS
Excel spreadsheet:
-0.673n
1
+ n
2
= 35 (1)
-0.306n
1
+ n
3
= 50 (2)
-0.021n
1
+ n
4
= 15 (3)
-0.059n
5
+ n
6
n
2
= 0 (4)
-0.926n
5
+ n
7
n
3
= 0 (5)
-0.015n
5
+ n
8
n
4
= 0 (6)
-n
7
= 5.0 (7)
-n
8
= 13.5 (8)
H82CYS - Computer System Simulation of Recycle Streams 21
Reactor
Separation &
recycle
Heat exchange
network
Utilities
(Linnhoff et al., 1982;
Smith 1995, 2005)
H82CYS - Computer System Simulation of Recycle Streams 22
Material recycle
Heat recycle
H82CYS - Computer System Simulation of Recycle Streams 23
Simulation of recycling system with SM
A B C D E F
Recycle stream
Unit operation
in simulator
Tear recycle stream
r
1
r
2
(Turton et al., 1998)
H82CYS - Computer System Simulation of Recycle Streams 24
Simulation of recycling system with SM
OBasic algorithms in handling a recycle
stream:
-Before the Equipment C is solved, some
estimation of stream r must be made a tear
stream occurs.
-Provided information is supplied about Stream
r
2
, we can solve the flowsheet all the way to
Stream r
1
by using sequential modular approach.
-Compare Streams r
1
and r
2
.
-If r
1
& r
2
agree within some specified tolerance
we have a converge solution
-Or else, r
2
is modified & simulation is repeated
until convergence is obtained.

H82CYS - Computer System Simulation of Recycle Streams 25
O In an isomerisation process, component A is converted to
component B.
O The mixture from the reactor is separated into relatively
pure A (which is recycled) & relatively pure product B.
O No by-products are formed and the reactor performance can
be characterised by its conversion.
O The performance of the separator is characterised by the
recovery of A to the recycled stream (r
A
) and recovery of B
to the product (r
B
).
(Smith, 2005)
H82CYS - Computer System Simulation of Recycle Streams 26
O Given the following variables:
- m
i ,j
= molar flowrate of Component i in Stream j
- X = reactor conversion
- r
i
= fractional recovery of Component i
O Mass balance equations for each unit may be written as:
- Mixer:


- Reactor:


- Separator:




m
A,2
= m
A,1
+ m
A,5
m
B,2
= m
B,1
+ m
B,5
m
A,3
= m
A,2
(1 X)
m
B,3
= m
B,2
+ Xm
A,2
m
A,4
= m
A,3
(1 r
A
)
m
A,5
= r
A
m
A,3

m
B,4
= r
B
m
B,3

m
B,5
= m
B,3
(1 r
B
)
H82CYS - Computer System Simulation of Recycle Streams 27
O Calculation sequence in SM: CCCCC.
O However, problem is encountered at the mixer, as the
flowrate & composition of the recycle are unknown.
O Strategy using SM approach:
- Tear the recycle streams
- Add a recycle convergence unit/solver in the tear stream.
- Estimate the component molar flowrates of the tear stream. This
allows the material balance in the reactor and separator to be solved,
& provide the molar flowrates for the recycle stream.
- The calculated and estimated values of the tear stream are compared
to test whether errors are within a specified tolerance.
(Smith, 2005)
H82CYS - Computer System Simulation of Recycle Streams 28
O Given the following values:
- m
A,1
= 100 kmol; m
B,1
= 0 kmol
- X = 0.7
- r
A
= 0.95; r
B
= 0.95
O Assume the flowrate of component A and B in the recycled
stream (stream 5) as follow:
- m
A,5
= 50 kmol
- m
B,5
= 5 kmol
O Setting at the recycle convergence unit/solver iteration
stops when the scaled residue is smaller than a specified
tolerance (1 x 10
-5
for this case). Scaled residue is given as:
(Smith, 2005)
value Estimated
value estimated - value Calculated
residue Scaled
H82CYS - Computer System Simulation of Recycle Streams 29