Back to  Predictive Control with Constraints  home page.
 
 

1. Introduction

Motivation
The `Receding Horizon' Idea
Computing the Optimal Inputs
A Simple Offset-free Tracking
Unstable Plant
Early History and Terminology
Predictive Control in the Control Hierarchy
General Optimal Control Formulation
What Is In This Book
Problems

 2. A Basic Formulation of Predictive Control

State-space models

Form of the model
Linear model, nonlinear plant
First-principles models and system identification

 A Basic Formulation
 General Features of Constrained Predictive Control
 Alternative State Variable Choices
 Allowing for Computational Delay
 Prediction

 No disturbances, Full state measurement
 Constant output disturbance
 Using an observer
 Independent and re-aligned models

Example: Citation Aircraft Model
Problems

3.  Solving Predictive Control Problems

 Unconstrained Problems

 Measured State, No Disturbances
 Formulation as a Least-Squares Problem
 Structure of the Unconstrained Controller
 Estimated State

 Constrained Problems

 Formulation as a QP Problem
 Controller Structure

 Solving QP Problems

 Active set methods
 Interior point methods

 Softening the Constraints
 Problems

4.  Step Response and Transfer Function Formulations

 Step and Pulse Response Models

 Step and Pulse Responses
 Relations to State-Space Models
 Prediction Using Step Response Model
 State-Space Models From Step Responses

 Transfer Function Models

 The Basics
 Prediction Using Transfer Functions
 Prediction with a Disturbance Model
 The GPC Model
 State-Space Interpretations
 Multivariable Systems

Problems

Measured Disturbances and Feedforward
Stabilised Predictions
Decomposition of Unstable Model
Non-quadratic Costs

 Characteristics of LP and QP problems
 Absolute value formulations
 Min-max formulations

 Zones, Funnels and Coincidence Points
 Predictive Functional Control (PFC)
 Continuous-time Predictive Control
 Problems

6.  Stability

 Terminal Constraints Ensure Stability
 Infinite Horizons

 Infinite horizons give stability
 Constraints and infinite horizons - stable plant
 Constraints and infinite horizons - unstable plant

 Fake Algebraic Riccati Equations
 Using the Youla Parametrization
 Problems

7. Tuning

 What Are We Trying To Do?
 Some Special Cases

 Effect of Control Weighting
 Mean-Level Control
 Deadbeat Control
 `Perfect' Control

 Frequency-Response Analysis
 Disturbance Models and Observer Dynamics

 Disturbance models
 Observer dynamics

 Reference Trajectory and Pre-Filter
 Problems

8.  Robust Predictive Control

 Formulations of Robust Control

 Norm-bounded uncertainty
 Polytope uncertainty

 The Tuning Procedure of Lee and Yu

 Simplified disturbance and noise model
 Tuning procedure

 The LQG/LTR Tuning Procedure
 LMI Approach

 Overview
 Robustness without constraints
 Robustness with constraints

 Robust feasibility

 Maximal output admissible sets
 Robust admissible and invariant sets

 Problems

9.  Two Case Studies

 Shell Oil Fractionator

 Process description
 Control specifications
 Initial controller design
 Controller performance and refinement
 Constraint softening
 Robustness to model errors

 Newell and Lee Evaporator
 Problems

 Spare Degrees of Freedom

 Ideal resting values
 Multiobjective formulations
 Fault tolerance

 Constraint Management
 Nonlinear Internal Models

 Motivation and approaches
 Sequential quadratic programming
 Neural net models
 Sub-optimal nonlinear MPC

 Moving-Horizon Estimation
 Concluding Remarks

 References

Appendices:

A.  Some Commercial MPC Products

Aspentech:  DMCPlus
Honeywell:  RMPCT
Simulation Sciences: Connoisseur
Adersa: PFC and HIECON
ABB:  3dMPC
Pavilion Technologies Inc: Process Perfecter

C. The MPC Toolbox

General remarks
Functions  scmpc2 and scmpcnl2
Functions  scmpc3 and scmpc4

Back to  Predictive Control with Constraints home page.

This page updated 22 June 2001.