This presentation was made at the NAFEMS European Conference on Simulation-Based Optimisation held on the 15th of October in London.
Optimisation has become a key ingredient in many engineering disciplines and has experienced rapid growth in recent years due to innovations in optimisation algorithms and techniques, coupled with developments in computer hardware and software capabilities. The growing popularity of optimisation in engineering applications is driven by ever-increasing competition pressure, where optimised products and processes can offer improved performance and cost-effectiveness which would not be possible using traditional design approaches. However, there are still many hurdles to be overcome before optimisation is used routinely for engineering applications.
The NAFEMS European Conference on Simulation-Based Optimisation brings together practitioners and academics from all relevant disciplines to share their knowledge and experience, and discuss problems and challenges, in order to facilitate further improvements in optimisation techniques.
Thanks to their increased propulsion efficiency, ducted propellers are promising candidates for electrical aircrafts where the ratio of thrust and electricity consumption must be highly optimised. The aerodynamic performance and efficiency of a ducted propeller can be obtained through models and experiments of various fidelity, ranging from cheap analytic formulas to expensive Computational Fluid Dynamics (CFD) simulations.
The ducted propeller design problem involves several uncertainty sources, among which environmental conditions (that affect the blade loading under operation) or uncertainties due to the manufacturing process (e.g. material or geometrical imperfections that affect the performance).
Most optimisation procedures require a high number of performance analyses to find an optimal design, especially when uncertainties are taken into account. The (usually limited) computational budget makes impracticable the use of high-fidelity CFD simulations during the entire optimisation process. The use of surrogate models can alleviate this problem. However, the accuracy of results depends on the surrogate model quality. In principle, the surrogate model accuracy can be increased by increasing the training dataset size, but the computational budget limits the number of high-fidelity simulations. To alleviate this issue, the literature has recently proposed several multi-fidelity optimisation techniques. Cheap low-fidelity simulations are used to obtain more information about unexplored locations of the design landscape and information stemming from the low- and high-fidelity numerical experiments are fused together.
We propose a new approach for the design optimisation of the ducted propeller under uncertainty. The proposed approach combines different surrogate models for the design and probabilistic spaces, thus allowing a dimensional reduction which further reduces the computational cost. Several implementations are possible for both modellisations, e.g. [1, 2]. The design space is modelled with multi-fidelity techniques to approximate results arising from CFD simulations with meshes of different sizes.
The method is validated on simple test functions and on the ducted propeller use case.
|Date||15th October 2019|