COMPUTER AIDED ENGINEERING
Computer Aided Engineering (CAE) is an attractive alternative for ascertaining and optimizing functional requirements of engineering systems as it enables simulations across virtual platforms.
An optimal Winglet design requires large number of Computational Fluid Dynamics (CFD) simulations for parametric and optimization studies. Genetic Algorithm can be used to optimize the winglet design. The Genetic Algorithms using cross-breeding and local mutations run iteratively on large population size to yield potential winglet designs.
A Winglet Optimization System on GARUDA was implemented with the collaboration of Zeus Numerix Private Limited. The algorithm runs on a machine and spawns a population of about 100 potential designs on the grid nodes (or clusters) to do CFD simulation so that each of them can be judge towards optimality. Simulation of 6000 winglets which was taking nearly 30 days sequential computing time was able to complete in about 3 hours by running concurrently over large computing resource of Garuda grid.