Development and application of linear state estimators in control of reactive distillation.

摘要

Effective control and monitoring of a process requires sufficient information on the state of the process, which is uniquely specified by the process state variables. In practice, online measurements of all of the variables of a process are rarely available, and in such cases, reliable information on the immeasurable state variables is obtained by using the state estimator. This work presents the design, implementation and application of linear state estimators, which can infer the column composition from the temperature measurements or other process states in a reactive distillation process.; The accuracy of the developed estimators is checked by comparing the estimated states to the actual states as predicted by the process model of a reactive distillation system. The robustness and reliability of the linear state estimators are demonstrated against erroneous initial conditions, the measurement noise and plant-model mismatch. The estimator-based control system is developed and implemented on a reactive distillation process. The control performance of the system that relies on the estimator is examined and compared to that of the system which takes direct measurement from the process model using online perfect analyzer. It is found that a robust linear state estimator can be successfully designed and implemented on the feedback control of reactive distillation system.