1991 年 24 巻 1 号 p. 51-57
An on-line dynamic optimizing control procedure for the operation of a binary distillation column is proposed, and its performance was examined experimentally.
A linear input-output model for a multi-input/multi-output system was employed to represent the column, and its parameters were recursively updated using the instrumental variable method. The model was used in the prediction of output process values dynamically. The optimization of profit function was carried out using the golden section search method in every sampling step while the feed flow rate and the set point of top-tray temperature were varied. For the control of column operation, an adaptive control technique using quadratic programming was implemented.
A satisfactory application of the optimizing control procedure was obtained, and it was found that the procedure gives the optimal solution in a short time and is applicable to every sampling time of advanced control to maximize profit while the system is changing dynamically. Also, good convergence in the procedures of identification, optimization and adaptive control was observed.
