Conceptual Design of Chemical Plants by Multi-Objective Optimization - - PowerPoint PPT Presentation

▶
conceptual design of chemical plants by multi objective
SMART_READER_LITE
LIVE PREVIEW

Conceptual Design of Chemical Plants by Multi-Objective Optimization - - PowerPoint PPT Presentation

Apr 07, 2023 323 likes •540 views

SCUOLA DI INGEGNERIA INDUSTRIALE E DELLINFORMAZIONE DIPARTIMENTO DI CHIMICA, MATERIALI E INGEGNERIA CHIMICA GIULIO NATTA LAUREA MAGISTRALE IN INGEGNERIA CHIMICA Conceptual Design of Chemical Plants by Multi-Objective Optimization of

slide-1
SLIDE 1

Conceptual Design of Chemical Plants by Multi-Objective Optimization of Economic and Environmental Criteria

Relatore: Prof. Davide MANCA Tesi di Piernico SEPIACCI matr. 837275

Anno Accademico 2015-2016 SCUOLA DI INGEGNERIA INDUSTRIALE E DELL’INFORMAZIONE DIPARTIMENTO DI CHIMICA, MATERIALI E INGEGNERIA CHIMICA «GIULIO NATTA» LAUREA MAGISTRALE IN INGEGNERIA CHIMICA

slide-2
SLIDE 2

Piernico Sepiacci – Milan, 21 December 2016

Definition of sustainability

A sustainable product or process:

  • constraints resource consumption and waste generation to an acceptable level;
  • makes a positive contribution to the satisfaction of human needs;
  • provides enduring economic value to the business enterprise.

2 Economy Society Environment Sustainable Development Economic System Goods and Services Materials and Energy Emissions and Wastes

slide-3
SLIDE 3

Piernico Sepiacci – Milan, 21 December 2016

Definition of sustainability

A sustainable product or process:

  • constraints resource consumption and waste generation to an acceptable level;
  • makes a positive contribution to the satisfaction of human needs;
  • provides enduring economic value to the business enterprise.

3 Economy Environment Sustainable Development Economic System Materials and Energy Emissions and Wastes

slide-4
SLIDE 4

Piernico Sepiacci – Milan, 21 December 2016

Structure of the work

4 Economic Sustainability

Under Market Uncertainty

Economic Objective Function Environmental Sustainability

by Application of WAR Algorithm

Environmental Objective Function Process Modeling and Simulation Multi-Objective Optimization

Reference Component Selection

Time Series Analysis

Time Delay Correlation

Econometric Modeling

Energy Generation Chemical Process

( ) ep

  • ut

I

( ) cp

  • ut

I

( ) cp in

I

( ) ep in

I

( ) cp we

I

( ) ep we

I

Impact Indicators Selection Impact Specific to Chemical k

k

Profitability Pollution

Statistical Analysis of Pareto Fronts

  • n Several Forecast Scenarios

HTPI HTPE TTP ATP GWP PCOP AP ODP

slide-5
SLIDE 5

Piernico Sepiacci – Milan, 21 December 2016

Process modeling and simulation

5 Process Modeling and Simulation Multi-Objective Optimization

Profitability Pollution

Statistical Analysis of Pareto Fronts

  • n Several Forecast Scenarios
slide-6
SLIDE 6

Piernico Sepiacci – Milan, 21 December 2016

The cumene manufacturing process

6 Reactions Kinetics 1) Cumene reaction 2) DIPB reaction 3) Transalkylation

3 6 6 6 9 12

C H C H C H  

3 6 9 12 12 18

C H C H C H  

12 18 6 6 9 12

2 C H C H C H   

7 1

2.8 10 exp 104,181/ ( )

B P

r RT C C   

 

9 2

2.32 10 exp 146,774 / ( )

C P

r RT C C   

   

8 3, 9 2 3,

2.529 10 exp 100,000 / ( ) 3.877 10 exp 127,240 / ( )

f B D b C

r RT x x r RT x      

Pathak, A. S., Agarwal, S., Gera, V., & Kaistha, N. (2011). Design and control of a vapor-phase conventional process and reactive distillation process for cumene

  • production. Industrial & Engineering Chemistry Research, 50(6), 3312-3326.
slide-7
SLIDE 7

Piernico Sepiacci – Milan, 21 December 2016

Economic sustainability

7 Economic Sustainability

Under Market Uncertainty

Economic Objective Function

Reference Component Selection

Time Series Analysis

Time Delay Correlation

Econometric Modeling

slide-8
SLIDE 8

Piernico Sepiacci – Milan, 21 December 2016

The feasibility assessment of chemical plants traditionally follows the economic guidelines suggested in the Conceptual Design of Chemical Processes by Douglas (1988). Hierarchy of decisions 1. Batch vs. Continuous; 2. Input-Output Structure of the flowsheet; 3. Recycle structure of the flowsheet; 4. General structure of the separation system; 5. Heat exchanger networks.

Conceptual design

8

  • 1,50E+08
  • 1,00E+08
  • 5,00E+07

0,00E+00 5,00E+07 1,00E+08 1,50E+08 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Cumulated EP4 [USD] Time [mo]

20 40 60 80 100 120 140 160 180 200 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Price [USD/kmol] Time [mo]

Cumene Raw materials

2 ( ) ( ) EP Product Price Raw Materials Cost   3 2 ( ) EP EP Reactor Cost CAPEX OPEX    4 3 ( ) EP EP Separation Cost CAPEX OPEX   

1

4 4 4

nMonths t t

Cumulated EP EP nMonths EP

  

slide-9
SLIDE 9

Piernico Sepiacci – Milan, 21 December 2016

Methodology

1. Selection of a suitable reference component, which must be:

  • chosen according to the market field of the chemical plant;
  • a key component for either the process or the sector where the plant operates.

For the O&G sector and petrochemical industry, a good candidate for the reference component is crude oil.

2. Definition of the sampling time and time horizon of the economic assessment; 3. Identification of an econometric model for the reference component; 4. Identification of an econometric model for the raw materials and (by)products; 5. Identification of an econometric model for the utilities; 6. Use of the identified econometric models to determine the economic impact of the designed plant in terms of Dynamic Economic Potentials (DEPs).

Predictive Conceptual Design (PCD)

9

slide-10
SLIDE 10

Piernico Sepiacci – Milan, 21 December 2016

Time series analysis

10

Use of moving-averaged values

Moving average allows eliminating most of the high- frequency fluctuations and catching the price trend.

(Auto)correlation analysis

(Auto)correlograms report the (auto)correlation index of the time series X and Y based on the time lag between them.

Identification of the candidate econometric model Evaluation of the adaptive parameters

It requires a linear regression procedure that minimizes the sum of square errors between real and model prices.

1 1

1 n

t i

m Y n

  

 

Price series Time

Real data Moving-averaged data

       

cov , corr , var var X Y X Y X Y 

   

 

 

cov , corr , var var

t t j t t j t t j

Y Y Y Y Y Y

  

1 1 2 3 4 5 6 7

Time lag Correlation index

1 1 2 2 1 1 2 2

... ...

t t t q t q t t p t p

Y a a X a X a X bY b Y b Y

     

        

1 1 2 2

...

t t t q t q

X a a X a X a X

  

     Autoregressive model Autoregressive Distributed Lag model

 

2 , 1

min

nMonths t t a b t

Y Y

Price series Time

slide-11
SLIDE 11

Piernico Sepiacci – Milan, 21 December 2016

Econometric models

11

20 40 60 80 100 120 140 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

WTI price [USD/bbl]

10 30 50 70 90 110 130 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Benzene price [USD/kmol]

50 100 150 200 2013 2014 2015 2016 2017 2018

WTI price [USD/bbl]

20 40 60 80 100 120 140 160 180 2013 2014 2015 2016 2017 2018

Benzene price [USD/kmol]

  

, 1 , 1 2 , 2

1

Crude Oil t Crude Oil t Crude Oil t Crude Oil Crude Oil

P a a P a P RAND X 

 

     

  

, 1 , 1 , 1

1

Benzene t Crude Oil t Benzene t Benzene Benzene

P a a P b P RAND X 

     

Commodity a0 a1 b1 σ Crude oil 3.3207 1.8286

  • 98.13%

0.0313 −0.0028 Benzene 2.6984 0.0754 0,9012 94.07% 0.0843 −0.0098

2

R X

slide-12
SLIDE 12

Piernico Sepiacci – Milan, 21 December 2016

Environmental sustainability

12 Environmental Sustainability

by Application of WAR Algorithm

Environmental Objective Function

Energy Generation Chemical Process

( ) ep

  • ut

I

( ) cp

  • ut

I

( ) cp in

I

( ) ep in

I

( ) cp we

I

( ) ep we

I

Impact Indicators Selection Impact Specific to Chemical k

k

HTPI HTPE TTP ATP GWP PCOP AP ODP

slide-13
SLIDE 13

Piernico Sepiacci – Milan, 21 December 2016

The Waste Reduction (WAR) algorithm is a tool to determine the potential environmental impact (PEI) of a chemical process.

The Waste Reduction algorithm

13

Energy Generation Process Chemical Manufacturing Process Energy Supply

(

) cp in

I

(

) cp

  • ut

I

(

) ep in

I

(

) ep

  • ut

I

(

) cp we

I

(

) ep we

I

Waste Energy Waste Energy

Mass Inputs Mass Outputs

( ) ( ) cp cp

  • ut
  • ut

j kj k j k

I M x   

( ) ( ) ep-g ep

  • ut
  • ut

j kj k j k

I M x   

k

( ) ( ) ( ) tot cp ep

  • ut
  • ut
  • ut

I I I  

Total rate of PEI output PEI output from the chemical process

Where:

  • Mj

(out) is the output mass flow rate of stream j;

  • xkj is the mass fraction of chemical k in stream j.

PEI output from the energy generation Overall PEI for chemical k

?

slide-14
SLIDE 14

Piernico Sepiacci – Milan, 21 December 2016

The Waste Reduction algorithm

14

General impact category Impact category Measure of impact category Human toxicity Ingestion LD50 Inhalation/dermal OSHA PEL Ecological toxicity Aquatic toxicity Fathead minnow LC50 Terrestrial toxicity LD50 Global atmospheric impacts Global warming potential GWP Ozone depletion potential ODP Regional atmospheric impacts Acidification potential AP Photochemical oxidation potential PCOP

Overall PEI for chemical k

Where:

  • ψkl

s is the specific PEI of chemical k for

the impact category l;

  • αl is the weighing factor of the impact

category l.

s k l kl l

   

Chemical Mass flow [kg/h] [PEI/kg] [PEI/h] Benzene 22.537 0.837 18.863 Propylene 53.473 3.838 205.251 Propane 232.293 0.155 36.010 Cumene 12.682 1.196 15.164 DIPB 0.001 7.898 0.010 NO2 3.624 2.483 8.999 CO 1.594 0.305 0.486 CO2 2275.783 0.001 2.387 SO2 0.012 0.719 0.008 Methane 0.045 0.472 0.021

k

( ) , tot

  • ut k

I

slide-15
SLIDE 15

Piernico Sepiacci – Milan, 21 December 2016

Multi-objective optimization

15 Process Modeling and Simulation Multi-Objective Optimization

Profitability Pollution

Statistical Analysis of Pareto Fronts

  • n Several Forecast Scenarios
slide-16
SLIDE 16

Piernico Sepiacci – Milan, 21 December 2016

16

Multi-objective optimization

General formulation: Objective functions: Optimization algorithm: grid-search method.

           

1 2

, , , . . 1,2, , 1,2, ,

n j e j i

Optimize F F ... F s t h j ... n g j ... n          

x

x x x x x x x x x

l u

F <

, 1

4 1, ,

nMonths k t k t

Cumulated DEP4 DEP k ... nScenarios

 

( ) tot

  • ut

Cumulated PEI I nHpM nMonths    To be maximized To be minimized

Design variable Lower bound Upper bound Step size Steps number Reactor length [m] 4 10 1 7 Inlet temperature [°C] 300 390 5 19

slide-17
SLIDE 17

Piernico Sepiacci – Milan, 21 December 2016

17

Pareto optimal solutions

These plots show the non-Pareto and Pareto optimal solutions for an arbitrarily chosen economic scenario. Each solution corresponds to a discrete point of the grid-search domain, i.e. a plant configuration. As reactor length and inlet temperature increase, propylene conversion increases, thus the total rate of PEI output

  • decreases. At the same time, capital and energy costs increase, and selectivity decreases.
  • 2,E+06

0,E+00 2,E+06 4,E+06 6,E+06 8,E+06 1,E+07 3,E+06 5,E+06 7,E+06 9,E+06 1,E+07 1,E+07 2,E+07

Cumulated DEP4 [USD] Cumulated PEI [PEI]

  • 2,E+06

0,E+00 2,E+06 4,E+06 6,E+06 8,E+06 1,E+07 3,E+06 4,E+06 5,E+06 6,E+06 7,E+06

Cumulated DEP4 [USD] Cumulated PEI [PEI]

Environmental sustainability Economic sustainability

Environmental

  • ptimum

Economic

  • ptimum

Configuration Inlet temperature [°C] Reactor length [m] Cumulated DEP4k [MUSD] Cumulated PEI [MPEI] Economic optimum 365 7 8.66 7.11 Environmental optimum 390 10 −0.99 3.42 Equidistant solution 385 5 5.13 4.77

slide-18
SLIDE 18

Piernico Sepiacci – Milan, 21 December 2016

18

Pareto optimal solutions over 3000 scenarios

0% 5% 10% 15% 20% 25%

  • 25 -20 -15 -10
  • 5

5 10 15 20 25 30 35 40 45 50 55 60 65

Frequency Cumulated DEP4 [MUSD]

Economic optimum Equidistant solution Environmental optimum

2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Pareto set cardinal number

Cumulated PEI [MPEI]

4 6 8 10 12 14 16 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Pareto set cardinal number

Average Cumulated DEP4 [MUSD]

5 10 15 20 25 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Pareto set cardinal number

% Negative Cumulated DEP4

*Cumulated PEI does not depend on market volatility.

slide-19
SLIDE 19

Piernico Sepiacci – Milan, 21 December 2016

Conclusions and future developments

19

Conclusions

  • The economic sustainability of the cumene plant is heavily conditioned by the fluctuations of commodity and

utility prices.

  • The WAR algorithm can be used in conjunction with PCD to achieve both environmental and economic

sustainability.

  • Whenever a modification is proposed to improve the environmental friendliness of a process, it is useful to

question its economic viability under market uncertainty.

Future developments

  • It will be worth considering new processes based on a higher number of design variables to make the
  • ptimization procedure more compliant with real plants.
  • A further development could be expanding the boundaries of the study to include the upstream and

downstream activities related to the main process.

  • As far as the social attribute of sustainability is concerned, it will be worth developing practical ways to measure

social sustainability for both single-site (e.g., process synthesis) and multi-site applications (i.e. SCM/EWO).

slide-20
SLIDE 20

Piernico Sepiacci – Milan, 21 December 2016

20

Thank you for your attention