Premium
Comparing two neural networks for pattern based adaptive process control
Author(s) -
Cooper Douglas J.,
Megan Lawrence,
Hinde Ralph F.
Publication year - 1992
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.690380105
Subject(s) - controller (irrigation) , computer science , artificial neural network , control theory (sociology) , constant (computer programming) , process (computing) , internal model , adaptation (eye) , set (abstract data type) , adaptive control , artificial intelligence , pattern recognition (psychology) , control (management) , algorithm , physics , optics , agronomy , biology , programming language , operating system
An adaptation strategy based on an analysis of patterns exhibited in the recent controller error history is presented. The strategy is a two parameter adaptation, where the gain and time constant of the controller's internal model are adjusted to make the closed loop error response match a target pattern. Both a back‐propagation network and a vector quantizer network (VQN) are compared as pattern analysis tools. This strategy is established for a number of model based controllers and is demonstrated here using the Generalized Predictive Control algorithm. Details of this set point tracking strategy are presented along with demonstrations on both simulated and real single loop processes that experience significant changes in process gain and time constant. Results show both networks to be equally capable at pattern recognition with the VQN's ease of training and implicit ability to assess the accuracy of the pattern match as deciding factors in network selection.