^^his is a book about the design of distillation column control systems. It is written primarily from the standpoint of an engineering design organization, and is based on years of experience with large design projects as well as on personal plant experience. Most new investment dollars go into new or modernized facilities, and it is in the design phase of projects for these facilities that the most opportunities occur and flexibility exists to influence process control. Consequently this book is aimed primarily at design personnel. It is our hope, however, that it will also be useful to those who have to operate or troubleshoot existing plants.
Part I is an introduction, including a perspective on control and a brief review of fundamentals of distillation, with emphasis on topics that will be of interest to the control engineer rather than to the column design engineer. The distillation review, it is hoped, will be particularly useful to nonchemical engineers.
Part II of the book, on concepts and configurations, discusses some practical aspects of distillation control. Once the requirements for a particular column in a particular process are understood, design engineers must make at least a preliminary choice of equipment arrangements and control system configuration. In this section we have mostly avoided the use of mathematics and control theory. It is our hope that our discussions of equipment and control system arrangements will be useful to process engineers, production supervisors, maintenance engineers, and instrument engineers seeking guidelines, alternatives, and perspectives.
Part III focuses on the quantitative design of distillation control systems. It is aimed at professional control engineers and any others concerned with the numerical definition and specification of control system performance. Probably the most important development in process control system design since about 1950 was the evolution of a substantial body of theory and mathematics, plus a large catalog of control system studies. Together, these permit quantitative design of most process control systems with a considerable degree of multivariable control. It is the purpose of this book to indicate the range of this technology, which has been developed for distillation control, to the point where it can be economically and reliably used for design. The ultimate economic advantages include lower plant investment (particularly in tankage), lower operating costs, and closer control of product quality. For the most part, we have stayed with the modest theory of single-input, single-output (SISO) systems presented in previous books: Techniques of Process Control by P. S. Buckley (Wiley, 1964) and Process Modeling, Simulation, and Control for Chemical Engineers by W. L. Luyben (McGraw-Hill, 1973). This kind of theory and mathematics not only is adequate for noninteracting systems and for simple interacting systems, but it has the advantages of requiring minimum formal training and of permitting low design costs. "Modern" or "optimal" control techniques are mentioned only briefly here because their use on real, industrial-scale distillation columns has been quite limited to date. These techniques are still being actively researched by a number of workers, and it is hoped that they eventually will be developed into practical design methods. As of the date of the writing of this book, however, these mathematically elegant methods are litde used in industry because of their complexity, high engineering cost, and limitation to relatively low-order systems. Simulation techniques also are not covered since there are several texts that treat this topic extensively.
In the past five years, we have witnessed the introduction and proliferation of microprocessor-based digital controls of various sorts that are intended to replace analog controls. In fact, most of the newly installed control systems are of this type. In addition, we are seeing more control being implemented in process control computers. Sampled-data control theory has taken on new importance because of these developments and so we have included a chapter on previous work we have done in this area as it relates specifically to distillation columns. The concepts we present are quite basic as opposed to the recent advances in adaptative, multivariable, and predictive control, but we hope they will benefit those interested in synthesizing single-loop sampled-data controllers.
Many thanks are due our associates in the Du Pont Company, particularly R. K. Cox, and throughout the industrial and academic communities for helpful comments and suggestions. Many of the concepts presented in this book have been vigorously debated (over untold cans of beer) during the Distillation Control Short Courses held at Lehigh University every other spring since 1968.
We also wish to thank Leigh Kelleher for major assistance in formatting and editing, Arlene Little and Elaine Camper for typing, and Ned Beard and his Art Group for preparing the illustrations.
Page S. Buckley William L. Luyben Joseph P. Shunta
Was this article helpful?