a hybrid simulation-optimization approach for the design
TRANSCRIPT
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh.
A HYBRID SIMULATION-OPTIMIZATION APPROACH FOR THE DESIGN OFINTERNALLY HEAT-INTEGRATED DISTILLATION COLUMNS
Juan A. Reyes-Labarta*, Jose A. Caballero, M.A. Navarro ([email protected])
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 2
• Concepts and general configuration of HIDiC with single heat exchangers• Concepts and general configuration of HIDiC with single heat exchangers
• Motivation• Motivation
A Hybrid Simulation-Optimization Approach for the Design ofInternally Heat-integrated Distillation Column
A Hybrid Simulation-Optimization Approach for the Design ofInternally Heat-integrated Distillation Column
Outline
• Examples and conclusions• Examples and conclusions
• Future work (inclusion of LCA and environmental impact evaluation in the optimization problem)
• Future work (inclusion of LCA and environmental impact evaluation in the optimization problem)
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 3
A Hybrid Simulation-optimization Approach For The Design OfInternally Heat-integrated Distillation Column
A Hybrid Simulation-optimization Approach For The Design OfInternally Heat-integrated Distillation Column
- Increasing energy efficiency in chemical processes not only provides economic benefits but also reduces emissions resulting for process operations => more efficient and sustainable chemical processes
Environmental aspects (THERE IS NOT PLANET “B”)
Motivation (I)
- Distillation columns are one of the major energy consumers in chemical processes and therefore represent a key point in terms of energy efficiency
Feed
Distillate
Residue
LD
...
...
n=1
n=N
LD+DQD
QR
Motivations
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 4
Optimal design of generalized distillation columnsOptimal design of generalized distillation columns
n=i
Motivation (II): Simulation-Optimization approaches
Lk’,n+1
Vk’,n+1-(VGFk)
Vk’,n+2
(LGFk)
Lk,n(=Lk,NTk) (Lk+1,0)
k’= k (or k+1)
Hysys Flowsheet (tray by tray calculations)
Reyes-Labarta, Caballero & Marcilla. Computer Aided Chemical Engineering. 2012, 30: 1257-1261.
(multiple feeds, products and/or intermediate heat exchangers)
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh.
Schematic representations of the internal existing streams at the zone connecting consecutive sectors in the case of a generalized feed side stream (GFk).
5
Optimal location of all the side streams:
Optimal design of generalized distillation columns
•Intermediate Heat exchanger
•Side product stream (liquid or vapor)
• (two phases) Side feed stream
where zopt refers to the phase composition at the optimal change point of sector k. L and H are the indexes for light and heavy key components, respectively.
Marcilla, Reyes-Labarta & Serrano. “Review and extension of the McCabe-Thiele method covering multiple feeds, products and heat transfer stages” (2012). http://hdl.handle.net/10045/23195
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 6
Optimal design of absorption systems (MINLP)Optimal design of absorption systems (MINLP)
• Solution strategy: an logic-based Outer Approximation algorithm (NLP subproblems[Matlab-Aspen]-MILP master problem[Gams]).
Distillation Column(binary variables
associated with the existence of the trays)
Motivation (II): Simulation-Optimization approaches
Reyes-Labarta et al. Computer Aided Chemical Engineering. 2011, 29: 301-305.
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 7
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
PossibleHeatIntegration!!
Feed
Distillate
Residue
LD
...
...
n=1
n=N
LD+DQD
QR
Conventional distillation column
Q B
(bottom)
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 8
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
Inter-Condenser and inter-reboiler benefits (McCabe-Thiele Method)
B
F 2
Q D
L 1, 0+ D
2
3
L 1, 0
Q
1 QE1
xB xDB,xB
D,xD
y
zF
s1
s2
s3
xB
F
QEyF
xF
xopt,F
D
yQE
yQE
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 9
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
Inter-Condenser and inter-reboiler benefits (McCabe-Thiele Method)
B
F 1
Q D
L 1, 0+ D
2
3
L 1, 0
Q
1
QA2
xQA zFxB xDB,xB
D,xD
y
s1
s2
s3
xB
F
QA
yF
xF
xopt,F
D
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 10
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
General representation of an internally heat-integrated distillation columnGeneral representation of an internally heat-integrated distillation column
B,xBB
F 2
Q D
D
L 1, 0+ D
3
4
L 1, 0
Q
2
QA3
1 Q E1
DD,xD
InternalHeat
Integration
(high pressure)
(low pressure)
Rectifying sector(light components)
Stripping sector(heavy components)
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 11
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
EV
cond
RSSS
Compp-H
Reb
F
IHER
D
QD
QR
Ln,RS
V1,SS
Ws
QIHT
-QIHT Ln,RS
(high pressure)
(low pressure)
General configuration of an internally heat-integrated distillation columnGeneral configuration of an internally heat-integrated distillation column
B
QB
Rectifying SectorStripping Sector
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 12
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
OBJECTIVE FUNCTIONSOBJECTIVE FUNCTIONS
opc CCTAC +=
1)1()1·(
)·(cos −++
+= n
n
compressorHXStcapitalannual iririr
CCC
∑=
+=hxsk
kHXS )cA·c(C 21
)](Wk3·log )k2·log(W [k1 S2
S10·8.5 ++=compressorC
lmkk
kk TU
QA
∆=
opDwSeRqop t)·Q·CW·CQ·C(C ++=
Capital recovery factor:ir = interest raten = life time of the equipment (years)
Cooling water
Electricity
Heat steam
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh.13
We propose an approach that combines the advantages of commercial process simulator (Aspen‐Hysys) to converge distillation columns, with the advantages of using a metaheuristic algorithm (PSO: particle swarm optimization) which initially has the ability to escape from local optima and not require gradient information
The original particle swarm optimization algorithm (Kennedy & Eberhart, 1995)
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
SIMULATION-OPTIMIZATION APPROACHSIMULATION-OPTIMIZATION APPROACH
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 14
Systems studied:
- benzene + toluene
- propylene + propane
- methanol + water
- benzene + cyclohexane + toluene
- methanol + acetone + water
- methanol + ethyl-acetate + water
- propylene + i-butane + propane
- benzene + toluene + ethyl-benzene + styrene
NUMERICAL EXAMPLESNUMERICAL EXAMPLES
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
PSO algorithm
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 15
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
Condenser and reboiler duties, and compressor shaft work vs overall internal heat transferred (QIHT)
0 200 400 600 800 1000 1200 1400 1600 1800 20000
500
1000
1500
2000
2500
QIHT (kW)
Q (
kW)
QD
QR
WS
TRADEOFF: It is possible to optimize the TAC reducing also the environmental impacts by heat integration!!!
Slightly increase of the compressor
work
Internal heat transferred
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh.
Services (kW) and TAC with total number of trays Optimal solutions Example 1
0
500
1000
1500
2000
2500
3000
20 30 40 50 60 70 80
Total number of Trays
kW
0
100000
200000
300000
400000
500000
600000T
AC
Total Ext. ServicesQDQRWcompQIHTTAC
TAC
Total Ext. ServicesQIHT
QD
QRWcomp
16
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh.
Services (kW) and TAC evolution with Feed flow rateOptimal solutions Example 2
0
1000
2000
3000
4000
5000
80 100 120 140 160 180 200 220 240 260 280 300 320
Feed flow rate (kmol/h)
kW
0
100000
200000
300000
400000
500000
600000
700000
800000
900000
TA
C
Total Ext. ServicesQDQRWcompQIHTTAC
TAC
Total Ext. Services
QIHT
QD
QR
Wcomp
17
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh.
Services (kW) and TAC evolution with Feed flow rateOptimal solutions Example 3
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
80 100 120 140 160 180 200 220 240 260 280 300 320Feed flow rate (kmol/h)
kW
0
200000
400000
600000
800000
1000000
1200000
1400000
1600000
TA
C
Total Ext. ServicesQDQRWcompQIHTTAC
TAC
Total Ext. Services
QIHT
QDQR
Wcomp
18
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh.
Services and TAC evolution with ∆P (Optimal solution)
0
500
1000
1500
2000
2500
3000
1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3∆P(atm)
kW
0
50000
100000
150000
200000
250000
300000
350000
TA
C
Total Ext. ServicesQDQRWcompQIHTTAC
TAC
Total Ext. Services
QIHT
QD
QRWcomp
19
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 20
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
General Conclusions (I)General Conclusions (I)
• The economic interest of HIDiC depends on the ratio between the cost of the electricity and the cost of the heating steam. However, environmentally friendly and more energy efficient technologies should be introduced to provide more sustainable processes that minimize the environmental impacts.
• HIDiC with a small number of intermediate heat exchangers can present interesting benefits, providing savings around 20-40% comparing with conventional distillation columns, depending on the components of the system, the flow rate and composition of the feed stream, and the product purities desired.
• The possibility of heat exchange between stages that are not at the same high in both sectors can also improve the efficiency of the configuration, together with the optimal location of the feed stream that is generally at the first stages of the stripping sector (but not always!!).
• The pressure difference between both sectors is also a key point of the process. The pressure of the rectifying sector should be high enough to allow the heat transfer between both sectors, but it should be as low as possible to minimize the compressor costs.
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 21
Design of Internally Heat-Integrated Distillation Columns (HIDiC)
General Conclusions (II)General Conclusions (II)
• Traditional HIDiC (with heat transfer along the total length of each sector) are less efficient when we have flat temperature profiles inside the column (i.e. with small variation in the equilibrium composition of the different trays) and high products purity.
However, HIDiC with a small number of intermediate heat exchangers can improve the feasibility and efficiency of the configuration in this kind of separations.
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 22
Design of sustainable processes
General configuration of a classical vapour recompression distillation column (VRC)
General configuration of a classical vapour recompression distillation column (VRC)
EV
cond
Comp
Reb
F1
B
DQp-cond
Qp-H
QIHT
(high pressure)
(low pressure)
R=L1,0
WsF2
p-H
p-cond
Mainly for systems with small temperature difference between the top and bottom products
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh.
A>B>C
A
B
C
I
II
III
IV
V
VI
Petlyuk configuration
A>B>C
A
B
C
I
II
III
IV
V
VI
Divided Wall Column
…Thermodynamically equivalent to…
• Thermally coupled columns/sequences or its thermodynamically equivalent Divided Wall Column (DWC) configurations
• Thermally coupled columns/sequences or its thermodynamically equivalent Divided Wall Column (DWC) configurations
Energy efficiency: heat integration configurations
Caballero, Reyes-Labarta & Grossmann. Computer Aided Chemical Engineering. 2003, 14(C): 59-64 and 2004, 18(C): 361-366.
A
B
C
D
E
ABCDE
ABC
CDE
AB
BC
DE
CD
A
B
C
D
E
ABCDE
ABC
CDE
AB
BC
DE
CD
A
B
C
D
E
ABCDE
ABC
CDE
AB
BC
DE
CD
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 24
Motivation (II): Simulation-Optimization approaches
New Vapor Compression Alternatives in Thermally Coupled Distillation ColumnsNew Vapor Compression Alternatives in Thermally Coupled Distillation Columns
Caballero, Navarro & Reyes-Labarta. Paper 399b. AIChE Meeting 2012. https://aiche.confex.com/aiche/2012/webprogram/Paper267695.html
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 25
-Reyes-Labarta, J.A. “Diseño de Columnas de Rectificación Y Extracción Multicomponentes”. Biblioteca Virtual Miguel de Cervantes (Universidad de Alicante), 1998.
http://www.cervantesvirtual.com/FichaObra.html?Ref=4845&ext=pdf
http://hdl.handle.net/10045/10023
-A. Marcilla, A. Gómez, J.A. Reyes, M.M. Olaya.; New Method for Quaternary Systems Liquid-liquid Extraction Tray to Tray Design. Industrial & Engineering Chemistry Research, 38, 3083-3095 (1999). http://dx.doi.org/10.1021/ie9900723-A. Marcilla, A. Gómez, J.A. Reyes; New Methods for Designing Distillation Columns ofMulticomponent Mixtures. Latin American Applied Research and International Journal of ChemicalEngineering, 27, 51-60 (1997). http://hdl.handle.net/10045/24679-Reyes, J.A.; Gomez, A.; Marcilla, A. Graphical concepts to orient the minimum reflux ratio calculation on ternary mixtures distillation. Industrial & Engineering Chemistry Research39(10),3912-3919 (2000). http://dx.doi.org/10.1021/ie9907021-J.A. Reyes-Labarta, I.E. Grossmann; Disjunctive Programming Models for the Optimal Design Of Liquid-liquid Multistage Extractors and Separation Sequences. AIChE Journal. 2001, 47 (10), 2243-2252. http://dx.doi.org/10.1002/aic.690471011
- J.A. Reyes-Labarta y I.E. Grossmann. Optimal Synthesis of Liquid-liquid Multistage Extractors. Escape-11 (European Symposium of Computer Aided Process Engineering), Capec (Computer Aided Process Engineering Center). ISBN: 0-444-50709-4 (Dinamarca, 2001). http://dx.doi.org/10.1016/S1570-7946(01)80076-6
Bibliography
© Juan A. Reyes-Labarta. AIChE meeting. October, 2012. Pittsburgh. 26
-Reyes-Labarta, J.A.; Caballero, J.A.; Marcilla, A. A Novel Hybrid Simulation-Optimization Approach for the Optimal Design of Multicomponent Distillation Columns. Computer Aided Chemical Engineering. 2012, 30, 1257-1261. http://dx.doi.org/10.1016/B978-0-444-59520-1.50110-X-Reyes-Labarta, J.A.; Brunet, R.; Caballero, J.A.; Boer, D.; Jiménez, L. Integrating process simulation and MINLP methods for the optimal design of absorption cooling systems. Computer Aided Chemical Engineering. 2011, 29, 301-305. http://dx.doi.org/10.1016/B978-0-444-53711-9.50061-4-Brunet, R.; Reyes-Labarta, J.A.; Guillén-Gosálbez, G.; Jiménez, L.; Boer, D. Combined Simulation-Optimization Methodology for the Design of Environmental Conscious Absorption Systems. Computers and Chemical Engineering, 2012, 46, 205-216http://hdl.handle.net/10045/24678-Marcilla et al. “Review and extension of the McCabe-Thiele method covering multiple feeds, products and heat transfer stages” (2012). http://hdl.handle.net/10045/23195
Bibliography
-Reyes-Labarta, J.A.; Navarro M.A.; Caballero, J.A. A Hybrid Simulation-OptimizationApproach for the Design of Internally Heat-Integrated Distillation Columns. Paper 488e. AIChE2012 Annual Meeting (Energy Efficiency by Process Intensification). https://aiche.confex.com/aiche/2012/webprogram/Paper267732.html-Caballero, Reyes-Labarta & Grossmann. Synthesis of Integrated Distillation Systems. Computer Aided Chemical Engineering. 2003, 14(C): 59-64.-Caballero, Reyes-Labarta & Grossmann. Simultaneous Design of Heat Integrated and Thermally Coupled Distillation Systems. Computer Aided Chemical Engineering. 2004, 18(C): 361-366.-Caballero, Navarro & Reyes-Labarta. New Vapor Compression Alternatives In Thermally Coupled Distillation. Paper 399b. AIChE Meeting 2012 (Advances in process intensification) https://aiche.confex.com/aiche/2012/webprogram/Paper267695.html