Get access
Advertisement

Translation of variables and implementation of efficient logic-based techniques in the MINLP process synthesizer MIPSYN

Authors

  • Marcel Ropotar,

    1. Faculty of Chemistry and Chemical Engineering, University of Maribor, Maribor, SI 2000, Slovenia
    2. Tanin Sevnica kemična industrija d.d., Sevnica, SI 8290, Slovenija
    Search for more papers by this author
  • Zdravko Kravanja

    Corresponding author
    1. Faculty of Chemistry and Chemical Engineering, University of Maribor, Maribor, SI 2000, Slovenia
    • Faculty of Chemistry and Chemical Engineering, University of Maribor, Maribor, SI 2000, Slovenia
    Search for more papers by this author

Abstract

This article describes alternative GDP formulation and convex hull representations for process synthesis problems and their implementation in a unique MINLP process synthesizer MIPSYN. A special translation of variables in mixed-integer, relaxed, and logic-based variations has been proposed, which enables modeling and solving process alternatives in a narrowed lifted space of variables, defined by nonzero lower and upper bounds. Based on these translation variations, alternative formulations have been developed for convex hulls, multiple-term generalized disjunctive programming problems, and logic-based outer-approximation algorithm, all of them being specialized for the synthesis of process flowsheets. Several studies were performed and three different large-scale synthesis problems were solved to test the performance and efficiency of different formulations. This initial research indicates that the proposed alternative convex hull representation usually outperforms the conventional one when solving both MILP and NLP steps in highly combinatorial MINLP process networks problems. © 2009 American Institute of Chemical Engineers AIChE J, 2009

Get access to the full text of this article

Ancillary