DEVELOPMENT OF AN ESTIMATOR FOR BATCH DISTILLATION COLUMN INFERENTIAL CONTROL

Abstract

Industrial interest in batch distillation has increased significantly in recent years as more batch processing is being used for low-volume high-value specialty chemicals. The operation and control of the batch distillation column is different particularly in predicting the time of switching from total-reflux operation to producing different product cuts and slopcuts of desired composition. However, in most industrial applications of batch distillation, perfect on-line composition measurements are not available. Due to this fact inferential control system that relies on secondary measurements, e.g., temperature, flow measurements must be used. Models of the batch distillation process can be used to provide estimates of the required compositions. Many articles dealing with batch distillation have appeared in the literature, but little attention has been paid to the problems of control. The objective of this work was to design an estimator for batch distillation system for inferential control and study the performance of estimator. Earlier some workers have developed estimators based on the steady-state and quasi-dynamic behaviour of the column. In the present work, a more realistic dynamic model of the estimator was studied. Computational technique and computer program were developed. The component material balance equations were written in tridiagonal matrix form. The resulting set of equations alongwith other constitutive equations of total material balance, equilibrium relationship and sum equations were solved by combination of two point implicit method and 0-method of convergence. Recurrence formulas and Newton-Raphson methods were applied to solve tridiagonal matrix and Jacobian matrix forms, respectively. Kb method has been used to update temperature profile. The response of the estimator was studied using the program and compared with actual values reported. It was observed that the model for estimator is good enough to predict reasonably accurate value of top product compositions of all the components using measured values of the tray temperatures. This estimator being a critical component of the inferential control closed-loop system of a batch distillation column may be further used for the design of the control system.