- This event has passed.
OPTIMIZATION METHODS AND DIMENSIONALITY REDUCTION TECHNIQUES FOR AERODYNAMIC DESIGN
21/06/2022 | 09:00 - 11:30
21 and 23 June 2022
Speaker: Dr. Fernando Tejero, lecturer in Propulsion Systems Design at Cranfield University, Cranfield, UK.
Fernando Tejero was awarded a 5-year full-time diploma in Mechanical Engineering from the University of Zaragoza (Spain) in 2011. Upon the successful completion of his postgraduate studies, he pursued a PhD in the topic of Flow Control in Aeronautics at the “Institute of Fluid Flow Machinery – Polish Academy of Sciences” in 2015. Fernando is involved in a wide range of research areas: computational fluid dynamics, numerical methods, multidisciplinary design optimization, data-driven methods and machine learning. He has an international standing built upon a well-established research track record and has been involved in research projects within the FP7 EU and Horizon 2020 programmes.As part of his teaching activities, Dr. Tejero is leading the CFD for gas turbine module which is part of the Thermal Power MSc programme. He is also the module leader for the Thermofluids course within premaster’s in Engineering. Dr. Tejero has received several fellowships throughout his career. This includes two grants for young researchers approved by the Spanish government and a Marie Curie fellowship to undertake his doctoral studies.
Aerodynamic shape optimization is a complex task due to thehigh dimensionality of theproblem, the associated non-linearity and its large computational cost. These threeaspects have an impact on the overall time of the optimization routine. For this reason, emerging methods are being developed to accelerate the design process. They include advanced geometry parameterizations, low order modeling (e.g. Kriging interpolation or neural networks), dimensionality reduction and multi-fidelity techniques. The proposed activities for this scholarship will cover the main mathematical aspects of these methods and worked examples for aeronautical applications.
The activity consists of two presentations that address the following topics:
1) “Design and optimization methods for CFD” (2,5h)
– Different geometry parameterizations to describe complex shapes
– Definition of numerical methods to predict objective functions with numerical simulations (CFD) or low order models (Kriging, radial basis functions or neural networks)
– Description of optimization methods
– Application of all described techniques to a real-world problem.
2) “Dimensionality reductiontechniques for aerodynamicdesign” (2,5h)
– Dimensionality reduction for aerodynamic design, e.g. neural network auto-encoders or the active subspace method
– Multi-fidelity methods, e.g. co-kriging
– Application of all described techniques to a real-world problem.
21 June 2022
09: 00-11: 30 – Dr. Fernando Tejero, Design and optimization methods for CFD.
23 June 2022
09: 00-11: 30 –Dr. Fernando Tejero, Dimensionality reduction techniques for aerodynamic design.