On-line monitoring of process transient states by the intelligent multivariable filtering system based on the rule-based method

Abstract In this paper an intelligent multivariable filtering system (IMFS) based on the rule-based method is suggested for on-line monitoring of the transient behavior of chemical processes. Occasionally, it may be ineffective to observe the transient states of the chemical process only by a fixed filter. In that case, a few filters could be properly switched along with the inconsistent situation during the overall period of state estimation. A proper filter could be selected by an alliance between the inference engine and the knowledge database in IMFS. In practice, IMFS is established on SIMULINK for filtering a continuous polymerization reactor. Dynamic behavior of the continuous polymerization reactor was simulated stochastically, and the filters installed were switched to estimate the transient state of the process.