Tutorial - Simple Nonlinear Model-Based Process Control

Summary

ISA 2012 Automation Week Track: Control Performance Thursday, 27 September, 3:15-4:45 pm, Room 204B R. Russell Rhinehart School of Chemical Engineering, Oklahoma State University Often, nonlinearity is the primary problem for single-input-single-output (SISO) chemical process control, and practicable techniques for nonlinear control can ease design constraints and permit the operation of more competitive processes. If the engineer’s process knowledge is expressed as either a dynamic or steady-state nonlinear model, then process-model based control (PMBC) can be easily implemented in-house. Several simple process-model based control approaches have found success in industrial applications. SISO PMBC has several advantages over either PI(D) or gain scheduled PI(D). PMBC has a single tuning parameter, nonlinear compensation throughout the entire operating range, preservation of process knowledge, and continuous monitoring of the process—for health, predictive maintenance, and constraint recognition. This tutorial will show how to develop two effective SISO PMBC techniques. The PMBC techniques can be extended to multi-input-multi-output (MIMO) processes, where PMBC can decouple nonlinear interaction, balance deviations from setpoints when manipulated variable (MV) constraints are hit, and determine economic optimum MV values when there are extra degrees of freedom (DoF). However, the first step is being able to implement SISO versions. In contrast to the “large model,” multi-step-ahead, APC techniques characterized by horizon predictive controller or model predictive control, the one-step-ahead controllers presented here can solve many problems and can be implemented by a process engineer.   This tutorial will present the concepts behind two SISO nonlinear process-model based controllers that can be implemented in-house. One form of Generic Model Control (GMC) uses steady-state (SS) models, and ends up being implemented as output characterization to a standard PI controller. The integration of steady-state process models and conventional PI control makes this the simpler to implement.  The second approach, Process-Model Based Control (PMBC), uses a simple dynamic model of the process, which provides somewhat better control action, a single tuning parameter, and tracking of process characteristics for process condition monitoring.  Finally, this tutorial will show how to implement SISO and MIMO PMBC and handle the cases of excess and inadequate DoF.  This tutorial will show: 1.    The basis for the controller equations and examples of how they are implemented in Structured Text 2.    How to achieve bumpless MAN-AUTO transfer and output limits 3.    How to formulate equations to prevent executable errors 4.    How to choose tuning values 5.    Process characteristics that make nonlinear controllers preferable to linear controllers 6.    How to implement SISO PMBC 7.    The extension to MIMO Participants will receive the software demonstrations used in the session (VBA in Excel).       Who will benefit?  1.    Operations engineers who will recognize where nonlinear control will improve the operation 2.    Project groups that can implement advanced techniques 3.    Systems integrators who will see examples of how to implement the techniques within control equipment that permits programming is Structured Text 4.    Control system vendors who could develop attractive product features that permit users to use Structured Text programming, or provide a PMBC or GMC framework in which users could simply enter the models Session presenter Dr. R. Russell Rhinehart, professor in the School of Chemical Engineering at Oklahoma State University, holds the Amoco Endowed Chair and has experience in industry (13 years) and academe (26 years). He was Head of the School (1997 to 2008) and Interim Head (2011 to 2012). Rhinehart is president of the American Automatic Control Council. He is a Fellow of ISA, a CONTROL Automation Hall of Fame inductee, and received the 2009 ISA Distinguished Service Award. He was editor-in-chief of ISA Transactions (1998-2011). His 1968 B.S. in Chemical Engineering and subsequent M.S. in Nuclear Engineering are from the University of Maryland. His 1985 Ph.D. in Chemical Engineering is from North Carolina State University. He is coauthor of the textbook, Applied Engineering Statistics, and has authored several handbook chapters on modeling, process control, and optimization. He published about 50 journal articles and 65 full-paper refereed periodical or conference papers, five chapters, one book, and three patents. Rhinehart teaches modeling, optimization, and process control courses; and he has developed short courses for industrial participants offered through ISA or directly to companies related to statistical process control, instrument and control systems, modeling, and model-based control. Students working with Rhinehart implement the techniques in computer-controlled pilot-scale unit operations. He believes practical demonstration is the key to preparing students for practice careers.