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
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
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
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 Iout(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
Process modeling and simulation
Process Modeling and Simulation
Multi-Objective Optimization
Profitability
Pollution
Statistical Analysis of Pareto Fronts on Several Forecast Scenarios
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.
Economic sustainability
Economic Sustainability
Under Market Uncertainty
Economic Objective Function
Reference Component
Selection Time Series Analysis
Time Delay Correlation
Econometric Modeling
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
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)
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
t t j
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
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
Environmental sustainability
Environmental Sustainability
by Application of WAR Algorithm
Environmental Objective Function
Energy Generation
Chemical Process
(ep)
Iout Iout(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
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
inI
out(cp)(ep)
I
inI
out(ep)(cp)
I
we (ep)I
we Waste EnergyWaste 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:
• Mj(out)is the output mass flow rate of stream j;
• xkjis the mass fraction of chemical k in stream j.
PEI output from the energy generation
Overall PEI for chemical k
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:
• ψklsis the specific PEI of chemical k for the impact category l;
• αlis 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 Iout k(tot,)
Multi-objective optimization
Process Modeling and Simulation
Multi-Objective Optimization
Profitability
Pollution
Statistical Analysis of Pareto Fronts on Several Forecast Scenarios
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 xl x xu 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
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 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
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
Pareto set cardinal number
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
Pareto set cardinal number
% Negative Cumulated DEP4
*Cumulated PEI does not depend on market volatility.
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