Title :
Application of statistical clustering to mathematical description and control of continuous processes with discrete event output
Author :
Skormin, Victor A.
Author_Institution :
Watson Sch. of Eng., State Univ. of New York, Binghamton, NY, USA
Abstract :
Statistical clustering is used for the mathematical description of continuous manufacturing processes with discrete event outputs. The resultant mathematical model defines the correspondence between various outcomes of the process and particular regions in the space of process variables. An adaptive version of the clustering model intended for online implementation is presented. Prediction, regime selection, and control procedures, utilizing the clustering model, are formulated.<>
Keywords :
discrete time systems; manufacturing processes; pattern recognition; process computer control; statistical process control; continuous process control; correspondence; discrete event output; manufacturing; mathematical description; model; online; process variables; statistical clustering; Control systems; Delay; Discrete event systems; Force measurement; Human factors; Manufacturing processes; Mathematical model; Pattern recognition; Predictive models; Process control;
Conference_Titel :
Industry Applications Society Annual Meeting, 1992., Conference Record of the 1992 IEEE
Conference_Location :
Houston, TX, USA
Print_ISBN :
0-7803-0635-X
DOI :
10.1109/IAS.1992.244232
