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

(1)

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

Relatore: Prof. Davide MANCA

Tesi di

SCUOLA DI INGEGNERIA INDUSTRIALE E DELL’INFORMAZIONE DIPARTIMENTO DI CHIMICA, MATERIALI E INGEGNERIA CHIMICA

«GIULIO NATTA»

LAUREA MAGISTRALE IN INGEGNERIA CHIMICA

(2)

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.

Economy Society

Environment

Sustainable Development Economic

System

Goods and Services Materials

and Energy

Emissions and Wastes

(3)

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.

Economy

Environment

Sustainable Development Economic

System

Materials and Energy

Emissions and Wastes

(4)

Structure of the work

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)

Iout I_out⁽^cp⁾

(cp)

Iin (ep)

Iin

(cp)

Iwe (ep)

Iwe

Impact Indicators Selection

Impact Specific to Chemical k

k

Profitability

Pollution

Statistical Analysis of Pareto Fronts on Several Forecast Scenarios

HTPI HTPE TTP ATP

GWP PCOP AP ODP

(5)

Process modeling and simulation

Process Modeling and Simulation

Profitability

Pollution

(6)

The cumene manufacturing process

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 2 9 12

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.

(7)

Economic sustainability

Economic Sustainability

Under Market Uncertainty

Economic Objective Function

Reference Component

Selection Time Series Analysis

Time Delay Correlation

Econometric Modeling

(8)

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

-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





 

(9)

Methodology

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

• chosen according to the market fieldof the chemical plant;

• a key componentfor 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)

(10)

Time series analysis

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 0

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

Y Y Y Y

Y Y





0 1

0 1 2 3 4 5 6 7

Time lag Correlation index

0 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_ b Y_ b Y_  b Y_

0 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

(11)

Econometric models

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]

0 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]

  

, 0 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

  

, 0 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 R² σ X

(12)

Environmental sustainability

Environmental Sustainability

by Application of WAR Algorithm

Environmental Objective Function

Energy Generation

Chemical Process

(ep)

Iout I_out⁽^cp⁾

(cp)

Iin (ep)

Iin

(cp)

Iwe (ep)

Iwe

Impact Indicators Selection

Impact Specific to Chemical k

k

HTPI HTPE TTP ATP

GWP PCOP AP ODP

(13)

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

The Waste Reduction algorithm

Energy Generation Process

Chemical Manufacturing

Process

Energy Supply

(cp)

I

in

I

_out⁽^cp⁾

(ep)

I

in

I

_out⁽^ep⁾

(cp)

I

we (ep)

I

we Waste Energy

Waste Energy Mass

Inputs

Mass Outputs

( ) ( )

cp

cp out

out j kj k

j k

I   M  x 

( ) ( )

ep -g

ep out

out j kj k

j k

I   M  x 

(tot) (cp) (ep)

out out out

I  I  I

Total rate of PEI output

PEI output from the chemical process

Where:

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

• x_kjis the mass fraction of chemical k in stream j.

PEI output from the energy generation

Overall PEI for chemical k

(14)

The Waste Reduction algorithm

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:

• ψ_kl^sis the specific PEI of chemical k for the impact category l;

• α_lis 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 _I_{out k}⁽^tot_,⁾

(15)

Multi-objective optimization

Process Modeling and Simulation

Profitability

Pollution

(16)

Multi-objective optimization

General formulation:

Objective functions:

Optimization algorithm: grid-search method.

       

 

1 , 2 , ,

. . 0 1, 2, ,

0 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_l x x_u F

<

, 1

4 1, ,

nMonths

k t k

t

Cumulated DEP4 DEP k ... nScenarios







(tot)

Cumulated PEI  Iout 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

(17)

Pareto optimal solutions

These plots show the non-Pareto and Pareto optimal solutionsfor an arbitrarily chosen economic scenario. Each solution corresponds to a discrete point of the grid-search domain, i.e. a plant configuration.

-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 optimum

Economic optimum

Configuration Inlet temperature [°C] Reactor length [m] Cumulated DEP4_k[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

(18)

Pareto optimal solutions over 3000 scenarios

0%

5%

10%

15%

20%

25%

-25 -20 -15 -10 -5 0 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

Average Cumulated DEP4 [MUSD]

0 5 10 15 20 25 30

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

% Negative Cumulated DEP4

*Cumulated PEI does not depend on market volatility.

(19)

Conclusions and future developments

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 PCDto 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 variablesto make the optimization procedure more compliant with real plants.

• A further development could be expanding the boundaries of the studyto 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

(20)

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

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

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.

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.

Structure of the work

Process modeling and simulation

The cumene manufacturing process

Economic sustainability

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



Methodology

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

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)

Time series analysis

Use of moving-averaged values

(Auto)correlation analysis

Identification of the candidate econometric model

Evaluation of the adaptive parameters



 



Econometric models

  

  

Environmental sustainability

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

The Waste Reduction algorithm

I

I

I

I

I

I

I   M  x 

I   M  x 

I  I  I

The Waste Reduction algorithm

    

Multi-objective optimization

Multi-objective optimization

Objective functions:

Optimization algorithm: grid-search method.

       

 

 



Pareto optimal solutions

Pareto optimal solutions over 3000 scenarios

Conclusions and future developments

Conclusions

Future developments

Thank you for your attention