Exergetic optimization of shell and tube heat exchangers using a genetic based algorithm Yavuz O ¨ zc ¸elik * Ege University, Engineering Faculty, Chemical Engineering Department, Bornova, Izmir, Turkey Received 13 September 2006; accepted 5 January 2007 Available online 26 January 2007 Abstract In the computer-based optimization, many thousands of alternative shell and tube heat exchangers may be examined by varying the high number of exchanger parameters such as tube length, tube outer diameter, pitch size, layout angle, baffle space ratio, number of tube side passes. In the present study, a genetic based algorithm was developed, programmed, and applied to estimate the optimum values of discrete and continuous variables of the MINLP (mixed integer nonlinear programming) test problems. The results of the test problems show that the genetic based algorithm programmed can estimate the acceptable values of continuous variables and optimum values of integer variables. Finally the genetic based algorithm was extended to make parametric studies and to find optimum configuration of heat exchangers by minimizing the sum of the annual capital cost and exergetic cost of the shell and tube heat exchangers. The results of the example problems show that the proposed algorithm is applicable to find optimum and near optimum alternatives of the shell and tube heat exchanger configurations. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Exergetic optimization; Genetic algorithms; Shell and tube heat exchangers 1. Introduction Heat exchangers are used to transfer the heat between two fluids that are at different temperatures and shell-and tube heat exchangers (STHE) are the most common type of heat exchangers that find widespread use in refrigera- tion, power generation, heating, air conditioning chemical processes, manufacturing and medical applications. They consist of a bundle of tubes and enclosed within a cylindri- cal shell and one fluid flows through the tubes and another one flows between tubes and shell. There are many previous studies on the optimization of STHE’s. The concept of entropy generation minimization has been proposed by Bejan [1]. The extension to this approach which includes material use has been taken into account by Aceves-Saborio et al. [2], and a thermoeco- nomic analysis is used to determine the optimum baffle spacing by Eryener [3]. Several investigators have used different strategies based on simulated annealing (SA) [4] and GA, [5] for large scale combinatorial optimization problems for optimal design of STHE’s. They have developed a methodology based on command procedure to run the design program coupled with the SA and GA’s. In these studies, discrete variables have been varied using the SA and GA. In the present study, a large scale mixed integer nonlinear programming (MINLP) problem was considered taking into account the sizing and exergy cost of the STHE and a hybrid algorithm based on GA’s was developed to esti- mate the values of continuous variables of the problem effectively. 1359-4311/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.applthermaleng.2007.01.007 * Tel.: +90 232 388 40 00/1488; fax: +90 232 388 76 00. E-mail address: yavuz.ozcelik@ege.edu.tr www.elsevier.com/locate/apthermeng Applied Thermal Engineering 27 (2007) 1849–1856