A framework for the design optimisation of aerospace platforms using intelligent technologies

A systematic and computationally efficient method for the global shape optimisation of aerospace platforms is discussed. Soft computing methodologies and Intelligent Technologies are used extensively throughout the framework to enable the solving of complex multiple-objective and multi-disciplinary problems in the design of these platforms. These technologies include simulated annealing for design optimisation, heuristics for the handling of objectives and constraints in order to guide the optimiser to a feasible solution, neural networks for analytical system and experimental data representation and adaptive function evaluation to decrease the computation time of the objective function. The application of this framework to a complex design problem in aerospace engineering is detailed, being the design of quiet and efficient propellers. The interpretation of optimiser behaviour is achieved through the stochastic analysis of the design states produced by the optimiser during its search. This is performed using principal component analysis and linear regression, which are very effective in reducing large multidimensional data sets. These methods allow the designer to establish the effect of objectives and constraints in the cost function and previously unknown physical characteristics of the system can also be inferred. The framework is shown to succeed in producing effective designs significantly improving on their noise requirements whilst maintaining or improving performance