%0 Journal Article %A Kravanja, Z %A Grossmann, Ignacio E. %A Center., Carnegie Mellon University.Engineering Design Research %D 1992 %T PROSYN : an automated topology and parameter process synthesizer %U https://kilthub.cmu.edu/articles/journal_contribution/PROSYN_an_automated_topology_and_parameter_process_synthesizer/6467363 %R 10.1184/R1/6467363.v1 %2 https://kilthub.cmu.edu/ndownloader/files/11895815 %K Chemical engineering Data processing. %X Abstract: "This paper describes an improved, user friendly version of the computer package PROSYN -- a mixed-integer nonlinear programming (MINLP) process synthesizer. PROSYN is an implementation of the modeling and decomposition (M/D) strategy by Kocis and Grossmann (1989) and the outer approximation and equality relaxation algorithm (OA/ER) by Kocis and Grossmann (1987). Main characteristic [sic] of the new version of PROSYN is that it enables automated execution of simultaneous topology and parameter optimization of processes. Optimization of each NLP subproblem is performed only on the existing units rather than on the entire superstructure which substantially reduces the size of the NLP subproblems. In order to reduce undesirable effects of nonconvexities involved in the master problem the OA/ER algorithm has been improved by the use of an augmented penalty function. A simple process simulator has been built in to perform automated initialization of the first NLP step. A comprehensive PROSYN's library of models for basic process units and interconnection nodes, and a comprehensive library of basic physical properties for the most common chemical components have been developed. This enables to carry out [sic] an automated generation of a complex model representation for a superstructure in which the topology is specified by a concise interface. PROSYN allows to run in interactive mode and thus provides the user with a good control and supervision of calculations instead of having the procedure to be totally automated. PROSYN allows to carry out process synthesis at two basic levels of complexity: MINLP optimization through M/D and simultaneous heat integration including HEN costs. Applications with PROSYN are demonstrated with two example problems." %I Carnegie Mellon University