On Heuristic Computation and Application of Flexibility Indices for Unsteady Process Design

Yi Chung Kuo, Chuei Tin Chang

Research output: Contribution to journalArticlepeer-review

8 Citations (Scopus)

Abstract

The unsteady chemical processes, designed according to the nominal operating conditions and fixed model parameters, have traditionally been evaluated with economic criteria. This design strategy often ends up with a plant which may become inoperable if some of the process conditions/parameters unexpectedly deviate from their normal levels. Two alternative performance measures, the dynamic and temporal flexibility indices, have thus been devised in recent years to characterize designs from a different viewpoint. However, not only their significances still have not been adequately interpreted and their roles clarified, but also their numerical values cannot be determined efficiently with the available algorithms. In this study, the simple trapezoidal rule is applied to discretize and integrate the dynamic models and a heuristic strategy is then utilized to implement the vertex method for computing the aforementioned metrics efficiently. To demonstrate the effectiveness of the proposed computation approaches, comprehensive case studies concerning two dynamic systems, that is, the buffer system and the solar-driven membrane distillation desalination system (SMDDS), have been carried out. The resulting insights have also been summarized in a heuristic design procedure to incorporate both indices for general applications.

Original languageEnglish
Pages (from-to)670-682
Number of pages13
JournalIndustrial and Engineering Chemistry Research
Volume55
Issue number3
DOIs
Publication statusPublished - 2016 Jan 27

All Science Journal Classification (ASJC) codes

  • General Chemistry
  • General Chemical Engineering
  • Industrial and Manufacturing Engineering

Fingerprint

Dive into the research topics of 'On Heuristic Computation and Application of Flexibility Indices for Unsteady Process Design'. Together they form a unique fingerprint.

Cite this