0% found this document useful (0 votes)
2 views

Modeling_Simulation_Lecture20

The document discusses methods for optimizing the partitioning and precedence order in chemical process design, specifically focusing on stream tearing. It outlines steps for identifying and minimizing tear streams using algorithms and optimization techniques, including the Barkley and Motard Algorithm. Additionally, it highlights the challenges of exponential growth in cases as the number of streams increases, emphasizing the need for efficient solvers.

Uploaded by

zarandluxurey
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
2 views

Modeling_Simulation_Lecture20

The document discusses methods for optimizing the partitioning and precedence order in chemical process design, specifically focusing on stream tearing. It outlines steps for identifying and minimizing tear streams using algorithms and optimization techniques, including the Barkley and Motard Algorithm. Additionally, it highlights the challenges of exponential growth in cases as the number of streams increases, emphasizing the need for efficient solvers.

Uploaded by

zarandluxurey
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

12/31/2020

Ali M. Sahlodin
Department of Chemical Engineering
AmirKabir University of Technology
1397 S.H

 Partitioningand precedence order


 Stream tearing

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 2

1
12/31/2020

 How to find tear streams optimally?


 Criterion 1: minimum number of tear streams

 Barkley and Motard Algorithm

Systematic Methods of Chemical Process


Design (Biegler et al)

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 3

 Step 1: Reduce the partition by removing any


unit with single input/output

Systematic Methods of
Chemical Process Design
(Biegler et al)

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 4

2
12/31/2020

 Step 2: Draw the graph of the reduced partition

Systematic Methods of Chemical Process Design (Biegler et al)

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 5

 Step 3: Draw the dual network for the


reduced partition

Systematic Methods of Chemical Process


Design (Biegler et al)

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 6

3
12/31/2020

Replace single-input Remove streams with


streams with their self loops
inputs

Systematic Methods of Chemical Process


Design (Biegler et al)

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 7

Tear the streams that form a self-loop

Systematic Methods of Chemical Process


Design (Biegler et al)

Any alternate minimum tearing set?

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 8

4
12/31/2020

 Prerequisite: Loop incidence array


 identify all recycle loops in the partition

A Streams
1 if stream j is in loop i Loops 1 2 3 …
aij   1 * *
0 otherwise
2 *
3 * *

Example
Courtesy of Benoit Chachuat

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 9

1. Start with any unit


2. Trace the outputs until a unit is
repeated (loop)
3. Collect the streams in the loop
4. Repeat the procedure from the
last unit before the repeated
one
Systematic Methods of Chemical Process
5. Continue until all outputs are Design (Biegler et al)

explored.

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 10

5
12/31/2020

K  (1)  L  (2)  M  (3)  L


  (7)  S  (8)  K
 (4)  O  (5)  K
 (6)  S  (8)  K

Systematic Methods of Chemical Process Design (Biegler et al)

A Streams
Loops 1 2 3 4 5 6 7 8
1 * *
2 * * * *
3 * * *
4 * * * *
Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 11

 Formulate as an optimization problem (set-covering


problem)
 Minimize the number of streams torn (wj=1)
 All loops must be torn at least once
n
min
yj
 wj y j
j 1 Integer programming problem:
n
s.t.  aij y j  1 i  1,, L
2n cases!
j 1
y j  {0,1}

5 streams in a partition32 cases


30 streams in a partition1 billion cases!
Efficient solvers needed.

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 12

6
12/31/2020

 Formulate and solve using Octave

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 13

 Number of cases grow exponentially with n.


 A very costly optimization problem

Worst-case
scenario!

n
 Can we simplify the optimization problem?

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 14

7
12/31/2020

Apply these rules repeatedly


 If a row has only one nonzero element aik≠0
 Set yk=1 and delete the row (self-loop)

 If a row dominates another row


 Delete the dominant row (why?)
 If Column Ck is dominated by Column Cj AND wk≥wj
 Delete Ck
 If Column Ck is dominated by  jJ C j AND wk   jJ w j
 Tearing Stream k is less efficient than tearing any stream in J
 Delete Ck

Apply optimization to the reduced problem.


Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 15

Minimum tear stream problem (wj=1)


A Streams
Loops 1 2 3 4 5 6 7 8
1 * *
2 * * * *
3 * * *
4 * * * *

A Streams
Loops 1 2 A Streams
1 * Loops 1 2
2 * * 1 *
3 * 3 *
4 *
Minimal tear set: {1,2}
No optimization needed!
Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 16

8
12/31/2020

 Precedence order within the torn partition

Systematic Methods of Chemical Process


Design (Biegler et al)

M L O S K

 Double tearing a loop causes delay in computations!


 Slow rate of convergence
 Potential convergence problems

Copyright © Ali M. Sahlodin, Dept. of Chemical Engineering, AmirKabir Univ. of Tech. 17

You might also like