Computational techniques for aircraft evolvability exploration during conceptual design.
| dc.contributor.advisor | Guenov, Marin D. | |
| dc.contributor.advisor | Molina-Cristobal, Arturo | |
| dc.contributor.author | Van Heerden, Albert Stevan Johan | |
| dc.date.accessioned | 2022-10-26T13:26:29Z | |
| dc.date.available | 2022-10-26T13:26:29Z | |
| dc.date.issued | 2018-02 | |
| dc.description.abstract | Evolvability is a critical consideration during the design of an aircraft. It refers to the extent to which a baseline design could be reused, or `easily' modified to create descendant designs that would meet future requirements. Since a major fraction of the cost of an aircraft programme is determined by decisions made during conceptual design, it is essential that the design space is explored thoroughly during this stage to find evolvable designs. Existing computational methods to perform such exploration exist, but are limited in two respects. The first of these is that, with existing techniques, derivatives are usually generated by applying pre-specified modifications to a selected baseline, such that each derivative in a study is only linked to a single baseline. The designer must therefore evaluate large numbers of baseline-derivative pairs to adequately capture the evolution options available. The second limitation concerns an absence of appropriate down-selection criteria to narrow down the number of design points when evolvability is considered. The work presented in this thesis addresses these limitations. The aim was to develop computational techniques that would enable aircraft designers to explore the evolvability of their designs more efficiently and effectively during the conceptual design stage. The scope was limited to civil transport aircraft and specifically to airframes. The work is applicable to both single- and twin-aisle aircraft, but the focus was, to a small degree, more on single-aisles. The research resulted in two main contributions: 1) a framework to provide a means to link all derivatives to all the baselines; and 2) a set of techniques to filter out inferior designs systematically. The framework builds on the premise that the degree of `similarity' between two ,designs could be used as an estimate for the redesign e ort (i.e. resource expenditure) required to change one of these into the other. Case studies involving existing aircraft families were conducted to determine which design changes could be considered `reasonable'. Based on this information, a set of techniques to assess airframe similarity was developed, which involves automatically predicting possible commonality across two designs. Several algorithms were devised to achieve this, including one that solves a longest common subsequence problem to find common body segments and a simple optimisation procedure to find common wing elements. Notably, these techniques can be used to compare aircraft with dissimilar configurations. For testing purposes, the framework was applied to several existing and future aircraft. The results showed that the predicted commonality matches published information regarding commonality and design re-use between designs. The framework essentially removes the need to model each future design option based on a specific starting design. The design filtering techniques involve the application of set-based design to facilitate systematic down-selection of potential designs. Specifically, it is demonstrated how established set-based design criteria could be adapted to prune an evolvability design space progressively. To demonstrate the usefulness of the research, it was applied to an example, concerning design candidates for a new single-aisle, environmentally friendly passenger aircraft. The results of this study were presented to a panel of design specialists from Airbus UK. The panel concluded that the proposed similarity assessment provides reasonable initial estimates for redesign e ort and that the overall approach adds value to the evolvability exploration process. | en_UK |
| dc.description.coursename | PhD in Aerospace | en_UK |
| dc.identifier.uri | https://dspace.lib.cranfield.ac.uk/handle/1826/18615 | |
| dc.language.iso | en | en_UK |
| dc.rights | © Cranfield University, 2015. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright holder. | |
| dc.subject | Aircraft conceptual design | en_UK |
| dc.subject | evolvability | en_UK |
| dc.subject | changeability | en_UK |
| dc.subject | set-based design | en_UK |
| dc.subject | knowledge-based engineering | en_UK |
| dc.subject | multi-attribute tradespace exploration | en_UK |
| dc.subject | longest common subsequence | en_UK |
| dc.title | Computational techniques for aircraft evolvability exploration during conceptual design. | en_UK |
| dc.type | Thesis | en_UK |