Multi-objective shape optimization of a heat exchanger using parallel genetic algorithms

Generalities and methods.- A Few Illustrative Examples of CFD-based Optimization.- Mathematical Aspects of CFD-based Optimization.- Adjoint Methods for Shape Optimization.- Specific Applications of CFD-based Optimization to Engineering Problems.- Efficient Deterministic Approaches for Aerodynamic Shape Optimization.- Numerical Optimization for Advanced Turbomachinery Design.- CFD-based Optimization for Automotive Aerodynamics.- Multi-objective Optimization for Problems Involving Convective Heat Transfer.- CFD-based Optimization for a Complete Industrial Process: Papermaking.

Optimization and Computational Fluid Dynamics

Aerodynamic optimisation has become an indispensable component for any aerodynamic design over the past 60 years, with applications to aircraft, cars, trains, bridges, wind turbines, internal pipe flows, and cavities, among others, and is thus relevant in many facets of technology. With advancements in computational power, automated design optimisation procedures have become more competent, however, there is an ambiguity and bias throughout the literature with regards to relative performance of optimisation architectures and employed algorithms. This paper provides a well-balanced critical review of the dominant optimisation approaches that have been integrated with aerodynamic theory for the purpose of shape optimisation. A total of 229 papers, published in more than 120 journals and conference proceedings, have been classified into 6 different optimisation algorithm approaches. The material cited includes some of the most well-established authors and publications in the field of aerodynamic optimisation. This paper aims to eliminate bias toward certain algorithms by analysing the limitations, drawbacks, and the benefits of the most utilised optimisation approaches. This review provides comprehensive but straightforward insight for non-specialists and reference detailing the current state for specialist practitioners.

/pdf/state-of-the-art-in-aerodynamic-shape-optimisation-methods-4rbh5x1ctl.pdf

State-of-the-art in aerodynamic shape optimisation methods

Review of design optimization methods for turbomachinery aerodynamics

A review on deep reinforcement learning for fluid mechanics

Hybrid method for aerodynamic shape optimization in automotive industry

Various global optimization methods are compared in order to find the best strategy to solve realistic drag reduction problems in the automotive industry. All the methods consist in improving classical genetic algorithms, either by coupling them with a deterministic descent method or by incorporating a fast but approximated evaluation process. The efficiency of these methods (called HM and AGA respectively) is shown and compared, first on analytical test functions, then on a drag reduction problem where the computational time of a GA is reduced by a factor up to 7.

/pdf/comparison-of-global-optimization-methods-for-drag-reduction-l09zch2q0f.pdf

Comparison of global optimization methods for drag reduction in the automotive industry

L'objet de cet article est de presenter une methode automatique, robuste et multicritere, d'optimisation de forme en aerodynamique automobile. Cette methode est appliquee a l'optimisation du coefficient de trainee ( C x ) d'une forme simplifiee de vehicule et parametree par les angles de lunette, de diffuseur et de retreint. Dans un premier temps, un algorithme, base sur la methode stochastique des algorithmes genetiques, est developpe. Ces methodes presentent l'avantage d'etre globales et multicriteres mais possedent un temps de convergence prohibitif vers la solution optimale. Pour pallier ce probleme, une procedure d'evaluation approchee, basee sur les reseaux de neurones, est couplee aux algorithmes genetiques. Une etude numerique est realisee sur la forme simplifiee de vehicule precedemment decrite, dans le but d'une part, de minimiser le coefficient de trainee de la forme etudiee et d'autre part, d'estimer le gain en temps de convergence apporte par la nouvelle methode proposee.

/pdf/optimisation-de-forme-en-aerodynamique-automobile-3e0enmvk8b.pdf

Optimisation de forme en aérodynamique automobile

Hybrid method for aerodynamic shape optimization in automative industry shape optimization. Various global optimization methods are constructed in order to be able to tackle realistic drag reduction problems in the automotive industry. All the methods consist in improving classical genetic algorithms, by coupling them with a fast but approximated evaluation process. The efficiency of these methods is shown, fisrt on analytical test functions, then on a drag reduction problem where the computational time of a GA is highly reduced.

/pdf/optimisation-de-forme-en-aerodynamique-automobile-3oqu0m2she.pdf

Frédérique Muyl

Papers

Hybrid method for aerodynamic shape optimization in automotive industry

Comparison of global optimization methods for drag reduction in the automotive industry

Optimisation de forme en aérodynamique automobile

Optimisation de forme en aérodynamique automobile