Premium
Predicting conversion using temperature as the measured variable in the RIM process. Part I: Simulation and measurement structure
Author(s) -
Romagnoli J.,
Castro J.
Publication year - 1983
Publication title -
polymer engineering and science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.503
H-Index - 111
eISSN - 1548-2634
pISSN - 0032-3888
DOI - 10.1002/pen.760230210
Subject(s) - process (computing) , thermocouple , work (physics) , variable (mathematics) , field (mathematics) , line (geometry) , computer science , temperature measurement , state variable , process control , temperature control , control theory (sociology) , materials science , process engineering , control engineering , mechanical engineering , control (management) , engineering , mathematics , thermodynamics , mathematical analysis , geometry , physics , artificial intelligence , pure mathematics , composite material , operating system
In general the monitoring and control of many industrial processes is so complicated by problems associated with the on‐line measurement of the desired objectives that they must be inferred from available measurements. This leads to a state estimation problem in which the selection and adaptation of the structure of the measurements plays an important role. In particular, in the reaction injection molding (RIM) process, an accurate on‐line estimate of the conversion field is highly desirable. Since conversions cannot be determined readily by direct measurements, and a thermocouple can provide reliable dynamic temperature data, we can predict the conversion field from the solution of a state estimation problem using temperature as the measured variable. In this article, we describe an algorithm for designing the optimal arrangement of measuring sensors and analyze the RIM process dynamics which influence the structure depending on the operating conditions. The search for the optimal measurement structure for the purpose of state estimation makes up the bulk of the results. No particular estimation‐control strategy is investigated in this paper. Work is underway to develop the on‐line corrective system, which will use the temperature measurements to correct model predictions. The results of that work will appear in part II of this series.