Sheet and film processes include coating, papermaking, metal rolling, and polymer film extrusion. Products produced by these processes include paper, bumper stickers, plastic bags, windshield safety glass, and sheet metal. The total capitalization of industries that rely on these processes is well over $ 500 billion worldwide. These processes are notorious for being difficult to control.
The goal of this book is to present the theoretical background and practical techniques for the identification and control of sheet and film processes. It is explained why many existing industrial control systems perform poorly for sheet and film processes. Identification and control algorithms are described and illustrated which provide consistent and reliable product quality. These algorithms include an experimental design technique that ensures that informative data are collected during input-output experimentation, model identification techniques that produce a process model and an estimate of its accuracy, and control techniques that take into account actuator constraints as well as robustness to model uncertainties.
The algorithms covered in this book are truly the state of the art. Variations on some of the algorithms have been implemented on industrial sheet and film processes. Other algorithms are in various stages of implementation. All of the algorithms have been applied to realistic simulation models constructed from industrial plant data; many of these studies are included in this book.

Inhalt

I. Background.- 1. Sheet and Film Processes.- 1.1 Plastic Film Extrusion.- 1.2 Papermaking.- 1.3 Coating.- 1.4 Metal Rolling.- 1.5 Book Organization.- 2. Process Characteristics.- 2.1 Traversing and Full-scan Sensors.- 2.2 Actuators.- 2.3 Process Dynamics.- 2.4 Interactions and Model Structures.- 2.4.1 Toeplitz Symmetric.- 2.4.2 Circulant Symmetric.- 2.4.3 Centrosymmetric.- 2.4.4 Pseudo-singular Value Decomposition.- 2.4.5 Modal Decomposition.- 2.4.6 Other Model Structures.- 2.5 Large-scale Systems.- 2.6 Model Uncertainty.- 3. Literature Review.- 3.1 Linear Control.- 3.2 Linear Control with Antiwindup Compensation.- 3.2.1 Directionality Compensation.- 3.2.2 Observer-based Compensation.- 3.2.3 IMC-based Antiwindup Compensation.- 3.3 Model Predictive Control.- 3.4 Robust Control.- 3.5 Profile Estimation.- 3.6 Model Identification.- 3.7 Process Monitoring.- 3.7.1 Fault Detection.- 3.7.2 Fault Isolation.- 3.7.3 Fault Compensation.- II. Identification and Control.- 4. Model Requirements and Process Identification.- 4.1 Model Requirements.- 4.2 Coupling Model Identification and Control.- 4.2.1 Model Identification.- 4.2.2 Controller Design.- 4.3 Simulation Studies.- 4.3.1 An Illustrative Example.- 4.3.2 Blown Film Extruder.- 4.4 Conclusions and Implications for Input Design.- 4.5 Proof of Theorem 4.1.- 5. Design of Experiments.- 5.1 Previous Research on Experimental Design.- 5.2 Process Gain Estimation.- 5.3 Problem Formulation for Constrained Input Design.- 5.3.1 Constraints.- 5.3.2 Modification of the Objective Function.- 5.3.3 Constrained Input Design Via Simulated Annealing.- 5.4 Simulation Case Study.- 5.4.1 Model Identification.- 5.4.2 Time Domain Simulations.- 6. Robust Control.- 6.1 Background.- 6.2 Optimal Robust Controller Design.- 6.2.1 Processes with General Interaction Matrices.- 6.2.2 Symmetric Nominal Models.- 6.2.3 Remarks.- 6.3 Low-order Robust Controller Design.- 6.3.1 LTI Uncertainty.- 6.3.2 SLTV, NLTV, NLTI, LTV Uncertainties.- 6.3.3 Multiplicative or Additive LTI Uncertainties.- 6.3.4 Implementation.- 6.4 Applications.- 6.4.1 Paper Machine Model.- 6.4.2 The Inadequacy of Commercial Software.- 6.4.3 Repeated Scalar Input and Output Uncertainties.- 6.4.4 Low-order Robust Controller Design.- 6.4.5 Full-block Input and Output Uncertainties.- 6.5 Summary.- 6.6 Proofs.- 7. Model Predictive Control.- 7.1 Problem Formulation.- 7.2 Fast MPC Algorithm.- 7.3 Industrial Paper Machine Simulation Study.- 7.4 Summary.- 8. Afterword.- References.