Title: A hybrid computing scheme for shape optimisation in thermo-fluid problems

Authors: Suman Ghosh, Dilip Kumar Pratihar, Biswajit Maiti, Prasanta Kumar Das

Addresses: Department of Mechanical Engineering, Indian Institute of Technology Kharagpur, Kharagpur 721302, West Bengal, India. ' Department of Mechanical Engineering, Indian Institute of Technology Kharagpur, Kharagpur 721302, West Bengal, India. ' Department of Mechanical Engineering, Indian Institute of Technology Kharagpur, Kharagpur 721302, West Bengal, India. ' Department of Mechanical Engineering, Indian Institute of Technology Kharagpur, Kharagpur 721302, West Bengal, India

Abstract: A hybrid computing scheme has been developed in the paper, which uses CFD softwares: Gambit for mesh generation, Fluent for carrying out hydrodynamic analysis, and employs a Genetic Algorithm (GA) for optimisation. It has been utilised to solve optimisation problems related to thermo-fluid field in a fully automated way by using specific system commands. In the developed hybrid computing scheme, the GA and CFD softwares combine each other seamlessly and data transfer takes places without any manual intervention. In this study, the main challenge lies in embedding highly structured commercial softwares, like Fluent and Gambit, inside an indigenously developed hybrid computing scheme. Using the said hybrid computing scheme, two different methodologies have been adopted to optimise the shapes of 2D symmetric nozzle. Corresponding to the optimised shapes of the nozzle, streamline patterns have been developed and variations in pressure loss (PL) factor values have been studied.

Keywords: hybrid computing; Fluent software; Gambit; genetic algorithms; GAs; nozzles; shape optimisation; CFD; computational fluid dynamics; thermofluids; streamline patterns; pressure loss.

DOI: 10.1504/IJCISTUDIES.2010.034886

International Journal of Computational Intelligence Studies, 2010 Vol.1 No.3, pp.207 - 226

Published online: 26 Aug 2010 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article