Premium
Model‐independent algorithms for time‐optimal control of chemical processes
Author(s) -
Lin YeongIuan,
Beard John N.,
Melsheimer Stephen S.
Publication year - 1993
Publication title -
aiche journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.958
H-Index - 167
eISSN - 1547-5905
pISSN - 0001-1541
DOI - 10.1002/aic.690390608
Subject(s) - a priori and a posteriori , process (computing) , computer science , optimal control , spinning , algorithm , noise (video) , control theory (sociology) , control (management) , mathematical optimization , engineering , mathematics , artificial intelligence , mechanical engineering , philosophy , epistemology , image (mathematics) , operating system
The new time‐optimal control algorithms presented do not require a‐priori knowledge of the process dynamics. They are applicable to systems that can be described by a second‐order (or first‐order) plus deadtime model and generate a dynamic model as a byproduct. Simulations presented illustrate the use of these algorithms in time‐optimal control and the process models obtained. The algorithms are robust with respect to the measurement noise often encountered in real chemical processes and are inherently adaptive to changes in the process dynamics with time. An experimental application to the heatup of a laboratory fiber spinning apparatus is presented. Since these time‐optimal control algorithms are easy to apply and give rapid, smooth responses, they will be of value even when obtaining minimum‐time response is not a critical issue. Such applications include startup of continuous and batch processes. Moreover, they can be used to determine first‐ or second‐order process models (as appropriate) from open‐loop step response data without actually implementing time‐optimal control.