Izvorni znanstveni članak
https://doi.org/10.15255/CABEQ.2014.2163
Reactor Network Synthesis Using Coupled Genetic Algorithm with the Quasi-linear Programming Method
H. Soltani
; Department of Chemical Engineering, Ahar Branch, Islamic Azad University, Ahar, Iran
S. Shafiei
; Faculty of Chemical Engineering, Sahand University of Technology, PO Box 51335/1996, Tabriz, Iran
J. Edraki
; Department of Chemical Engineering, Ahar Branch, Islamic Azad University, Ahar, Iran
Sažetak
This research is an attempt to develop a new procedure for the synthesis of reactor networks (RNs) using a genetic algorithm (GA) coupled with the quasi-linear programming (LP) method. The GA is used to produce structural configuration, whereas continuous variables are handled using a quasi-LP formulation for finding the best objective function. Quasi-LP consists of LP together with a search loop to find the best reactor conversions (xi), as well as split and recycle ratios (yi). Quasi-LP replaces the nonlinear programming (NLP) problem, and is easier to solve. To prevent complexity and ensure
an optimum solution, two types of ideal reactors, namely plug flow reactor (PFR) and continuous stirred tank reactor (CSTR), were considered in the network. Since PFRs require the introduction of differential equations into the problem formulation, a CSTR cascade was used instead in order to eliminate differential equations. To demonstrate the effectiveness of the proposed method, three reactor-network synthesis case studies are presented.
Ključne riječi
reactor network synthesis; genetic algorithm; quasi-linear programming method; plug flow reactor; continuous stirred tank reactor
Hrčak ID:
161797
URI
Datum izdavanja:
14.7.2016.
Posjeta: 1.450 *