Nassim Khaled

Introducing the Book

This chapter introduces the authors briefly. Both the authors are academic and industrial experts who learned Model Predictive Control (MPC) on their own and implemented it in industrial applications. They have gone through the pain of failed designs and tunings in their industrial experiences. They have learned coding tricks, automated multiple MPC design techniques as well as robustness best practices that they wanted to share with the industrial and academic world. The chapter also describes the organization of the book and hardware and software requirements to implement the examples in the book, in addition to the free resources available for the reader.

Real Time Embedded Target Application of MPC

This chapter guides the user through the process of controlling the speed of a DC motor using model predictive control (MPC). The controller is implemented in real-time hardware (Arduino). The System Identification and Controller Design steps are performed similar to previous chapters. The emphasis of this chapter is to outline the process to deploy MPC in the hardware in order to control the motor speed in real time. Technical challenges of embedding the controller are highlighted as well as mitigation plans. Performance of the MPC controller to track a varying speed reference is analyzed in further details.

Single MPC Design for a Ship

This chapter guides the reader through the process of designing a linear Model Predictive Controller (MPC) for a ship. The turning rate and ship speed are controlled using the rudder and propeller. These actuators operate under physical constraints which will be used in the design of the controller. Tuning challenges and hurdles that might face the designer are outlined along with recommended solutions. The chapter concludes with an application problem that challenges the reader to implement the methods discussed. All the codes used in the chapter can be downloaded from MATLAB file exchange. Alternatively, the reader can also download the material and other resources from the dedicated website or contact the authors for further help. https://www.practicalmpc.com/

System Identification for a Ship

In this chapter, a dynamic model for a ship navigating in sea is introduced. The nonlinear model of the ship is used to collect maneuvering data. System identification is performed on the data using Matlab System Identification Toolbox. Details and best practices to do system identification are shared. The resulting linear model of the ship forward speed and turning rates is used in later chapters for MPC design.

Multiple MPC Design for a Ship

In most industrial applications, the dynamics of the plants is nonlinear. Despite nonlinearity, PID is the default controller in many of these cases. It is chosen without proper analysis of nonlinearity. In this chapter, the nonlinearity will be studied before the design of the Model Predictive Control (MPC) controller. The response surface across the range of operation of the plant will be generated. This surface will give insight into the number of MPC controllers needed for the full range. Without loss of generality, the plant that is being considered is that of a ship. The process, however, applies to other plants. Multiple simulations for the ship are carried out by using the Parallel Computing Toolbox. Methods such as changing Simulink parameters from the script, creating new file names, and checking for MATLAB licenses are introduced. The linearized models of the ship for various operating ranges are generated. Using these models, multiple MPC controllers are created and integrated with the ship model. Tracking results and controller performance is also analyzed. We identify and study the problem of frequent switching of MPC modes. A hysteresis logic is introduced to mitigate the problem. The limitations of such an approach are highlighted. Finally, this chapter concludes with an application problem to reinforce the learnings.

MPC Design for Air-Handling Control of a Diesel Engine

Abstract In this chapter, we showcase one of the industrial software that is used for industrial MPC design and deployment. We use Honeywell’s Automotive software, OnRamp Design Suite conjunction with MATLAB and Simulink. We design and deploy a Model Predictive Controller (MPC) for a diesel engine with dual exhaust gas recirculation (EGR) loops. This engine architecture requires the control of two EGR valves (high pressure and low pressure), an exhaust throttle and a variable geometry turbocharger. Following the same process reiterated throughout the book, we design and deploy the production-ready MPC. We also run robustness analysis to challenge the final design.