Multi-Objective Optimal Control of Autocatalytic Esterification Process Using Control Vector Parameterization (CVP) and Hybrid Strategy (HS)

Abstract

The semi-batch esterification of propionic anhydride (PA) with 2-butanol (BT) in the presence of catalyst can be optimised using an optimal control strategy, which utilises the reactor temperature (TR) and feed (FR) flowrate. However, the opposing objective functions, which are maximum conversion (XM) and minimum process time (tf) in the autocatalytic esterification process, could complicate the optimisation strategy. Simultaneous optimisation of various objectives results in a multi-objective optimal control (MOOC) problem with numerous solutions known as non-dominated (ND) points. In this paper, control vector parameterisation (CVP) and hybrid strategy (HS) are utilised to form Pareto Front (PF) for two opposite targets, which are first to increase XM and secondly to reduce tf. Each ND point comprises variant optimal dynamic tracks of TR and FR, which results in various targets of XM and tf. These solutions provide numerous options for evaluating trade-offs and deciding on the most efficient operating strategy. It is found that the ND point in zone II can be selected as the trade-off of the optimal TR and FR in this study.