Integrated Process Modeling, Advanced Control and Data Analytics for Optimizing Polyolefin Manufacturing, 2 Volume Set

Y. A. Liu, Alumni Distinguished Professor at Virginia Tech, is an award-winning teacher, author and scholar of sustainable design and industrial practice, and an advisor of global top ten chemical companies. Niket Sharma received his PhD in chemical engineering and M. Eng. in computer science with specialization in machine learning from Virginia Tech in 2021. He is currently a Senior Engineer at Aspen Technology, Boston, where he works on development of machine learning and hybrid modeling applications.

• Volume 1Foreword xviiPreface xxxiiiAcknowledgment xxxviiCopyright Notice xxxixAbout the Authors xliAbout the Companion Website xliii1 Introduction to Integrated Process Modeling, Advanced Control, and Data Analytics in Optimizing Polyolefin Manufacturing 11.1 Segment-Based Modeling of Polymerization Processes: Component Characterization and Polymer Attributes 11.1.1 Component Types in Polymer Process Modeling 11.1.2 Concept of Moments and Some Basic Polymer Attributes 31.1.3 Stream Initialization and Basic Polymer Attributes 51.2 Workshop 1.1: Finding the Resulting Stream Attributes After Mixing Two Copolymer Streams 71.2.1 Objective 71.2.2 Problem Statement 71.2.3 Process Flowsheet 71.2.4 Unit System, Components, and Characterization of Copolymers 71.2.5 Property Method and Property Parameters for Components 91.2.6 Specifications of Streams and Blocks 111.3 Workshop 1.2: A Simplified Simulation Model for a Slurry HDPE Process and the Workflow for Developing a Polymer Process Simulation Model 161.3.1 Objective 161.3.2 Step 1: Problem Setup 161.3.3 Step 2: Component Specifications 181.3.4 Step 3: Property Method 201.3.5 Step 4: Property Parameters – Obtaining Values from Databanks and Estimating Missing Parameters 201.3.6 Step 5: Verification of the Accuracy of the Selected Property Method by Comparing Predicted Pure-Component Property Values with Report Experimental Data 211.3.7 Step 6: Regress Component Liquid Density Data and Binary Vapor– Liquid Equilibrium (TPXY) Data to Estimate Missing Pure-component and Binary Interaction Parameters of Selected Property Method and Verify Predicted VLE Results with Experimental Data 221.3.8 Step 7: Develop Correlations for Polymer Product Quality Indices, Such as Density and Melt Index (Melt Flow Rate) Based on Plant Data 221.3.9 Step 8: Define the Polymerization Reactions and Enter the Initial Reaction Rate Constants 221.3.10 Step 9: Draw the Open-Loop Process Flowsheet and Enter the Inputs for Streams and Blocks 241.3.11 Step 10: Run the Initial Open-loop Process Simulation and Check if the Simulation Results Are Reasonable 251.3.12 Step 11: Close the Recycled Loops, Finalize a Converged Closed-loop Steady-state Simulation Model, and Investigate Applications to Improving Process Operations and Identifying Operating Conditions for New Product Design 261.3.13 Step 12: Convert the Steady-state Simulation Model in Aspen Plus to a Dynamic Simulation Model in Aspen Plus Dynamics; Add Appropriate Controllers; and Investigate Process Operability, Control, and Grade Changes 261.4 Industrial and Potential Applications of Integrated Process Modeling, Advanced Control, and Data Analytics to Optimizing Polyolefin Manufacturing 261.4.1 Industrial and Potential Applications of Process Modeling to Optimizing Polyolefin Manufacturing 261.4.2 Industrial and Potential Applications of Advanced Process Control to Optimizing Polyolefin Manufacturing 281.4.3 Industrial and Potential Applications of Data Analytics to Optimizing Polyolefin Manufacturing 311.4.4 Hybrid Modeling: Integrated Applications of Process Modeling, Advanced Control, and Data Analytics to Optimizing Polyolefin Manufacturing 34References 362 Selection of Property Methods and Estimation of Physical Properties for Polymer Process Modeling 412.1 Property Methods and Thermophysical Parameter Requirements for Process Simulation 412.2 Polymer Activity Coefficient Models (ACM): Polymer Nonrandom Two-liquid (POLYNRTL) Model 422.2.1 Vapor–Liquid Equilibrium for an Ideal Vapor Phase and a Nonideal Liquid Phase 422.2.2 General Vapor–Liquid Equilibrium Relationships Based on Fugacity Coefficient and Liquid-phase Activity Coefficient 432.2.3 Segment-based Mole Fraction Versus Species-based Mole Fraction 432.2.4 POLYNRTL: Polymer Nonrandom Two-liquid Activity Coefficient Model 442.2.5 Concept of Henry Components for Vapor–Liquid Equilibrium for a Vapor Phase and a Nonideal Liquid Phase Involving Supercritical Components 462.3 Workshop 2.1. Estimating POLYNRTL Binary Parameters Using UNIFAC 502.3.1 Objective 502.3.2 Estimating POLYNRTL Binary Parameters Using UNIFAC for Polystyrene Manufacturing 512.4 Prediction of Polymer Physical Properties by Van Krevelen Functional Group Method 532.5 Workshop 2.2. Estimating the Physical Properties of a Copolymer Using the Van Krevelen Group Contribution Method 552.5.1 Objective 552.5.2 Draw the Process Flowsheet and Specify the Unit Set and Global Options 552.5.3 Define Components, Segments, and Polymer and Characterize Their Structures 552.5.4 Choosing Property Method and Entering or Estimating Property Parameters 562.5.5 Specifications of Feed Stream and Flash Block 572.5.6 Creating Property Sets 582.5.7 Defining Property Analysis Run to Create Property Tables 582.6 Polymer Sanchez–Lacombe Equation of State (POLYSL) 612.7 Workshop 2.3. Estimating Property Parameters Using DataRegression Tool 642.7.1 Objective 642.7.2 Defining a DRS Run 642.7.3 Specifying a Unit Set and Global Options 642.7.4 Defining Components, Segments, Oligomers, and Polymer 652.7.5 Choose Property Method and Enter Known Property Parameters from Aspen Enterprise Databanks 662.7.6 Enter Experimental Data for Data Regression, Run the Regression, and Examine the Results 672.7.7 Specifying a Regression Run and the Parameters to be Regressed 692.7.8 Running the Regression Case and Examining the Results 692.8 Polymer Perturbed-chain Statistical Fluid Theory (POLYPCSF) Equation of State 722.9 Workshop 2.4. Regression of Property Parameters for POLYPCSF EOS 742.9.1 Objective and Data Sources 742.9.2 Regression of Pure Component Parameters for POLYPCSF EOS 752.10 Correlation of Polymer Product Quality Indices and Structure–Property Correlations 772.10.1 Polyolefin Product Quality Indices 772.10.2 Empirical Correlations of Polymer Product Quality Targets 802.10.3 Estimation of Apparent Newtonian Viscosity from MI-MWW Measurement 81References 833 Reactor Modeling, Convergence Tips, and Data-Fit Tool 873.1 Kinetic or Rate-Based Reactors 873.2 Continuous Stirred-Tank Reactor Model (RCSTR) 873.2.1 RCSTR Configurations 873.2.2 RCSTR Specifications 883.3 Plug-Flow Reactor Model (RPLUG) 893.3.1 RPLUG Configurations 893.3.2 RPLUG Specifications 903.4 Batch Reactor Model (RBATCH) 913.4.1 RBATCH Configuration 913.4.2 RBATCH Specifications 923.5 Representation of Nonideal Reactors 933.6 RCSTR Convergence 933.6.1 Initialization 933.6.2 Scaling Factors 953.6.3 Residence Time Loop 953.6.4 Energy Balance Loop 963.6.5 Mass Balance Loop 973.6.6 Flash Loop 983.6.7 Recommendation for RCSTR Mass Balance Algorithm for Polyolefin Process Simulation 993.7 RPLUG/RBATCH Model Convergence 1003.8 Data Fit (Simulation Data Regression) 1013.9 Workshop 3.1: Data Fit of Kinetic Parameters for Styrene Polymerization Using Concentration Profile Data 1033.9.1 Objective 1033.9.2 A Simplified Kinetic Model for Styrene Polymerization 1033.9.3 Datasets 1063.9.4 Simulation Data Regression (Data Fit) 1073.10 Workshop 3.2: Data Fit of Kinetic Parameters for Styrene Polymerization Using Point Data 1113.10.1 Objective 1113.10.2 Dataset 1113.10.3 Simulation Data Regression (Data Fit) 112References 1144 Free Radical Polymerizations: LDPE and EVA 1154.1 Polymers by Free Radical Polymerization 1154.2 Kinetics of Free Radical Polymerization 1154.2.1 Initiator and Its Decomposition-Rate Parameters 1164.2.2 Chain Initiation Reactions 1184.2.3 Chain Propagation Reactions 1194.2.4 Chain Transfer Reactions 1204.2.5 Termination Reactions 1214.2.6 Autoacceleration, Trommsdorff Effect, or Gel Effect 1224.2.7 Other Free Radical Polymerization Reactions 1234.3 Thermodynamic Methods and Property Parameter Requirements 1234.4 Workshop 4.1: Simulation of an Autoclave High-pressure LDPE Process 1244.4.1 Objectives 1244.4.2 Process Flowsheet and Simulation Representation 1244.4.3 Unit System, Components, and Characterization of Polymer 1264.4.4 Thermodynamic Methods and Property Parameters for Components, Segment, and Polymer 1294.4.5 PCES (Physical Constant Estimation System) for Estimating Missing-Property Parameters 1304.4.6 Defining Free Radical Polymerization Reactions for LDPE 1304.4.7 Specifications of Inlet Process Streams and Unit Operation and Reactor Blocks 1334.4.8 Methodology for Improving Simulation Convergence and for Kinetic Parameter Estimation 1334.4.9 Base-Case Simulation Results 1364.4.10 Model Applications 1384.4.11 Separation Section 1394.5 Workshop 4.2: Simulation of Tubular Reactors for HP LDPE Process 1404.5.1 Objectives 1404.5.2 Process Flowsheet and Simulation Representation 1414.5.3 Unit System, Components, and Characterization of Polymer 1414.5.4 Thermodynamic Method and Property Parameters for Components 1424.5.5 PCES (Physical Constant Estimation System) for Estimating Missing-Property Parameters 1444.5.6 Free Radical Polymerization Reactions for LDPE 1444.5.7 Specifications of Inlet Process Streams and Unit Operation and Reactor Blocks 1444.5.8 User FORTRAN Subroutine for Heat Transfer Calculations for the LDPE Reactor 1454.5.9 Base-Case Simulation Targets and Kinetic Parameter Estimation 1474.5.10 Model Applications 1494.6 Workshop 4.3: Simulation of Tubular Reactors for Ethylene–Vinyl Acetate (EVA) Copolymerization Process 1514.6.1 Objective 1514.6.2 Process Background 1514.6.3 Unit System, Components, and Characterization of Polymer 1534.6.4 Thermodynamic Method and Property Parameters for Components and Polymer 1554.6.5 Free Radical Polymerization Kinetics for EVA Copolymerization 1564.6.6 Specifications of Inlet Process Streams and Unit Operation and Reactor Blocks 1564.6.7 Base-Case Simulation Targets and Kinetic Parameter Estimation 158References 1605 Ziegler–Natta Polymerization: HDPE, PP, LLDPE, and EPDM 1635.1 Ziegler–Natta (ZN) Polymerization 1645.1.1 Introduction 1645.1.2 Ziegler–Natta Catalysts 1645.2 Ziegler–Natta Polymerization Kinetics 1655.2.1 Catalyst Activation (ACT) 1655.2.2 Chain Initiation (CHAIN-INI) 1665.2.3 Chain Propagation (PROP) 1665.2.4 Chain-Transfer Reaction (CHAT) 1675.2.5 Catalyst Deactivation (DEACT) 1675.2.6 Catalyst Inhibition (INH) 1675.2.7 Copolymerization Kinetics 1685.3 Modeling Considerations 1705.3.1 Reactor Types 1705.3.2 Process Flowsheets 1715.3.3 Polymer Types 1725.3.4 Molecular Weight Distribution (MWD) and Multi-Modal Distributions 1735.3.5 Thermodynamics 1745.3.6 Global Kinetics Versus Local Kinetics 1745.4 Commercial Polyolefin Production Targets 1755.4.1 General Production Targets 1755.4.1.1 Production Rate 1755.4.1.2 MWN 1755.4.1.3 MI 1765.4.1.4 Conversion 1765.4.1.5 PDI 1765.4.1.6 SMWN and SPFRAC 1765.4.1.7 SFRAC and SCB 1765.4.1.8 Rho 1765.4.1.9 Residence Time 1775.4.2 Polymer-Specific Targets 1775.4.2.1 CISFRAC 1775.4.2.2 ATFRAC 1775.5 Methodology for Polyolefin Kinetic Estimation 1785.5.1 Efficient Use of Software Tool: Data Fit 1795.5.2 Flowchart of the Methodology for Kinetic Parameter Estimation 1795.5.2.1 Multiple Product Grades and Single Active Catalyst Site 1805.5.2.2 Multisite Model and Deconvolution Analysis 1835.5.2.3 GPC Data and Deconvolution Analysis to Estimate the Number of Active Catalyst Sites 1855.5.3 Efficient Use of Software Tools: Sensitivity Analysis 1885.5.4 Efficient Use of Software Tools: Design Specification 1915.5.5 Model Applications 1945.6 Workshop 5.1: Simulation of a Slurry HDPE Process 1955.6.1 Objective 1955.6.2 Process Flowsheet 1955.6.3 Unit System, Components, and Characterization of Oligomer, Polymer, and Site-Based Species 1955.6.4 The Role of Solid Polymer in Phase-Equilibrium Calculations 1995.6.5 Thermodynamic Model and Parameters 1995.6.6 Pure-Component Parameters 2005.6.7 Feed Streams 2035.6.8 Ziegler–Natta Kinetics Specifications 2045.6.9 Specifications of Unit Operations and Chemical Reactor Blocks 2075.6.9.1 Mixers (Figure 5.38) 2075.6.9.2 Reactors (Figures 5.39–5.42) 2075.6.9.3 Specification of Flash Drums (Figure 5.42) 2085.6.10 Simulation Results 2095.6.11 Sensitivity Analysis 2095.6.12 Closing the Recycle Loops 2105.7 Workshop 5.2: Simulation of Stirred-Bed Gas-Phase PP Process 2145.7.1 Objective 2145.7.2 Process Description 2145.7.3 Modeling the Stirred-Bed Reactor 2155.7.4 Process Flowsheet 2185.7.5 Unit System, Components, and Characterization of Polymer and Site-Based Species 2185.7.6 Thermodynamic Model and Parameters 2205.7.7 Feed Streams 2225.7.8 Ziegler–Natta Kinetics Specifications 2235.7.9 Specifications of Unit Operation and Chemical Reactor Blocks 2255.7.9.1 Mixers MIX1 to MIX8 (Figure 5.62) 2255.7.9.2 Reactors R1 to R8 (Figures 5.63 and 5.64) 2255.7.9.3 Other Blocks 2255.7.9.4 Convergence Blocks 2265.7.10 Open-Loop Simulation Results and Closing the Loop 2275.7.11 Model Applications 2295.8 Workshop 5.3: Simulation of a Gas-Phase Fluid-Bed LLDPE Process with Condensed Mode Cooling 2295.8.1 Objective 2295.8.2 Condensed Mode Cooling in Ethylene Polymerization in a Fluidized-Bed Reactor 2305.8.3 Process Flowsheet 2345.8.4 Unit System, Components, and Characterization of Oligomer, Polymer, and Site-Based Species 2345.8.5 Deconvolution Analysis of GPC Data to Determine the Number of Active Catalyst Sites 2355.8.6 Thermodynamic Model and Parameters 2375.8.7 Inlet Stream Specifications for Grades A and B 2385.8.8 Specifications of Unit Operation and Chemical Reactor Blocks 2385.8.9 Ziegler–Natta Kinetics Specifications 2415.8.10 Reactor and Flowsheet Simulation to Match Plant Production Targets 2425.8.11 Model Applications 2425.9 Workshop 5.4: Simulation of a Solution Polymerization Process for Producing Ethylene–Propylene Copolymer (EPM) or anEthylene–Propylene–Diene Terpolymer (EPDM) with Metallocene Catalysts 2475.9.1 Objective 2475.9.2 Process Background 2475.9.3 EPM Copolymerization Kinetics and EPDM Terpolymerization Using a Metallocene Catalyst System 2495.9.4 Unit System, Components, and Characterization of Polymer 2505.9.5 Thermodynamic Method and Property Parameters for Components and Polymer 2545.9.6 Process Flowsheet and Inlet Stream and Block Specifications 2555.9.7 Base-Case Simulation Results 2565.9.8 Extension to EPDM (Ethylene–Propylene–Diene Terpolymer) 2575.10 Conclusions 260References 2616 Free Radical and Ionic Polymerizations: PS and SBS Rubber 2676.1 Workshop 6.1: Simulation of Polystyrene Reactors with Gel Effect andOligomer Formation 2686.1.1 Objective 2686.1.2 Process Flowsheet 2696.1.3 Unit System, Components, and Characterization of Polymer 2706.1.4 Characterization of Oligomers 2716.1.5 Thermodynamic Method and Property Parameters for Components and Oligomers 2766.1.6 PCES (Physical Constant Estimation System) for Estimating Property Parameters for Oligomers 2786.1.7 Defining Free Radical Reactions and Oligomer Reactions 2796.1.8 Specification of Inlet Process Streams and Unit Operation and Reactor Blocks 2846.1.9 Kinetic Parameter Estimation and Model Validation 2866.1.10 Model Applications 2886.2 Workshop 6.2: Production of Poly(Styrene–Butadiene–Styrene) or SBS Rubber by Ionic Polymerization 2926.2.1 Motivation and Objective for Modeling Ionic Polymerization Processes 2926.2.2 Reactor Configurations and Copolymer Products 2936.2.2.1 Tapered Block Copolymer 2936.2.2.2 Di-/Tri-Block Copolymer and a Star-Shaped Block Copolymer 2946.2.2.3 Random Copolymer 2946.2.3 Components, Segments, and Polymer in Anionic Copolymerization of Styrene and Butadiene 2946.2.4 Thermodynamic Method and Property Parameters of Components and Polymer 2946.2.5 Kinetics of Anionic Copolymerization of Styrene and Butadiene 2956.2.5.1 Initiator Disassociation (INIT-DISSOC) 2956.2.5.2 Chain Initiation (CHAIN-INI) 2966.2.5.3 Chain Propagation (PROPAGATION) 3006.2.5.4 Association or Aggregation (ASSOCIATION) 3006.2.5.5 Chain Transfer (CHAT) 3016.2.5.6 Chain Termination (TERM-AGENT) 3026.2.5.7 Equilibrium with Counter-Ion or Reversible Ionization (EQUILIB-CION) 3026.2.5.8 Batch Reactor for Producing a Tapered Block Copolymer 3026.2.5.9 Semi-Batch Reactor for Producing a Tri-Block SBS Copolymer by an Industrial Batch-Sequence Recipe 3066.2.5.10 Semi-Batch Reactor for Producing a Tri-Block SBS Copolymer by a Literature Batch-Sequence Recipe 313References 3187 Improved Polymer Process Operability and Control Through Steady-State and Dynamic Simulation Models 3217.1 Workshop 7.1: Workflow for Dynamic Process Modeling Using Aspen Plus and Aspen Plus Dynamics 3227.2 Running Simulation in Aspen Plus Dynamics 3257.2.1 Types of Dynamic Simulations: Flow-Driven and Pressure-Driven 3257.2.2 Graphical Interface of Aspen Plus Dynamics 3257.2.3 Simulation Run Modes and Run Control 3267.2.4 Viewing Simulation Results Using Predefined Tables and Plots 3287.2.5 Specification Status and Analysis 3297.2.6 Creating New Tables and Plots 3327.2.7 Variable Finding in Aspen Plus Dynamics (AD) 3347.3 Process Control in Aspen Plus Dynamics 3357.3.1 Workshop 7.2: Adding a PID Controller 3357.3.2 Configuring a PID Controller 3377.4 Snapshots 3447.5 Workshop 7.3: Tasks for Implementing Discrete Events 3447.6 Workshop 7.4: Dynamic Simulation and Grade Change of a Slurry HPDE Process 3507.6.1 Objectives 3507.6.2 Stepwise Procedure to Develop Aspen Plus Dynamics (AD) Simulation Model 3507.6.3 Simulation of Grade-Change Operations 3557.7 Workshop 7.5: Dynamic Simulation and Control of a Commercial Slurry HDPE Process 3597.7.1 Objectives 3597.7.2 Converting a Steady-State Simulation Model to a Dynamic Simulation Model 3597.7.3 Initial Adjustments of the AD Model 3617.7.3.1 Polymer Attributes for Streams and Blocks 3617.7.3.2 Implementation of Reactor Level Control Using Mechanical Weir 3617.7.3.3 Improvement of the Reactor Temperature Controller 3627.7.3.4 Deletion of Pressure Controllers 3637.7.3.5 Adding a Hydrogen–Ethylene Ratio Controller to the Recycle Gas 3637.8 Workshop 7.6: Dynamic Simulation and Control of a Gas-Phase Fluidized-Bed Process for Producing LLDPE in Condensed Mode Operation 3677.8.1 Objectives 3677.8.2 Converting a Steady-State Simulation Model to a Dynamic Simulation Model 3677.9 Workshop 7.7: Dynamic Simulation and Control of a Slurry HDPE Process Using an Inferential Controller 3707.9.1 Objective 3707.9.2 Inferential Control Theory and Recent Applications 3707.9.3 HDPE Process Description and Steady-State Model Empirical Correlation 3727.9.4 Grade-Change Transition Using Basic H2-Based Controller 3737.9.5 Open-Loop Inferential Controller Using Dynamic Model 3747.9.6 Closed-Loop Inferential Controller 375References 378Volume 2Foreword xvPreface xxxiAcknowledgment xxxvCopyright Notice xxxviiAbout the Authors xxxixAbout the Companion Website xli8 Model-Predictive Control of Polyolefin Processes 3819 Application of Multivariate Statistics to Optimizing Polyolefin Manufacturing 47710 Applications of Machine Learning to Optimizing Polyolefin Manufacturing 53311 A Hybrid Science-Guided Machine Learning Approach for Modeling Chemical and Polymer Processes 651Appendix A Matrix Algebra in Multivariate Data Analysis and Model Predictive Control 699Appendix B Introduction to Python for Chemical Engineers 737Aman AggarwalIndex 759