Use this URL to cite or link to this record in EThOS:
Title: Impact of a hybridisation approach on the efficiency of boundary layer stability optimisation schemes
Author: McRoberts , Richard James
Awarding Body: Queen's University Belfast
Current Institution: Queen's University Belfast
Date of Award: 2013
Availability of Full Text:
Full text unavailable from EThOS.
Please contact the current institution’s library for further details.
With rising global fuel prices impacting on airline operating costs and increasingly stringent environmental targets being enforced on the aviation industry as a whole, interest in the implementation of laminar flow control techniques has been renewed. However, there are many difficulties associated with the design of practical profiles with extended regions of laminar f low, in particular the trade-off between empirical transition prediction methods versus the computational intensiveness of the more accurate physics based methods. The aim of the work documented in this thesis is to demonstrate how improvements in both the robustness and efficiency of the design of aerodynamic components with extended regions of laminar flow can be achieved through the use of a hybrid physics-based laminar flow optimisation approach, implementing the linear form of the Parabolised Stability Equations. Hybridisation of the optimisation algorithm to improve the efficiency of a global search genetic algorithm with an adaptive low fidelity local search is discussed, and benchmarking against the more traditional gradient based and genetic algorithms demonstrates that the bespoke algorithm consistently outperforms the traditional search methods. Initial tests indicated that the success of the algorithm could be enhanced through increased populations, and extension of the algorithm to use trust region methods to exploit the solution speed of the solver is demonstrated. This is achieved while retaining the accuracy of the parabolised stability evaluations to facilitate larger population searches without sacrificing solution robustness. The viability of using an adaptive low fidelity local search is ensured through an updatable N factor scheme, the accuracy of the lower fidelity solver is improved through iterating the N factor to give a solution closer to that had the high fidelity solver been implemented. In all instances tested, the newly developed algorithm resulted in increased quality of the profiles generated in relation to the defined objective function
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available