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

SCUOLA DI INGEGNERIA INDUSTRIALE E DELL’INFORMAZIONE DIPARTIMENTO DI CHIMICA, MATERIALI E INGEGNERIA CHIMICA «GIULIO NATTA» LAUREA MAGISTRALE IN INGEGNERIA CHIMICA 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

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. Environment Materials Emissions and Energy and Wastes Sustainable Development Goods and Economic Services System Economy Society Piernico Sepiacci – Milan, 21 December 2016 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. Environment Materials Emissions and Energy and Wastes Sustainable Development Economic System Economy Piernico Sepiacci – Milan, 21 December 2016 3

Structure of the work Economic Sustainability Environmental Sustainability Process Modeling Under Market Uncertainty by Application of WAR Algorithm and Simulation ( cp ) ( ep ) I I in in Reference Component Selection Energy Chemical ( ep ) ( cp ) I I Generation Process we we Time Series Analysis Correlation ( ep ) ( cp ) I I Statistical Analysis of Pareto Fronts out out on Several Forecast Scenarios Impact Indicators Selection Time Delay Profitability Impact Specific to Chemical k Econometric Modeling HTPI HTPE TTP ATP  k Pollution GWP PCOP AP ODP Economic Objective Multi-Objective Environmental Function Optimization Objective Function Piernico Sepiacci – Milan, 21 December 2016 4

Process modeling and simulation Process Modeling and Simulation Statistical Analysis of Pareto Fronts on Several Forecast Scenarios Profitability Pollution Multi-Objective Optimization Piernico Sepiacci – Milan, 21 December 2016 5

The cumene manufacturing process Reactions Kinetics        7 1) Cumene reaction C H C H C H r 2.8 10 exp 104,181/ ( RT ) C C 1 B P 3 6 6 6 9 12        9 2) DIPB reaction r 2.32 10 exp 146,774 / ( RT ) C C C H C H C H 2 C P 3 6 9 12 12 18      8 r 2.529 10 exp 100,000 / ( RT ) x x  3) Transalkylation 3, f B D C H C H 2 C H   12 18 6 6 9 12    9 2 3.877 10 exp 127,240 / ( ) r RT x 3, b C 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. Piernico Sepiacci – Milan, 21 December 2016 6

Economic sustainability Economic Sustainability Under Market Uncertainty Reference Component Selection Time Series Analysis Correlation Time Delay Econometric Modeling Economic Objective Function Piernico Sepiacci – Milan, 21 December 2016 7

Conceptual design The feasibility assessment of chemical plants traditionally follows the economic guidelines suggested in the Conceptual Design of Chemical Processes by Douglas (1988). Cumene Raw materials Hierarchy of decisions 200 180 Price [USD/kmol] 160 1. Batch vs. Continuous; 140 120 100 80 2. Input-Output Structure of the flowsheet; 60 40   20 EP 2 ( Product Price ) ( Raw Materials Cost ) 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Time [mo] 3. Recycle structure of the flowsheet; 1,50E+08    EP 3 EP 2 Reactor Cost CAPEX ( OPEX ) 1,00E+08 Cumulated EP4 [USD] 4. General structure of the separation system; 5,00E+07 0,00E+00    EP 4 EP 3 Separation Cost CAPEX ( OPEX ) 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 -5,00E+07 nMonths     Cumulated EP 4 EP 4 nMonths EP 4 t -1,00E+08  t 1 -1,50E+08 5. Heat exchanger networks. Time [mo] Piernico Sepiacci – Milan, 21 December 2016 8

Predictive Conceptual Design (PCD) 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 ). Piernico Sepiacci – Milan, 21 December 2016 9

Time series analysis Real data Moving-averaged data Use of moving-averaged values Price series    1 n 1 m Y Moving average allows eliminating most of the high-  t 1 n  i 0 frequency fluctuations and catching the price trend. Time Correlation index   cov Y Y ,   1   t t j (Auto)correlation analysis corr Y Y ,      t t j var Y var Y  t t j (Auto)correlograms report the (auto)correlation index of the   cov X Y ,   0  time series X and Y based on the time lag between them. corr X Y ,     0 1 2 3 4 5 6 7 var X var Y Time lag Autoregressive model      X a a X a X ... a X Identification of the candidate econometric    t 0 1 t 1 2 t 2 q t q model Autoregressive Distributed Lag model          Y a a X a X ... a X bY b Y ... b Y       t 0 1 t 1 2 t 2 q t q 1 t 1 2 t 2 p t p Price series Evaluation of the adaptive parameters nMonths    2  min Y Y It requires a linear regression procedure that minimizes the t t a b ,  t 1 sum of square errors between real and model prices. Time Piernico Sepiacci – Milan, 21 December 2016 10

Econometric models            P a a P a P 1 RAND X   Crude Oil t , 0 1 Crude Oil t , 1 2 Crude Oil t , 2 Crude Oil Crude Oil 140 200 WTI price [USD/bbl] WTI price [USD/bbl] 120 150 100 80 100 60 50 40 20 0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2013 2014 2015 2016 2017 2018            P a a P b P 1 RAND X  Benzene t , 0 1 Crude Oil t , 1 Benzene t , 1 Benzene Benzene 130 180 160 110 Benzene price Benzene price 140 [USD/kmol] [USD/kmol] 90 120 70 100 80 50 60 30 40 10 20 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2013 2014 2015 2016 2017 2018 2 σ Commodity a 0 a 1 b 1 R X 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 Piernico Sepiacci – Milan, 21 December 2016 11

Environmental sustainability Environmental Sustainability by Application of WAR Algorithm ( cp ) ( ep ) I I in in Energy Chemical ( ep ) ( cp ) I I Generation Process we we ( ep ) ( cp ) I I out out Impact Indicators Selection Impact Specific to Chemical k HTPI HTPE TTP ATP  k GWP PCOP AP ODP Environmental Objective Function Piernico Sepiacci – Milan, 21 December 2016 12

The Waste Reduction algorithm The Waste Reduction ( WAR ) algorithm is a tool to determine the potential environmental impact ( PEI ) of a chemical process. Total rate of PEI output ( ep )   I ( tot ) ( cp ) ( ep ) I I I we out out out Waste Energy PEI output from the chemical process   cp  x  ( Energy Generation ( ( cp ) ( out ) ep ) ep ) I M I I Process out j kj k in out j k Mass Mass Energy Supply Where: (out) is the output mass flow rate of stream j ; Inputs Outputs • M j • x kj is the mass fraction of chemical k in stream j . Chemical ( ( ) ) cp cp I I Manufacturing in out PEI output from the energy generation Process   ep-g  x  ( ep ) ( out ) I M out j kj k Waste Energy j k ( cp ) I we Overall PEI for chemical k  ? k Piernico Sepiacci – Milan, 21 December 2016 13