news and events
abrocrumbie
 
News
 
  Faculty
 
  Graduate Students
 
  Undergraduate Students
 
  Events
Seminars
Featured Items
Downloads
Rice Events
Links
 
 
news and events

Seminars

Process Operability

Professor Christos Georgakis
Department of Chemical and Biological Engineering
Tufts University

 

When: Thursday, November 9, 2006
Time: 2:30 PM to 3:30 PM
Where: 1070 Duncan Hall

 

Abstract:

It is well known that the effectiveness of a chemical process is dependant both on its steady state and its dynamic characteristics. Traditional approaches address the issue in a sequential manner, by first designing the plant from the steady state point of view and addressing the dynamic characteristics and the controller design tasks after the plant is built. This often results in a substantially over-designed and sub-optimal plant with dynamic and control characteristics that are difficult to handle with simple single-input-single-output controllers. Efforts to build more optimal and less polluting processes have resulted in the introduction of a substantial number of heat integration schemes and a large number of recycle streams, making the process more complex than before by several orders of magnitude. Such processes are very difficult to operate or are totally inoperable, unless their dynamic and control characteristics are examined very early in the design phase. The need to examine the interaction between process design and control has recently started to be recognized in the literature.


This seminar presents a new and systematic approach that examines the operability of a plant design prior to its physical construction. This enables the examination of the operability characteristics of several alternative designs and suggests design changes that improve process operability. It can also be used to quantify the characteristics of an existing plant to achieve increases in plant productivity and flexibility. The proposed operability measure is multivariable and nonlinear, based on concepts such as the Available Input or the Expected Disturbance Spaces and their transformations by either the steady state or dynamic model of the process. After the initial definition of the new concepts, the proposed approach is applied to several example unit operations such as reactors and multi-unit processes, concluding with the examination of a plant-wide problem. The extension of the operability concept to non-square systems is used to design the output constraints for Multivariable Model-based controllers.

 

 

CHEMICAL & BIOMOLECULAR ENGINEERING DEPT. MS-362
Rice University PO Box 1892
Houston, Texas 77251-1892
E-mail: chbe@rice.edu
Phone: (713) 348-4902
FAX:(713) 348-5478
rice university