Engineering systems are designed to perform particular functions to a given specification. Optimisation is the process of finding the design which best achieves these goals or objectives. Until now, the aerodynamic design of missile systems has been performed using single-objective optimisation and semi-empirical models. An optimisation framework has been developed and tailored for the multi-objective optimisation of projectiles. The framework is demonstrated on three different configurations of projectile: body-tails, body-canards-tails and body-canards-tails with a rotating nose, with up to 7 input parameters and 6 objective functions. Incorporated in the framework are sampling schemes, Kriging surrogate modelling, Fourier decomposition, genetic algorithms and adaptive sampling. The MISL3 semi-empirical and Cart3D CFD aero-prediction codes are used to generate an aerodynamic performance dataset. Three aerodynamic optimisation problems are solved for objective functions based on lateral acceleration, time-to-target, range, static stability and roll moment coefficient. Quantitative validation is carried out on the Kriging surrogate model, determining that the mean RMS error is less than 3%. Adaptive sampling is applied to refine the surrogate model. The results are presented in the form of Pareto fronts showing the trade-off in two and three-dimensional objective space. This gives the designer key information on the trade-off and performance of a range of projectile designs. With the exception of range, all of the objective functions were found to be competing. A set of optimal designs were identified, with lateral accelerations of up to 40g. The framework was demonstrated to be robust when dealing with unexpected conditions such as failed simulations. Implementing such a framework in the design of missile systems provides the designer with a wide range of design options and has the potential to provide shorter time-to-market and cost savings for the industry.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:679038 |
| Date | January 2014 |
| Creators | Lisk, D. M. |
| Publisher | Queen's University Belfast |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
Page generated in 0.0017 seconds