Files in this item



application/pdf9624321.pdf (6MB)Restricted to U of Illinois
(no description provided)PDF


Title:A hybrid direct-iterative linear solver for chemical process simulation
Author(s):Cofer, Haruna Nakamura
Doctoral Committee Chair(s):Stadtherr, Mark A.
Department / Program:Chemical and Biomolecular Engineering
Discipline:Chemical Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Engineering, Chemical
Abstract:This dissertation presents the results of developing a hybrid direct-iterative linear solver to solve the sparse linear equation systems arising in chemical process simulation. The motivation for this work is to enable simulations of large, complex, and realistic models requiring the computing power of high performance machines. The objective is carried out by combining the reliability of direct sparse linear methods (e.g., Gaussian elimination) with the efficiency of iterative sparse linear methods (e.g., Krylov subspace methods). The new hybrid solver is implemented in the SEQUEL-II, ASPEN PLUS$\sp{\rm TM}$, and SPEEDUP$\sp{\rm TM}$ programs and evaluated in terms of its overall effectiveness in reducing the total simulation time.
The results of the above computational experiments indicate that the hybrid solver is generally successful in reducing the total solution time of the SEQUEL-II and ASPEN PLUS$\sp{\rm TM}$ programs. The effect of the hybrid solver on improving the performance of SPEEDUP$\sp{\rm TM}$, though, is not definitive and is still unclear. Further work to improve the performance of the hybrid solver is, however, concluded to be both beneficial and necessary to solve future large-scale problems. In particular, the hybrid strategy may be very well-suited for traditional scalar workstations and symmetric multiprocessing computers.
Issue Date:1995
Rights Information:Copyright 1995 Cofer, Haruna Nakamura
Date Available in IDEALS:2011-05-07
Identifier in Online Catalog:AAI9624321
OCLC Identifier:(UMI)AAI9624321

This item appears in the following Collection(s)

Item Statistics