Improved Multiobjective Particle Swarm Optimization Integrating Mutation and Changing Inertia Weight Strategy for Optimal Design of the Extractive Single and Double Dividing Wall Column

The extractive dividing wall column (EDWC) has received more and more attention because of the advantages of energy saving and high efficiency for separating mixtures with multiple azeotropes. Nevertheless, the optimization of the EDWC is challenging due to its highly nonlinear behaviors and inherent strong interactions caused by the decrease in the degree of freedom. This work proposes a multiobjective optimization framework that combines the particle swarm algorithm and the technique for order preference by similarity to the ideal solution to determine the optimal decision variable of the EDWC to improve economic performance. In this contribution, the particle mutation and linearly decreasing inertia weight strategies are introduced in the conventional multiobjective particle swarm optimization (MOPSO) to increase population diversity and feasible solutions for the decision-maker. The proposed optimization framework is validated through two case studies [i.e., EDWC for separating acetonitrile/N-propanol and extractive double dividing wall column (EDDWC) for separating tetrahydrofuran/methanol/water]. The results demonstrate that the improved MOPSO presents unique advantages in terms of maintaining population diversity compared to sequential iterative optimization and the genetic algorithm. Compared with the sequential iterative optimization, the total annual cost of the EDWC and EDDWC is respectively decreased by 12.34 and 36.03% via the proposed optimization strategy.

Automated summary

This study proposes a multiobjective optimization framework that combines the particle swarm algorithm and the technique for order preference by similarity to the ideal solution for improving the economic performance of the extractive dividing wall column (EDWC). The framework introduces particle mutation and linearly decreasing inertia weight strategies to increase population diversity and feasible solutions. The results demonstrate the unique advantages of the improved MOPSO in maintaining population diversity and reducing total annual cost compared to sequential iterative optimization.