Use this URL to cite or link to this record in EThOS:
Title: Combining shaping and flow control for performance optimization
Author: Zhang, Mai
ISNI:       0000 0004 6498 0095
Awarding Body: University of Oxford
Current Institution: University of Oxford
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Conventionally, shaping and flow control are treated as two separate disciplines. Both have been extensively studied and demonstrated to lead to very effective performance improvement for aerodynamic surfaces in the past. Despite of this, there are very few works that have combined the two together, a stark contrast to the huge number of publications within each discipline respectively. This thesis presents a first-of-the-kind systematic investigation into the combination of the two disciplines for performance improvement in a general sense. To provide efficient high-fidelity flow modelling as required by the flow control analysis, a novel method to efficiently utilize the large-eddy simulation (LES) to analyse the complicated flow is developed, successfully making the LES much more affordable to be used in the design optimizations, while maintaining the fidelity and accuracy of the LES solutions. A computational study is first carried out with 2D geometries and fast RANS performance evaluation as a starting point for the subsequent 3D investigation with the high-fidelity flow modelling. Five optimization approaches are systematically investigated, which are 1) shape optimization alone; 2) flow control optimization alone; 3) optimizing the shape then optimizing the flow control (sequential I); 4) optimizing the flow control then optimizing the shape (sequential II); 5) simultaneously optimizing the shape and the flow control (concurrent). The results show the combined optimizations of the two disciplines produce higher performance than the maximized performances from the conventional shape optimization alone or flow control optimization alone. In the combined optimizations, the concurrent approach produces higher performance than the two sequential approaches. Moving the investigation onto 3D realistic configurations, the novel efficient LES method is developed to provide efficient LES-based numerical solution of the complicated flow control flow field. As only the time-mean performance of the unsteady flow field is of interest, and as the time-mean flow field is time-dependent, an attempt to improve the computational efficiency has been proposed, which is obtain the time-mean LES solution by solving the steady-state equations instead of solving the unsteady LES governing equations. In doing so, the key is to obtain the correct turbulence correlation source terms (e.g. Reynolds stress gradients) for the steady-state equations for each design. A few LES runs are still required at a few initial sample designs to compute the initial set of correct source terms that would enable the steady-state simulations to reproduce the time-mean LES solutions at these sample designs. The methods to iteratively extract the source terms from the sample LES solutions and produce the converged steady-state solutions are developed. Then, various source term propagation methods are developed to adapt the sample source terms for the remaining designs in the design space, so that they can be evaluated with the fast steady-state simulations instead of the expensive time-accurate LES, whilst the steady-state solutions will still have a as high-fidelity and accuracy as the LES. The total computational cost thus can be greatly reduced comparing to the brute-force LES for all the designs. The developed efficient LES method is successfully implemented in the combined shape and flow control optimization of a 3D realistic blade trailing edge cutback injection configuration. The steady-state performance evaluations fit well with into the LES-predicted trend of variations, illustrating its ability of achieving a significant reduction of computational costs compared to the brute-force LES.
Supervisor: He, Li Sponsor: China Scholarship Council
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available