A Method for Concept and Technology Exploration of Aerospace Architectures

Design;Graph;Ant colony;Technology;Concept

Villeneuve, Frédéric ; Aerospace Engineering

University:Georgia Institute of Technology

Department:Aerospace Engineering

关键词: Design;    Graph;    Ant colony;    Technology;    Concept;   

Others  :  https://smartech.gatech.edu/bitstream/1853/16212/1/villeneuve_frederic_200708_phd.pdf

美国|英语

来源: SMARTech Repository

【 摘 要 】

This dissertation presents the development of a new concept and technology exploration methodology for aerospace architectures. The methodology is based on modeling the design space by a graph, and optimizing the graph using Ant Colony Optimization. The results show that the proposed design methodology can explore more efficiently the concept and technology space of a launch vehicle architecture than traditional optimization approaches such as Genetic Algorithm and Simulated Annealing.The purpose of the method is to introduce quantitative and simultaneous exploration of concept and technology alternatives during the early phases of conceptual design. To achieve this goal, technical challenges such as expanding the size of the design space, exploring more efficiently the design options, and simultaneously considering technologies and concepts are overcome.The total number of design alternatives grows factorially with the number of concepts in the design space. Under these circumstances, the design space is difficult to explore in its totality. Considering more alternatives has been the focus of several researchers, using Genetic Algorithms and Simulated Annealing. The large number of incompatibilities between alternatives, however, limits these optimization algorithms and reduces the number of concepts or technologies that can be considered.To address these problems, a concept and technology selection methodology is developed. The methodology proposes a way to automatically generate aerospace architectures, and to model concept and technology incompatibilities by means of a graph. In conjunction with this new modeling approach, a graph-based stochastic optimization algorithm is used to efficiently explore the design space. This design methodology is applied to the simultaneous concept and technology exploration of an expendable launch vehicle architecture.This study demonstrates that the consideration of more design alternatives can help design engineers to make more informed decisions during the concept and technology selection process. Moreover, the simultaneous exploration of concepts and technologies has the potential to identify a different set of solutions than the standard approach where the technologies are explored after the concepts have initially been selected.