Advances in Optimal Design of Shapes and Processes by Computational Fluid Dynamics
Immonen, Eero (2023-10-09)
Väitöskirja
Immonen, Eero
09.10.2023
Lappeenranta-Lahti University of Technology LUT
Acta Universitatis Lappeenrantaensis
School of Energy Systems
School of Energy Systems, Energiatekniikka
Kaikki oikeudet pidätetään.
Julkaisun pysyvä osoite on
https://urn.fi/URN:ISBN:978-952-335-971-0
https://urn.fi/URN:ISBN:978-952-335-971-0
Tiivistelmä
In many engineering applications, it is desirable to design objects which display good performance characteristics when interfacing with fluid flow, either internal or external to the objects. For example, an aircraft wing designer attempts to maximize lift while simultaneously maintaining a small wing drag (external flow). On the other hand, S-duct diffusers are designed to produce maximal pressure recovery and as uniform a flow as possible towards a compressor inlet (internal flow).
During the past decades, advances in simulation software and proliferation of parallel computing have made it possible to use Computational Fluid Dynamics to carry out such design optimization. However, the integration of rigorous computational optimization to industrial design processes is an elaborate exercise and is typically feasible only for sufficiently mature products with incremental development over many years. Yet, in many practical instances, the design improvements should be available much faster, even if the application domain was new to the designer and the reasons for any performance issues were unknown at the outset. Such design challenges are, in practice, still typically resolved by relying on human experts.
This thesis addresses optimal design of fluid systems, shapes and processes using a computational approach that is largely automatable and covers both fluid representations and numerical optimization. The proposed approach is based on parametric surrogate models, identified by using Computational Fluid Dynamics simulations. Besides fluid flow, the proposed approach can also accommodate practical design objectives and constraints arising from manufacturability, structural mechanics or business concerns. Moreover, besides shape optimization, the proposed approach can address layout design and identification of redundant parts in fluid systems. The wide applicability of the proposed methodology is demonstrated in six case study publications.
During the past decades, advances in simulation software and proliferation of parallel computing have made it possible to use Computational Fluid Dynamics to carry out such design optimization. However, the integration of rigorous computational optimization to industrial design processes is an elaborate exercise and is typically feasible only for sufficiently mature products with incremental development over many years. Yet, in many practical instances, the design improvements should be available much faster, even if the application domain was new to the designer and the reasons for any performance issues were unknown at the outset. Such design challenges are, in practice, still typically resolved by relying on human experts.
This thesis addresses optimal design of fluid systems, shapes and processes using a computational approach that is largely automatable and covers both fluid representations and numerical optimization. The proposed approach is based on parametric surrogate models, identified by using Computational Fluid Dynamics simulations. Besides fluid flow, the proposed approach can also accommodate practical design objectives and constraints arising from manufacturability, structural mechanics or business concerns. Moreover, besides shape optimization, the proposed approach can address layout design and identification of redundant parts in fluid systems. The wide applicability of the proposed methodology is demonstrated in six case study publications.
Kokoelmat
- Väitöskirjat [1038]