A Real-Time Approach to Process Control, 3rd Edition

11

Advanced Process Control

In this final chapter we tackle the subject of advanced process control (APC) as is widely practiced in industry. Over the past 20 years or so, APC has become synonymous with model predictive control (MPC). Thus we will review why this is so by first considering what the term advanced process control actually has come to mean and why, and then introduce the reader to the fundamental concepts of MPC and industrial implementation of MPC.

11.1 Advanced Process Control

The reasons that APC may be necessary include the multivariable nature of processes, significant variations in process disturbances, variations in process dynamics and the ability to handle constraints which means that the performance of single-loop conventional control systems is not sufficient. APC can offer solutions to these issues and substantial economic benefits (savings of the order of 2–6% of annual operating costs typically (and reported up to 15%) and generate an extra 1% in revenue [1,2]).

The term advanced process control means different things to different people. It is a relative term, to both the process and technology used. It is usually multivariable in nature, typically involves computer-based mathematical models of the process and can involve manipulation of the tuning constants (adaptive control) or model parameters (MPC) and on-line optimization; see, for example, [3].

Although admittedly subjective, Seborg [4] presented the following useful classification of process ...