Use this URL to cite or link to this record in EThOS:
Title: Effective multiobjective MDO for conceptual design : an aircraft design perspective
Author: Fantini, Paolo
ISNI:       0000 0001 3457 1512
Awarding Body: Cranfield University
Current Institution: Cranfield University
Date of Award: 2007
Availability of Full Text:
Access from EThOS:
Access from Institution:
Once the requirements for a new aircraft have been defined, the Conceptual design phase is launched. During this phase one or more designers have the goal of defining and investigating a number of alternative solutions. Through discussion with industry it has become apparent that optimisation tools are seldom used, even though these could greatly enhance the work of the designers. The objective of the work carried forward has been of identifying, comparing and where necessary improving the most suitable techniques, as well as schemes for their integration, in order to perform effectively Multidisciplinary Design and Optimisation (MDO) in the Conceptual phase of the aircraft design. The techniques that have been investigated include: multi-objective optimisation algorithms, MDO algorithms for treating non-hierarchically decomposable systems and Automatic Differentiation (AD). As a result an algorithm for performing multiobjective MDO has been developed. Given a complete model for a complex non-hierarchically decomposable system and given a number of objectives and constraints, the algorithm is capable of determining a set of well distributed solutions, representative of both local and global Pareto frontiers. A number of test cases have been used for evaluating the alternative methodologies and the proposed algorithm. These include a set of complex algebraic test cases typically used for evaluating global optimisation algorithms and a simplified aircraft conceptual design model, which was provided by industry. The results demonstrate the unique capability of the algorithm of determining well distributed solutions on the global and local Pareto frontiers for global multiobjective continuous nonlinear constrained optimisation problems. The results also show this capability when the algorithm is applied to non-hierarchically decomposable systems, as typically encountered when performing MDO. Further work could extend the approach in order to handle mixed discrete/continuous variables.
Supervisor: Guenov, Marin D. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available