Use this URL to cite or link to this record in EThOS:
Title: Evolutionary computation based multi-objective design search and optimization of spacecraft electrical power subsystems
Author: Asif, Samina
ISNI:       0000 0004 2668 0630
Awarding Body: University of Glasgow
Current Institution: University of Glasgow
Date of Award: 2008
Availability of Full Text:
Access from EThOS:
Access from Institution:
Designing a spacecraft electrical power system (SEPS) is a complex and time-consuming engineering task that involves meeting several design objectives under constraints. A conceptual design of a spacecraft power system involves an optimal selection of available technologies for various components, such as solar cells, solar arrays, batteries, and bus voltages. Each technology has its own advantages and disadvantages that need to be taken into account in the search for an optimal design solution. This selection must meet certain criteria, the most important of which are cost-effectiveness, mass and performance. Traditionally, this task is a manual iterative process. At present, designs thus selected may not be realizable using the state-of-art design options available in the industry. However, advances in domain knowledge and in extra-numerical and multi-objective search techniques, such as evolutionary computation, offer a possibility of accelerating and improving this design cycle through a machine-automated design procedure. This thesis addresses the key issue of intelligent design automation and optimization of spacecraft power systems implemented in realistic design processes. The SEPS design is multi-objective in nature, a situation where a designer searches for solutions that are feasible with respect to all conflicting objectives. To facilitate the intelligent search process, meta-heuristics techniques are exploited in this work to provide computationally inexpensive design optimization. It extends the existing concept of computer-aided design to computer-automated design. To make the process of trade selection more efficient and reliable, a multi-objective design system for solving preliminary design problems for spacecraft electrical power subsystems is developed. It presents a system engineering framework that places design requirements at the core of the design activities. The thesis presents how simulation and optimization techniques can be used to automate and improve the design process of spacecraft power subsystems. The automated design procedure involves the design parameterization and the tools for system sizing and analysis. For the SEPS analysis, an inexpensive method for estimating design behavior is presented. Truly multi-objective and globally optimal design solutions are then artificially evolved as a result of interfacing evolutionary computation techniques with system sizing and analysis tools under practical constraints. Compared with conventional optimization techniques, the multi-objective design approach provides system designers with a clearer understanding of the effect of their design selections on all design variables simultaneously. In particular, the thesis extends a SEPS design problem from the basic technology selection to a detailed optimization based systematic design, which ensures the optimality and usability of designs from the beginning of the design process. Designs are made with implementation of solar cell modeling and parameter optimization using simulated annealing, which forms a very useful tool for simulating the behavior of solar arrays comprising of different types of solar cells. SEPS simulation is extended in MATLAB from existing work currently limited to Si solar cells and NiH2 batteries to a variety of solar cell and battery technologies. The thesis also develops a complete SEPS design and search framework, as a single tool and thus avoiding all compatibility issues involved. This feature makes this work very practical and efficient. It also keeps a way open for further improvements and modifications, both for optimization techniques and for the SEPS search space
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: TA Engineering (General). Civil engineering (General)