Intelligent control strategy for electrified pressure-swing distillation processes using artificial neural networks-based composition controllers

Daye Yang, Jingcheng Wang*, Huihuang Cai, Jun Rao, Chengtian Cui*

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

12 Downloads (Pure)

Abstract

This study introduces a novel artificial neural network (ANN)-based control strategy for pressure-swing distillation (PSD) systems, integrating heat pump-assisted distillation (HPAD) and self-heat recuperation technology (SHRT) to transition from thermally-driven to electrically-driven processes. While previous research has validated the dynamics and controllability of conventional PSD (PSD-CONV), PSD-HPAD, and PSD-SHRT for separating a maximum-boiling acetone/chloroform azeotrope, this work specifically focuses on enhancing product purity control through composition-temperature cascade control (CC-TC). Although similar control strategies have been proposed, our approach uniquely predicts temperature set points using easily measurable process variables, effectively bypassing the inaccuracies of composition measurements. Simulation results demonstrate that this ANN-based strategy significantly improves dynamic performance and adaptability in controlling product purity without requiring a composition analyzer. By leveraging the strengths of traditional Proportional-Integral-Derivative (PID) control alongside data-driven methods, this research highlights a critical advancement in the control of electrified PSD applications, paving the way for more efficient and reliable distillation processes.

Original languageEnglish
Article number130991
Number of pages16
JournalSeparation and Purification Technology
Volume360
DOIs
Publication statusPublished - 2025

Keywords

  • Artificial neural networks
  • Dynamics and control
  • Electrically-driven process
  • Intelligent composition control
  • Pressure-swing distillation

Fingerprint

Dive into the research topics of 'Intelligent control strategy for electrified pressure-swing distillation processes using artificial neural networks-based composition controllers'. Together they form a unique fingerprint.

Cite this