Multivariable Predictive Control

Sandip Kumar Lahiri, PhD, is a chemical engineer with more than twenty one years of experience in operations and technical services at leading petrochemical industries around the globe. His areas of expertise include simulation, process modelling, artificial intelligence and neural networks in process industry, APC, soft sensor, and slurry flow modelling.

• Figure List xixTable List xxiPreface xxiii1 Introduction of Model Predictive Control 11.1 Purpose of Process Control in Chemical Process Industries (CPI) 11.2 Shortcomings of Simple Regulatory PID Control 21.3 What Is Multivariable Model Predictive Control? 31.4 Why Is a Multivariable Model Predictive Optimizing Controller Necessary? 41.5 Relevance of Multivariable Predictive Control (MPC) in Chemical Process Industry in Today’s Business Environment 61.6 Position of MPC in Control Hierarchy 61.6.1 Regulatory PID Control Layer 61.6.2 Advance Regulatory Control (ARC) Layer 81.6.3 Multivariable Model‐Based Control 81.6.4 Economic Optimization Layer 81.6.4.1 First Layer of Optimization 81.6.4.2 Second Layer of Optimization 91.6.4.3 Third Layer of Optimization 91.7 Advantage of Implementing MPC 101.8 How Does MPC Extract Benefit? 131.8.1 MPC Inherent Stabilization Effect 131.8.2 Process Interactions 141.8.3 Multiple Constraints 151.8.4 Intangible Benefits of MPC 171.9 Application of MPC in Oil Refinery, Petrochemical, Fertilizer, and Chemical Plants, and Related Benefits 172 Theoretical Base of MPC 232.1 Why MPC? 232.2 Variables Used in MPC 252.2.1 Manipulated Variables (MVs) 252.2.2 Controlled Variables (CVs) 252.2.3 Disturbance Variables (DVs) 252.3 Features of MPC 262.3.1 MPC Is a Multivariable Controller 262.3.2 MPC Is a Model Predictive Controller 262.3.3 MPC Is a Constrained Controller 262.3.4 MPC Is an Optimizing Controller 272.3.5 MPC Is a Rigorous Controller 272.4 Brief Introduction to Model Predictive Control Techniques 272.4.1 Simplified Dynamic Control Strategy of MPC 282.4.2 Step 1: Read Process Input and Output 292.4.3 Step 2: Prediction of CVs 302.4.3.1 Building Dynamic Process Model 302.4.3.2 How MPC Predicts the Future 322.4.4 Step 3: Model Reconciliation 332.4.5 Step 4: Determine the Size of the Control Process 342.4.6 Step 5: Removal of Ill‐Conditioned Problems 342.4.7 Step 6: Optimum Steady‐State Targets 352.4.8 Step 7: Develop Detailed Plan of MV Movement 363 Historical Development of Different MPC Technology 433.1 History of MPC Technology 433.1.1 Pre‐Era 433.1.1.1 Developer 433.1.1.2 Motivation 443.1.1.3 Limitations 443.1.2 First Generation of MPC (1970–1980) 443.1.2.1 Characteristics of First‐Generation MPC Technology 443.1.2.2 IDCOM Algorithm and Its Features 453.1.2.3 DMC Algorithm and Its Features 463.1.3 Second‐Generation MPC (1980–1985) 463.1.4 Third‐Generation MPC (1985–1990) 473.1.4.1 Distinguishing Features of Third‐Generation MPC Algorithm 483.1.4.2 Distinguishing Features of the IDCOM‐M Algorithm 493.1.4.3 Evolution of SMOC 503.1.4.4 Distinctive Features of SMOC 503.1.5 Fourth‐Generation MPC (1990–2000) 503.1.5.1 Distinctive Features of Fourth‐Generation MPC 513.1.6 Fifth‐Generation MPC (2000–2015) 513.2 Points to Consider While Selecting an MPC 524 MPC Implementation Steps 554.1 Implementing a MPC Controller 554.1.1 Step 1: Preliminary Cost–Benefit Analysis 554.1.2 Step 2: Assessment of Base Control Loops 554.1.3 Step 3: Functional Design of Controller 564.1.4 Step 4: Conduct the Preliminary Plant Test (Pre‐Stepping) 574.1.5 Step 5: Conduct the Plant Step Test 574.1.6 Step 6: Identify a Process Model 574.1.7 Step 7: Generate Online Soft Sensors or Virtual Sensors 584.1.8 Step 8: Perform Offline Controller Simulation/Tuning 584.1.9 Step 9: Commission the Online Controller 584.1.10 Step 10: Online MPC Controller Tuning 594.1.11 Step 11: Hold Formal Operator Training 594.1.12 Step 12: Performance Monitoring of MPC Controller 594.1.13 Step 13: Maintain the MPC Controller 604.2 Summary of Steps Involved in MPC Projects with Vendor 605 Cost–Benefit Analysis of MPC before Implementation 635.1 Purpose of Cost–Benefit Analysis of MPC before Implementation 635.2 Overview of Cost–Benefit Analysis Procedure 645.3 Detailed Benefit Estimation Procedures 655.3.1 Initial Screening for Suitability of Process to Implement MPC 655.3.2 Process Analysis and Economics Analysis 665.3.3 Understand the Constraints 675.3.4 Identify Qualitatively Potential Area of Opportunities 675.3.4.1 Example 1: Air Separation Plant 685.3.4.2 Example 2: Distillation Columns 695.3.5 Collect All Relevant Plant and Economic Data (Trends, Records) 695.3.6 Calculate the Standard Deviation and Define the Limit 695.3.7 Estimate the Stabilizing Effect of MPC and Shift in the Average 705.3.7.1 Benefit Estimation: When the Constraint Is Known 715.3.7.2 Benefit Estimation: When the Constraint Is Not Well Known or Changing 725.3.8 Estimate Change in Key Performance Parameters Such as Yield, Throughput, and Energy Consumption 725.3.8.1 Example: Ethylene Oxide Reactor 725.3.9 Identify How This Effect Translates to Plant Profit Margin 735.3.10 Estimate the Economic Value of the Effect 735.4 Case Studies 735.4.1 Case Study 1 735.4.1.1 Benefit Estimation Procedure 735.4.2 Case Study 2 745.4.2.1 Benefit Estimation Procedure 746 Assessment of Regulatory Base Control Layer in Plants 776.1 Failure Mode of Control Loops and Their Remedies 776.2 Control Valve Problems 776.2.1 Improper Valve Sizing 786.2.1.1 How to Detect a Particular Control Valve Sizing Problem 786.2.2 Valve Stiction 796.2.2.1 What Is Control Valve Stiction? 796.2.2.2 How to Detect Control Valve Stiction Online 806.2.2.3 Combating Stiction 806.2.2.4 Techniques for Combating Stiction Online 806.2.3 Valve Hysteresis and Backlash 816.3 Sensor Problems 826.3.1 Noisy 826.3.2 Flatlining 826.3.3 Scale/Range 826.3.4 Calibration 826.3.5 Overfiltered 836.4 Controller Problems 836.4.1 Poor Tuning and Lack of Maintenance 836.4.2 Poor or Missing Feedforward Compensation 836.4.3 Inappropriate Control Structure 846.5 Process‐Related Problems 846.5.1 Problems of Variable Gain 846.5.2 Oscillations 846.5.2.1 Variable Valve Gain 856.5.2.2 Variable Process Gain 856.6 Human Factor 856.7 Control Performance Assessment/Monitoring 866.7.1 Available Software for Control Performance Monitoring 866.7.2 Basic Assessment Procedure 876.8 Commonly Used Control System Performance KPIs 876.8.1 Traditional Indices 886.8.1.1 Peak Overshoot Ratio (POR) 886.8.1.2 Decay Rate 886.8.1.3 Peak Time and Rise Time 886.8.1.4 Settling Time 886.8.1.5 Integral of Error Indexes 886.8.2 Simple Statistical Indices 886.8.2.1 Mean of Control Error (%) 896.8.2.2 Standard Deviation of Control Error (%) 896.8.2.3 Standard Variation of Control Error (%) 896.8.2.4 Standard Deviation of Controller Output (%) 896.8.2.5 Skewness of Control Error 896.8.2.6 Kurtosis of Control Error 896.8.2.7 Ratio of Standard of Control Error and Controller Output 896.8.2.8 Maximum Bicoherence 906.8.3 Business/Operational Metrics 906.8.3.1 Loop Health 906.8.3.2 Service Factor 906.8.3.3 Key Performance Indicators 906.8.3.4 Operational Performance Efficiency Factor 906.8.3.5 Overall Loop Performance Index 906.8.3.6 Controller Output Changes in Manual 906.8.3.7 Mode Changes 906.8.3.8 Totalized Valve Reversals and Valve Travel 906.8.3.9 Process Model Parameters 906.8.4 Advanced Indices 906.8.4.1 Harris Index 916.8.4.2 Nonlinearity Index 916.8.4.3 Oscillation‐Detection Indices 916.8.4.4 Disturbance Detection Indices 926.8.4.5 Autocorrelation Indices 926.9 Tuning for PID Controllers 926.9.1 Complications with Tuning PID Controllers 936.9.2 Loop Retuning 936.9.3 Classical Controller Tuning Algorithms 946.9.3.1 Controller Tuning Methods 946.9.3.2 Ziegler‐Nichols Tuning Method 946.9.3.3 Dahlin (Lambda) Tuning Method 946.9.4 Manual Controller Tuning Methods in Absence of Any Software 956.9.4.1 Pre‐Tuning 956.9.4.2 Bring in Baseline Parameters 976.9.4.3 Some Like It Simple 976.9.4.4 Tuning Cascade Control 987 Functional Design of MPC Controllers 1017.1 What Is Functional Design? 1017.2 Steps in Functional Design 1027.2.1 Step 1: Define Process Control Objectives 1027.2.1.1 Economic Objectives 1027.2.1.2 Operating Objectives 1037.2.1.3 Control Objectives 1047.2.2 Step 2: Identify Process Constraints 1047.2.2.1 Process Limitations 1047.2.2.2 Safety Limitations 1047.2.2.3 Process Instrument Limitations 1057.2.2.4 Raw Material and Utility Supply Limitation 1057.2.2.5 Product Limitations 1057.2.3 Step 3: Define Controller Scope 1057.2.4 Step 4: Select the Variables 1067.2.4.1 Economics of the Unit 1067.2.4.2 Constraints of the Unit 1077.2.4.3 Control of the Unit 1077.2.4.4 Manipulated Variables (MVs) 1077.2.4.5 Controlled Variables (CVs) 1077.2.4.6 Disturbance Variables (DVs) 1087.2.4.7 Practical Guidelines for Variable Selections 1087.2.5 Step 5: Rectify Regulatory Control Issues 1097.2.5.1 Practical Guidelines for Changing Regulatory Controller Strategy 1097.2.6 Step 6: Explore the Scope of Inclusions of Inferential Calculations 1107.2.7 Step 7: Evaluate Potential Optimization Opportunity 1107.2.7.1 Practical Guidelines for Finding out Optimization Opportunities 1117.2.8 Step 8: Define LP or QP Objective Function 1117.2.8.1 CDU Example 1128 Preliminary Process Test and Step Test 1138.1 Pre‐Stepping, or Preliminary Process Test 1138.1.1 What Is Pre‐Stepping? 1138.1.2 Objective of Pre‐Stepping 1138.1.3 Prerequisites of Pre‐Stepping 1138.1.4 Pre‐Stepping 1148.2 Step Testing 1158.2.1 What Is a Step Test? 1158.2.2 What Is the Purpose of a Step Test? 1158.2.3 Details of Step Testing 1168.2.3.1 Administrative Aspects 1168.2.3.2 Technical Aspects 1168.2.4 Different Step‐Testing Method 1178.2.4.1 Manual Step Testing 1178.2.4.2 PRBS (Pseudo Random Binary Sequence) 1178.2.4.3 General Guidelines of PRBS Test 1178.2.5 Difference between Normal Step Testing and PRBS Testing 1188.2.6 Which One to Choose? 1188.2.7 Dos and Don’ts of Step Testing 1188.3 Development of Step‐Testing Methodology over the Years 1209 Model Building and System Identification 1239.1 Introduction to Model Building 1239.2 Key Issues in Model Identifications 1249.2.1 Identification Test 1249.2.2 Model Structure and Parameter Estimation 1259.2.3 Order Selection 1269.2.4 Model Validation 1279.3 The Basic Steps of System Identification 1279.3.1 Step 0: Experimental Design and Execution 1289.3.2 Step 1: Plan the Case that Needs to Be Modeled 1309.3.2.1 Action 1 1309.3.2.2 Action 2 1309.3.3 Step 2: Identify Good Slices of Data 1309.3.3.1 Looking at the Data 1319.3.4 Step 3: Pre‐Processing of Data 1319.3.5 Step 4: Identification of Model Curve 1329.3.5.1 Hybrid Approach to System Identification 1329.3.5.2 Direct Modeling Approach of System Identification 1339.3.5.3 Subspace Identification 1349.3.5.4 Detailed Steps of Implementations 1359.3.6 Step 5: Select Final Model 1369.4 Model Structures 1379.4.1 FIR Models 1389.4.1.1 FIR Structures 1389.4.2 Prediction Error Models (PEM Models) 1399.4.2.1 PEM Structures 1399.4.3 Model for Order and Variance Reduction 1409.4.3.1 ARX Parametric Models (Discrete Time) 1409.4.3.2 Output Error Models (Discrete Time) 1409.4.3.3 Laplace Domain Parametric Models 1419.4.3.4 Final Model Form 1419.4.4 State‐Space Models 1419.4.5 How to Know Which Structure and Method to Use 1429.5 Common Features of Commercial Identification Packages 14210 Soft Sensors 14510.1 What Is a Soft Sensor? 14510.2 Why Soft Sensors Are Necessary 14510.2.1 Process Monitoring and Process Fault Detection 14610.2.2 Sensor Fault Detection and Reconstruction 14610.2.3 Use of Soft Sensors in MPC Application 14610.3 Types of Soft Sensors 14710.3.1 First Principle‐Based Soft Sensors 14710.3.1.1 Advantages 14710.3.1.2 Disadvantages 14710.3.2 Data‐Driven Soft Sensors 14810.3.2.1 Advantages 14810.3.2.2 Disadvantages 14810.3.3 Gray Model‐Based Soft Sensors 14810.3.3.1 Advantages 14910.3.4 Hybrid Model‐Based Soft Sensors 14910.3.4.1 Advantages 14910.4 Soft Sensors Development Methodology 14910.4.1 Data Collection and Data Inspection 14910.4.2 Data Preprocessing and Data Conditioning 15010.4.2.1 Outlier Detection and Replacement 15110.4.2.2 Univariate Approach to Detect Outliers 15110.4.2.3 Multivariate Approach to Detect Outliers (Lin 2007) 15110.4.2.4 Handling of Missing Data 15210.4.3 Selection of Relevant Input Output Variables 15310.4.4 Data Alignment 15310.4.5 Model Selection, Training, and Validation (Kadlec 2009; Lin 2007) 15310.4.6 Analyze Process Dynamics 15410.4.7 Deployment and Maintenance 15510.5 Data‐Driven Methods for Soft Sensing 15610.5.1 Principle Component Analysis 15610.5.1.1 The Basics of PCA 15610.5.1.2 Why Do We Need to Rotate the Data? 15610.5.1.3 How Do We Generate Principal Components? 15610.5.1.4 Steps to Calculating Principal Components 15710.5.2 Partial Least Squares 15710.5.3 Artificial Neural Networks 15810.5.3.1 Network Architecture 15910.5.3.2 Back Propagation Algorithm (BPA) 15910.5.4 Neuro‐Fuzzy Systems 16010.5.5 Support Vector Machines 16110.5.5.1 Support Vector Regression–Based Modeling 16110.6 Open Issues and Future Steps of Soft Sensor Development 16210.6.1 Large Effort Required for Preprocessing of Industrial Data 16210.6.2 Which Modeling Method to Choose? 16310.6.3 Agreement of the Developed Model with Physics of the Process 16310.6.4 Performance Deterioration of Developed Soft Sensor Model 16311 Offline Simulation 16711.1 What Is Offline Simulation? 16711.2 Purpose of Offline Simulation 16711.3 Main Task of Offline Simulation 16811.4 Understanding Different Tuning Parameters of Offline Simulations 16811.4.1 Tuning Parameters for CVs 16911.4.1.1 Methods for Handling of Infeasibility 17011.4.1.2 Priority Ranking of CVs 17011.4.1.3 cv Give‐Up 17011.4.1.4 cv Error Weight 17011.4.2 Tuning Parameters for MVs 17111.4.2.1 mv Maximum Movement Limits or Rate‐of‐Change Limits 17111.4.2.2 Movement Weights 17111.4.3 Tuning Parameters for Optimizer 17211.4.3.1 Economic Optimization 17211.4.3.2 General Form of Objective Function 17311.4.3.3 Weighting Coefficients 17311.4.3.4 Setting Linear Objective Coefficients 17311.4.3.5 Optimization Horizon and Optimization Speed Factor 17411.4.3.6 Optimization Speed Factor 17411.4.3.7 mv Optimization Priority 17411.4.4 Soft Limits 17511.4.4.1 How Soft Limits Work 17511.4.4.2 cv Soft Limits 17511.4.4.3 mv Soft Limits 17611.5 Different Steps to Build and Activate Simulator in an Offline PC 17611.6 Example of Tests Carried out in Simulator 17711.6.1 Control and Optimization Objectives 17711.6.1.1 Test 1 17811.6.1.2 Test 2 17911.6.1.3 Test 3 17911.6.1.4 Test 4 18011.6.1.5 Test 5 18011.6.1.6 Test 6 18011.6.1.7 Others Tests 18111.7 Guidelines for Choosing Tuning Parameters 18111.7.1 Guidelines for Choosing Initial Values 18111.7.2 How to Select Maximum Move Size and MV Movement Weights During Simulation Study 18212 Online Deployment of MPC Application in Real Plants 18312.1 What Is Online Deployment (Controller Commissioning)? 18312.2 Steps for Controller Commissioning 18312.2.1 Set up the Controller Configuration and Final Review of the Model 18312.2.2 Build the Controller 18412.2.3 Load Operator Station on PC Near the Panel Operator 18412.2.4 Take MPC Controller in Line with Prediction Mode 18612.2.5 Put the MPC Controller in Close Loop with One CV at a Time 18712.2.6 Observe MPC Controller Performance 18712.2.7 Put Optimizer in Line and Observe Optimizer Performance 18912.2.8 Evaluate Overall Controller Performance 18912.2.9 Perform Online Tuning and Troubleshooting 19012.2.10 Train Operators and Engineers on Online Platform 19012.2.11 Document MPC Features 19012.2.12 Maintain the MPC Controller 19113 Online Controller Tuning 19313.1 What Is Online MPC Controller Tuning? 19313.2 Basics of Online Tuning 19313.2.1 Key Checkout Regarding Controller Performance 19313.2.2 Steps to Troubleshoot the Problem 19413.3 Guidelines to Choose Different Tuning Parameters 19514 Why Do Some MPC Applications Fail? 19914.1 What Went Wrong? 19914.2 Failure to Build Efficient MPC Application 20114.2.1 Historical Perspective 20114.2.2 Capability of MPC Software to Capture Benefits 20214.2.3 Expertise of Implementation Team 20214.2.3.1 MPC Vendor Limitations 20314.2.3.2 Client Limitations 20414.2.4 Reliability of APC Project Methodology 20414.3 Contributing Failure Factors of Postimplementation MPC Application 20514.3.1 Technical Failure Factors 20614.3.1.1 Lack of Performance Monitoring of MPC Application 20614.3.1.2 Unresolved Basic Control Problems 20614.3.1.3 Poor Tuning and Degraded Model Quality 20714.3.1.4 Problems Related to Controller Design 20714.3.1.5 Significant Process Modifications and Enhancement 20714.3.2 Nontechnical Failure Factors 20814.3.2.1 Lack of Properly Trained Personnel 20814.3.2.2 Lack of Standards and Guidelines to MPC Support Personnel 20814.3.2.3 Lack of Organizational Collaboration and Alignment 20814.3.2.4 Poor Management of Control System 20914.4 Strategies to Avoid MPC Failures 21014.4.1 Technical Solutions 21114.4.1.1 Development of Online Performance Monitoring of APC Applications 21114.4.1.2 Improvement of Base Control Layer 21214.4.1.3 Tuning Basic Controls 21214.4.1.4 Control Performance Monitoring Software 21314.4.2 Management Solutions 21414.4.2.1 Training of MPC Console Operators 21414.4.2.2 Training of MPC Control Engineers 21514.4.2.3 Development of Corporate MPC Standards and Guidelines 21614.4.2.4 Central Engineering Support Organization for MPC 21714.4.3 Outsourcing Solutions 21915 MPC Performance Monitoring 22115.1 Why Performance Assessment of MPC Application Is Necessary 22115.2 Types of Performance Assessment 22215.2.1 Control Performance 22215.2.2 Optimization Performance 22215.2.3 Economic Performance 22215.2.4 Intangible Performance 22215.3 Benefit Measurement after MPC Implementation 22215.4 Parameters to Be Monitored for MPC Performance Evaluation 22315.4.1 Service Factors 22415.4.2 KPI for Financial Criteria 22415.4.3 KPI for Standard Deviation of Key Process Variable 22515.4.3.1 Safety Parameters 22515.4.3.2 Quality Giveaway Parameters 22515.4.3.3 Economic Parameters 22515.4.4 KPI for Constraint Activity 22615.4.5 KPI for Constraint Violation 22615.4.6 KPI for Inferential Model Monitoring 22615.4.7 Model Quality 22615.4.8 Limit Change Frequencies for CV/MVs 22715.4.9 Active MV Limit 22715.4.10 Long‐Term Performance Monitoring of MPC 22715.5 KPIs to Troubleshoot Poor Performance of Multivariable Controls 22815.5.1 Supporting KPIs for Low Service Factor 22815.5.2 KPIs to Troubleshoot Cycling 22915.5.3 KPIs for Oscillation Detection 23015.5.4 KPIs for Regulatory Control Issues 23015.5.5 KPIs for Measuring Operator Actions 23115.5.6 KPIs for Measuring Process Changes and Disturbances 23115.6 Exploitation of Constraints Handling and Maximization of MPC Benefit 23116 Commercial MPC Vendors and Applications 23516.1 Basic Modules and Components of Commercial MPC Software 23516.1.1 Basic MPC Package 23516.1.2 Data Collection Module 23616.1.3 MPC Online Controller 23616.1.4 Operator/ Engineer Station 23716.1.5 System Identification Module 23716.1.5.1 Different Modeling Options 23916.1.5.2 Reporting and Documentation Function 23916.1.5.3 Data Analysis and Pre‐Processing 23916.1.6 PC‐Based Offline Simulation Package 24016.1.7 Control Performance Monitoring and Diagnostics Software 24016.1.7.1 Control Performance Monitoring 24016.1.7.2 Basic Features of Performance Monitoring and Diagnostics Software 24016.1.7.3 Performance and Benefits Metrics 24116.1.7.4 Offline Module 24116.1.7.5 Online Package 24116.1.7.6 Online Reports 24116.1.8 Soft Sensor Module (Also Called Quality Estimator Module) 24216.1.8.1 Soft Sensor Offline Package 24216.1.8.2 Soft Sensor Online Package 24316.1.8.3 Soft Sensor Module Simulation Tool 24316.2 Major Commercial MPC Software 24316.3 AspenTech and DMCplus 24416.3.1 Brief History of Development 24416.3.1.1 Enhancement of DMC Technology to QDMC Technology in 1983, Regarded as Second‐Generation of MPC Technology (1980–1985) 24416.3.1.2 Introduction of AspenTech and Evolvement of Third‐Generation MPC Technology (1985–1990) 24516.3.1.3 Appearance of DMCplus Product with Fourth‐Generation MPC Technology (1990–2000) 24516.3.1.4 Improvement of DMCplus Technology for Quicker Implementation in Shop Floor, Regarded as Fifth‐Generation MPC (2000–2015) 24516.3.2 DMCplus Product Package 24616.3.2.1 Aspen DMCplus Desktop 24616.3.2.2 Aspen DMCplus Online 24616.3.2.3 DMCplus Models and Identification Package 24716.3.2.4 Aspen IQ (Soft Sensor Software) 24716.3.2.5 Aspen Watch: AspenTech MPC Monitoring and Diagnostic Software 24716.3.3 Distinctive Features of DMCplus Software Package 24816.3.3.1 Automating Best Practices in Process Unit Step Testing 24816.3.3.2 Adaptive Modeling 24816.3.3.3 New Innovation 24916.3.3.4 Background Step Testing 25016.4 RMPCT by Honeywell 25116.4.1 Brief History of Development 25116.4.2 Honeywell MPC Product Package and Its Special Features 25116.4.3 Key Features and Functions of RMPCT 25116.4.3.1 Special Feature to Handle Model Error 25116.4.3.2 Coping with Model Error 25216.4.3.3 Funnels 25216.4.3.4 Range Control Algorithm 25216.4.4 Product Value Optimization Capabilities 25216.4.5 “One‐Knob” Tuning 25316.5 SMOC–Shell Global Solution 25316.5.1 Evolution of Advance Process Control in Shell 25316.5.1.1 1975–1998: The Beginnings 25316.5.1.2 1998–2008: Shell Global Solution and Partnering with Yokogawa Era 25416.5.1.3 2008 Onward: Shell Returns to Its Own Application 25416.5.2 Shell MPC Product Package and Its Special Features 25516.5.2.1 Key Characteristics of SMOC 25516.5.2.2 Applications 25516.5.3 SMOC Integrated Software Modules 25516.5.3.1 AIDA Pro Offline Modeling Package 25616.5.3.2 md Pro 25616.5.3.3 RQE Pro 25616.5.3.4 SMOC Pro 25716.5.4 SMOC Claim of Superior Distinctive Features 25916.5.4.1 Integrated Dynamic Modeling Tools and Automatic Step Tests 25916.5.4.2 State‐of‐the‐Art Online Commissioning Tools 25916.5.4.3 Online Tuning 25916.5.4.4 Advance Regulatory Controls 26016.5.4.5 Features of New Product 26016.6 Conclusion 261Index 263